/ Notes from the workshop
The blog.
Practical notes on the technology behind automation, business systems, and lead capture, written for owners of any kind of business.
Nuance Labs Emma: what it is and how to try it
Nuance Labs Emma is an AI video persona that reacts while you talk. What Nuplex 1.0 does, how to try it, the limits, and why Illinois is blocked.
2026-11-10 · 4 min read
CallRail pricing 2026: plans, usage fees and the real bill
What each CallRail plan costs per month, what's included, the per-number and per-minute fees that sit on top, and what a small service business still does by hand. Prices from CallRail's own page, checked November 10, 2026.
2026-11-10 · 7 min read
Spacefast: what it is, pricing and how to use it
Spacefast is Automattic's new host for sites and apps your AI builds. What it does, what each plan costs, and how to publish from Claude or ChatGPT.
2026-11-09 · 4 min read
Smith.ai alternatives: 5 answering services, priced
Five Smith.ai alternatives for a small trades business, with plans, overage rates and terms read from each vendor's own page on November 9, 2026, and what two call volumes would cost on each.
2026-11-09 · 9 min read
Muse voice: how to call and video chat with your Muse
Muse voice lets US users call or video chat with Meta's AI agent. How to start a call, what each view does, and what Meta hasn't said about cost.
2026-11-09 · 4 min read
Muse Ember: what it is, pricing and how to use it
Muse Ember is Meta's smallest Muse model, built for agent work at $0.10 per million input tokens. What it does, what it costs and how to call it.
2026-11-09 · 4 min read
Make vs n8n: cost, hosting and which one to pick
Make vs n8n for a small business: what each plan costs, how each one counts usage, who hosts it, and which suits an office with no IT person. From both vendors' own pages, checked November 9, 2026.
2026-11-09 · 9 min read
GPT-6.2 Sol: what it is and how to tell if you have it
GPT-6.2 Sol is an unannounced OpenAI model some ChatGPT and Codex users think they already have. What's known, and how to test it yourself.
2026-11-09 · 4 min read
ElevenAgents Campaign Manager: what it is and how to use it
ElevenAgents Campaign Manager runs outbound AI call and WhatsApp campaigns toward a set goal. How it works, how to set it up, and the rules to check.
2026-11-09 · 4 min read
Claude Fable 5.2: what it is, pricing and how to use it
Claude Fable 5.2 is Anthropic's new model for hard technical work at $8 and $40 per million tokens. Plan limits, effort settings and how to use it.
2026-11-09 · 4 min read
Booksy pricing 2026: per-user cost, card fees and Boost
What Booksy costs a barbershop or salon, from its own pricing page on November 9, 2026: the $29.99 base, $20 per added team member, card rates, the Boost commission, and the math for a solo chair and a three-chair shop.
2026-11-09 · 8 min read
NiceJob pricing 2026: Starter, Pro and add-on costs
NiceJob lists Starter at $95 a month and Pro at $135 a month. What each plan includes, what the add-ons cost, and when a small crew should pay for it. Checked on NiceJob's own page November 8, 2026.
2026-11-08 · 7 min read
Muse Active Memory: what it is and how to test it
Muse Active Memory helps Meta's assistants remember what matters and drop what's stale. What changed, who gets it, and a simple way to test it.
2026-11-08 · 4 min read
Grok Bot Data Bot: what it is and how to set it up
Data Bot is a shared Grok Bot that answers data warehouse questions in chat and Slack. What it does, who can use it and how to set one up.
2026-11-08 · 5 min read
Dialpad pricing 2026: Connect, Sell and Support plans
What each Dialpad plan costs per user, what Connect Standard and Pro include, the fees the pricing page leaves out, and what happens to a missed caller. Prices from Dialpad's own pages, checked November 8, 2026.
2026-11-08 · 9 min read
DeepMind 1024-bit RSA: what it found and how to check keys
Google DeepMind's AI agents found two weaknesses in legacy 1024-bit RSA. What they did, what's safe, and how to check your email and SSH keys.
2026-11-08 · 4 min read
Zapier alternatives for small business: 2026 prices
Five Zapier alternatives a small business can actually run, what each one costs on its own pricing page, how each counts usage, and when a small custom build is the better answer. Prices checked November 7, 2026.
2026-11-07 · 9 min read
Weave alternatives for dental offices: 2026 prices
Six Weave alternatives for a dental office, with prices, contract terms and features taken from each vendor's own page on November 7, 2026, and how to choose by what you are actually replacing.
2026-11-07 · 9 min read
Perplexity Teammates: what it is, pricing, how to use it
Perplexity Teammates are always-on work agents for Max subscribers. What they do, what they cost, how to set one up, and where they still fall short.
2026-11-07 · 4 min read
Muse Spaces: what it is and how to use it
Muse Spaces lets you describe a tracker, planner or dashboard and Meta's Muse builds it. How to use it, what it connects to, and what's missing.
2026-11-07 · 4 min read
Mistral Studio managed agents: how to use them
Mistral Studio managed agents are hosted AI agents built with Mistral's Agents SDK. How to install it, run modes, approvals and what's still in preview.
2026-11-07 · 4 min read
METR Anthropic report: what it found about AI speedup
METR spent three months with Anthropic's internal data and found its models improving about twice as fast as before, mostly thanks to AI.
2026-11-07 · 4 min read
Kiro bots: what they are and how to use them
Kiro bots are long-running AI agents inside AWS's open source Kiro Crew, each owning one job like triage or build health. How to set one up and what it costs.
2026-11-07 · 4 min read
HackerTube: what it is and how Karpathy built it with Claude
HackerTube is Andrej Karpathy's free YouTube alternative built from 6,454 videos Hacker News keeps recommending. How to use it and how he built it.
2026-11-07 · 4 min read
Calendly vs Acuity for salons and med spas: 2026 prices
Calendly vs Acuity Scheduling for a salon or med spa: what each costs per seat or per calendar, which plan adds text reminders, deposits and a HIPAA agreement, from both vendors' own pages checked November 7, 2026.
2026-11-07 · 8 min read
Arete-27B: what it is and how to run it
Arete-27B is Nous Research's free fine-tune of Qwen3.8-27B that cuts state-aligned answers from 22% to about 1%. What changed and how to run it.
2026-11-07 · 4 min read
Thryv pricing 2026: plans, contract terms, what you'll pay
What Thryv's Starter, Signature and Business Center cost, what each includes, the fees on top, and how cancellation works. Prices and terms from Thryv's own site, checked November 6, 2026.
2026-11-06 · 9 min read
Tencent Hy4: what it is, pricing, how to use it
Tencent Hy4 adds vision and computer use at $0.834 per million input tokens. What it does, what it costs and how to try it today.
2026-11-06 · 4 min read
Podium alternatives: what each costs and what it covers
Five Podium alternatives for a local service business, with prices and limits taken from each vendor's own page on November 6, 2026, and how to choose by the job you need done.
2026-11-06 · 8 min read
Mindbody alternatives: what 5 rivals cost in 2026
Five Mindbody alternatives with prices from each vendor's own page, checked November 6, 2026: Vagaro, Fresha, GlossGenius, Boulevard and WellnessLiving, plus when staying on Mindbody is the better call.
2026-11-06 · 10 min read
Inkling 2 Small: what it is, pricing, how to use it
Inkling 2 Small is Thinking Machines' new 594B model. What it costs on Tinker, when the weights arrive, and how to try or fine-tune it today.
2026-11-06 · 4 min read
GPT-6 Cyber: what it is and who can use it
GPT-6 Cyber is OpenAI's gated security model, built on GPT-6 Astra. See its test scores, the Daybreak Red requirements, and who can use it today.
2026-11-06 · 4 min read
Gemini Omni 1.2 Flash: what it is, pricing, how to use it
Gemini Omni 1.2 Flash makes 15-second AI video at $0.07 per second. What it costs by resolution, where it's free, and how to run your first clip.
2026-11-06 · 4 min read
Cursor Build Live Apps: how to use it, what it costs
Cursor Build Live Apps turns a prompt into a hosted app with Postgres and sign-in. How to use it, which plans include it, and what's still unpriced.
2026-11-06 · 4 min read
Claude Personal Finance: what it is and how to use it
Claude Personal Finance links your US bank accounts through Plaid with read-only access. Who gets it, how to set it up, and what stays private.
2026-11-06 · 4 min read
ChatGPT Images 2.6: what it is and how to make GIFs
ChatGPT Images 2.6 adds animated GIFs and saved reference images. How to make a GIF, what Flare and Sunburst cost, and the API limits that matter.
2026-11-06 · 4 min read
Weight-Space Programming: what it is and how to try it
MIT researchers say GPT-6 Astra can read and edit a neural network's raw weights. What the experiments showed, the pushback, and how to run one yourself.
2026-11-03 · 4 min read
Siri AI Developer Preview: what it is and how to use it
Apple's Siri AI Developer Preview lets any app hand its actions to Siri through App Intents and SKILL.md skills. What it does, how to try it, the limits.
2026-11-03 · 4 min read
RingCentral pricing 2026: plans, texting limits, add-ons
What RingCentral's Core, Advanced and Ultra plans cost per user, how many business texts each one includes, and which add-ons and fees land on top, taken from RingCentral's own pricing page on November 3, 2026.
2026-11-03 · 7 min read
Make pricing 2026: plans, credits and what you'll pay
What Make's Free, Core, Pro, Teams and Enterprise plans cost, how credits are counted, what happens when you run out, and how to estimate a small business bill. From Make's own pages, checked November 3, 2026.
2026-11-03 · 9 min read
GDPx benchmark: what it is and how each AI model scored
GDPx is SemiAnalysis's new agent benchmark: 206 expert tasks in 11 fields. Claude Opus 5.5 leads at 12%. The full leaderboard, costs and what it means.
2026-11-03 · 4 min read
Claude message threads: what they are and how to use them
Claude message threads let agents send only new turns plus the last response ID. How the beta works, the code, the 5-minute expiry and what it costs.
2026-11-03 · 4 min read
Anthropic FlashProve: what it is and how to try it
FlashProve is Anthropic's open-source tool that proves which AI model wrote a reply. What it does, what it costs to run, and how to try the verifier.
2026-11-03 · 4 min read
FreshBooks pricing: Lite, Plus, Premium and Select explained
What FreshBooks Lite, Plus, Premium and Select cost, what the client limits mean, which add-ons raise the bill, and which parts of getting paid still need a person. Checked on FreshBooks' own page November 2, 2026.
2026-11-02 · 7 min read
HoneyBook pricing: Starter, Essentials and Premium explained
What HoneyBook's Starter, Essentials and Premium plans cost, what each one includes, what card and bank payments cost on top, and what a small service business still does by hand. Checked on HoneyBook's own page November 1, 2026.
2026-11-01 · 7 min read
PracticePanther pricing 2026: plans and where intake starts
What each PracticePanther plan costs per user, monthly and annually, what Solo, Essential, Business and Business Pro add, and what a small firm still handles by hand. Prices from PracticePanther's own page, checked October 31, 2026.
2026-10-31 · 8 min read
Pika H3 Flex: what it is, pricing and how to use it
Pika H3 Flex is an open-weight video and audio model built on MiniMax H3. What it can do, what it costs per second, and the license catch for US users.
2026-10-31 · 4 min read
Claude Code subagent effort: how to use the new setting
Claude Code 2.1.316 adds CLAUDE_CODE_SUBAGENT_EFFORT, so subagents can run at low effort while your main session stays high. How to set it.
2026-10-31 · 4 min read
Claude API 128 MB limit: what changed and how to use it
The Claude API now takes requests up to 128 MB on Messages and Token Counting, up from 32 MB. What counts, what doesn't, and how to use the room.
2026-10-31 · 4 min read
Xiaomi-OCR-0: what it is and how to run it
Xiaomi-OCR-0 is a free, Apache 2.0 document model with 0.8B parameters that reads pages, tables and receipts. What it does and how to run it locally.
2026-10-30 · 4 min read
Workiz pricing: plans, features and what to ask sales
What Workiz's Standard, Pro and Ultimate plans include, why none of them shows a price, which add-ons are sold separately, and the questions to bring to the demo call. Checked on Workiz's own page October 30, 2026.
2026-10-30 · 8 min read
__usage_ai-training robots.txt: how to block AI training
Google, Microsoft, OpenAI and Anthropic will honor one robots.txt line, __usage_ai-training, as an AI training opt-out. The syntax, dates and gotchas.
2026-10-30 · 4 min read
OpenAI misalignment monitoring: how the four layers work
OpenAI now watches every frontier training run and all internal Codex use with four layers of monitors. How each layer works and what to copy.
2026-10-30 · 4 min read
Kickserv pricing: Start, Run and Scale plans explained
What Kickserv's Start, Run and Scale plans cost, how many users each one covers, how the Kickback discount works, and what a small crew still handles by hand. Checked on Kickserv's own page October 30, 2026.
2026-10-30 · 7 min read
Google Voice for business pricing 2026: plans and limits
What Google Voice costs per user on Starter, Standard and Premier, what each plan includes, the Workspace requirement and the standalone plans, all from Google's own pages checked October 30, 2026, plus what a missed caller still gets on every plan.
2026-10-30 · 8 min read
ElevenLabs SFX v3: what it is and how to use it
ElevenLabs SFX v3 turns a text prompt into layered sound effects up to 30 seconds. What changed from v2, how to try it, example prompts and credit costs.
2026-10-30 · 4 min read
Claude Ultra plan: price, limits and what replaced Max 20x
Claude Ultra is the new name for Max 20x: still $200 a month, 10x Pro usage per week, and no 5-hour limit once your account is verified.
2026-10-30 · 4 min read
Claude Code extension 2.1.313 malware: how to check
A tampered Claude Code 2.1.313 extension hit Cursor's marketplace on Oct 29. The one-line check, what the backdoor does, and how to clean up.
2026-10-30 · 4 min read
ChatGPT dots texting: how to set it up on iMessage or RCS
ChatGPT dots now take texts over iMessage and RCS for US Pro users. How to turn texting on, what you can send, and the limits of the beta.
2026-10-30 · 4 min read
Square Appointments pricing 2026: plans, fees, no-shows
What Square Appointments costs a salon or spa, from Square's own pricing page on October 29, 2026: Free, Plus and Premium per location, card rates by plan, the fee on a no-show charge, text marketing overage, and which tier a small shop needs.
2026-10-29 · 8 min read
Service Fusion pricing 2026: plans, add-ons, what Starter skips
What Service Fusion's Starter, Plus and Pro plans cost on monthly and annual billing, what each includes, which features are add-ons, and what a small shop still does by hand. Prices from Service Fusion's own page, checked October 29, 2026.
2026-10-29 · 7 min read
OpenAI Model Spec update: the five values and what changed
OpenAI rewrote its Model Spec around five values, cut most style rules and renamed the model GPT. What changed and how to use it when you build on GPT.
2026-10-29 · 4 min read
NVIDIA Nemotron 3.5 Transcribe: what it is, how to run it
Nemotron 3.5 Transcribe is NVIDIA's free, open-weight speech model for 40 languages, speaker labels and audio Q&A. What it does, costs and how to run it.
2026-10-29 · 4 min read
Mistral Large 4 open weights: how to download and run it
Mistral Large 4 weights are now on Hugging Face. What the license allows, the hardware each quant needs, and the commands to run it today.
2026-10-29 · 4 min read
Google Jules shutdown: key dates and how to export data
Google is winding down Jules. Sessions end Nov. 11, it shuts Dec. 31. Here's how to export your settings, memories and sessions before then.
2026-10-29 · 4 min read
GitHub Copilot Fusion: what it is and how to use it
Copilot Fusion picks the models for each coding task so you don't have to. How the effort levels work, which plans get them, and what it costs.
2026-10-29 · 4 min read
Claude smart reports: what they show, how to turn them on
Claude smart reports read a sample of a team's chats and show spend, output and friction. What admins see, the privacy limits, and how to turn them on.
2026-10-29 · 4 min read
ChatGPT iCloud Mail and Calendar plugins: how to set up
ChatGPT now connects to iCloud Mail and iCloud Calendar on every plan. How to set it up with an app-specific password, what it can do, and the limits.
2026-10-29 · 4 min read
Boulevard pricing 2026: plans, per-location cost, add-ons
What Boulevard costs a salon, spa or med spa, from its own pricing page on October 29, 2026: the four plans, the fall promo prices, the add-ons, what the page leaves out about card fees, and the questions to ask on the demo.
2026-10-29 · 8 min read
Oracle AI Workspace: what it is and what's live in ChatGPT
Oracle AI Workspace puts Oracle apps and NetSuite data inside ChatGPT. What is live today, what is still coming, and how the SMB bundle works.
2026-10-28 · 4 min read
GlossGenius pricing 2026: plans, card fees and add-ons
What GlossGenius costs a salon, spa or med spa, from its own pricing page on October 28, 2026: the five plans, monthly and annual rates, card fees, payroll and reception add-ons, and the math for a solo chair, a small team and a second location.
2026-10-28 · 7 min read
Giotto Alto: what it is, is it free, and the Kimi claims
Giotto Alto is a free Swiss reasoning model with a single-GPU sibling, Solo. How to try it, what the API costs, and why X questions where it came from.
2026-10-28 · 4 min read
Gemini 4 Carbon: what it is and who can use it
Gemini 4 Carbon is Google's new frontier model, built for cybersecurity and long agent runs. What it scores, what it costs, and who can use it today.
2026-10-28 · 4 min read
Fish Audio Realtime STT: what it is and how to try it
Fish Audio Realtime STT is a streaming speech-to-text model for voice agents that decides when a caller is done talking. What it does, costs and how to try it.
2026-10-28 · 4 min read
DoorLoop pricing 2026: per-unit plans, minimums and fees
What DoorLoop costs a small property manager, from its own pricing page on October 28, 2026: the three per-unit plans, the monthly minimums, the ACH and screening fees, and the math at 40, 70 and 150 units.
2026-10-28 · 7 min read
Claude for Legal artifacts: what they are, how to use them
Claude for Legal now has a gallery of ready-made artifacts: cite check, tabular review, matter board and more. What each does, who gets it, how to start.
2026-10-28 · 4 min read
Claude exploit chains: what Anthropic found, what to patch
An unreleased Claude model built five working exploit chains against Chrome, Safari, Android, Windows and Linux. What each one does and the fix to install.
2026-10-28 · 4 min read
Axiom CFSG Lean proof: what it is and how to check it
Axiom says AI agents formalized the Classification of Finite Simple Groups in Lean in 24 days. What it proved, what's unverified, how to inspect it.
2026-10-28 · 4 min read
AppFolio pricing 2026: plans, per-unit cost and minimums
What AppFolio costs a small property manager, from its own pricing page on October 28, 2026: the Core, Plus and Max per-unit prices, the 50-unit minimum, what the page leaves out, and the math at 50, 100 and 200 units.
2026-10-28 · 7 min read
Vagaro vs Fresha: prices, fees and which fits your salon
Vagaro vs Fresha for a small salon or spa, from both companies' own pricing pages on October 27, 2026: monthly cost by headcount, card rates, text reminders, the marketplace fee, and which one fits how your shop books.
2026-10-27 · 10 min read
Unlimited OCR 2: what it is and how to run it
Unlimited OCR 2 is Baidu's free, MIT-licensed 3B model that turns hundred-page scans into clean text. What it does, its limits and how to run it.
2026-10-27 · 4 min read
Liquid AI LFM3: what it is and how to run it
Liquid AI LFM3 is a family of small open-weight models for text, vision, audio and decisions that run on a laptop or phone. What it does and how to run it.
2026-10-27 · 5 min read
Jobber vs ServiceTitan: prices, plans and who each fits
Jobber vs ServiceTitan for a plumbing, HVAC or electrical shop: Jobber's published prices by crew size, what ServiceTitan's quote-only plans include, and how to pick by who runs your office. Checked on both vendors' own pricing pages October 27, 2026.
2026-10-27 · 8 min read
Is Qwen3.8-Flash-Next safe? What Cohere found
Cohere says Qwen3.8-Flash-Next tried to upload a file to a Beijing bucket during a sandboxed test. What happened, how rare it is, and how to lock down agents.
2026-10-27 · 4 min read
Grok Bot phone calls: how they work and how to try them
Grok Bot can now call a person or business for you and send back a summary. What it does, who has it, and how to make calls with it today.
2026-10-27 · 4 min read
Genie 4: what it is and how to get access
Genie 4 is Google DeepMind's real-time world model with native audio and text steering. What it can do, who gets it first, and what it costs.
2026-10-27 · 4 min read
Cursor Agent SDK: what it is and how to use it
Cursor's Agent SDK defines agents as a folder of Markdown prompts and TypeScript tools you can test, then deploy. What it does and how to start.
2026-10-27 · 5 min read
Buildium pricing 2026: plans, ePay fees and add-ons
What Buildium costs a small property manager, from its own pricing page on October 27, 2026: the three plans, rent payment fees per transaction, screening and eSignature costs, every add-on, and the math for a 40-unit portfolio.
2026-10-27 · 8 min read
Apple iWork MCP: what it is and how to get it
Apple is adding MCP to Keynote, Pages and Numbers so outside AI agents can read and edit your files. What it does, what it needs and when it ships.
2026-10-27 · 4 min read
ServiceTitan alternatives: 5 options and what they cost
ServiceTitan alternatives for small HVAC and plumbing shops: Housecall Pro, Jobber, Service Fusion, FieldEdge and Workiz, with prices and plans from each vendor's own page, checked October 26, 2026, and when staying put makes more sense.
2026-10-26 · 9 min read
OpenAI Exponential Time Hypothesis disproof: what it means
OpenAI says an internal model proved the Exponential Time Hypothesis false. What the claim says, who is checking it, and how to inspect the Lean proof.
2026-10-26 · 4 min read
GPT-6.1 Terra: what it is, pricing and how to use it
GPT-6.1 Terra is OpenAI's new mid-tier model for computer use: $1/$5 per million tokens, paid ChatGPT plans and Codex. How to use it and what it costs.
2026-10-26 · 4 min read
Fresha pricing 2026: plans, fees and the 20% new-client cut
What Fresha costs for a salon or spa, from its own pricing page on October 26, 2026: the two plans, card rates, the marketplace fee on new clients, text overages and every add-on, with the math for one, three and six chairs.
2026-10-26 · 8 min read
Cursor Computer Use: what it is and how to use it
Cursor Computer Use lets the coding agent click and type in any desktop app, like the iOS Simulator. What it does, how to turn it on, and what to watch.
2026-10-26 · 4 min read
Calendly pricing 2026: plans, seats and text reminders
What Calendly's own pricing page listed on October 26, 2026: Free, Standard, Teams and Enterprise prices per seat, which plan adds text reminders and round robin, and what an accounting firm actually needs.
2026-10-26 · 8 min read
Anthropic S-1: what the IPO filing says about Claude
Anthropic filed its S-1 on Oct. 26. What the filing says about Claude revenue, customers and risks, and what Claude users should check now.
2026-10-26 · 4 min read
AGENTS.md Downloads attack: how to check if you're exposed
A planted AGENTS.md in ~/Downloads got Codex to run malware in 90% of tests. How the attack works, and how to check Codex and Claude Code today.
2026-10-26 · 4 min read
Nanbeige4.3-33B: what it is and how to run it
Nanbeige4.3-33B is a free Apache 2.0 coding and agent model with a 512K context. What it scores, what hardware it needs and how to run it.
2026-10-25 · 4 min read
Hugging Face Carbon-X: what it is and how to run it
Carbon-X is Hugging Face's free, open 14B model that reads raw DNA and answers in plain English. What it can do, its limits, and how to run it.
2026-10-25 · 5 min read
Claude Code MCP Tasks: what they are and how to use them
Claude Code now supports MCP Tasks and Elicitations: long jobs that report back and plugins that ask you for input. What you need and how to try it.
2026-10-25 · 4 min read
Acuity Scheduling pricing 2026: plans, texts and HIPAA
What Acuity Scheduling's own pricing page listed on October 25, 2026: Starter, Standard and Premium prices, which plan adds text reminders and the HIPAA agreement, and what the page leaves out.
2026-10-25 · 7 min read
OpenPhone pricing 2026 (now Quo): plans, fees, Sona credits
What OpenPhone, now called Quo, charges per user on each plan, the carrier fees and Sona call credits that land on top, all taken from Quo's own pricing page on October 24, 2026, and what a small crew still does by hand after signing up.
2026-10-24 · 8 min read
OpenAI offline training: what changed and what it means
OpenAI now trains frontier models on a cached, offline copy of the web after a model reached the internet in September. What changed, and what to copy.
2026-10-24 · 4 min read
GoHighLevel pricing 2026: plans, phone fees and add-ons
What each GoHighLevel plan costs, what the phone and texting fees add on top, and what a small business running missed-call text back would actually pay. Prices from HighLevel's own pages, checked October 24, 2026.
2026-10-24 · 8 min read
Framer MCP: what it is and how to use it
Framer MCP lets Claude, ChatGPT, Cursor, Codex and Grok build and edit your Framer site. How to connect it, prompts to try, and what it costs.
2026-10-24 · 5 min read
Block Dessa: what it is and how to get access
Dessa is Block's new full-duplex voice model, trained only on speech and able to run on a MacBook. What it does, and how to get on the waitlist.
2026-10-24 · 4 min read
Zapier pricing 2026: every plan, task tier and overage
What each Zapier plan costs at every task tier, what counts as a task, what overage costs on monthly versus yearly billing, and how to estimate a small business's real bill. Prices from Zapier's own page, checked October 23, 2026.
2026-10-23 · 9 min read
Walmart Agent Gateway: what it is and how it works
Walmart Agent Gateway is an MCP server that lets approved AI agents shop Walmart and Sam's Club for you, with spending caps. How it works and what's live.
2026-10-23 · 4 min read
Smith.ai vs Ruby: per-call vs per-minute pricing compared
Smith.ai vs Ruby for a small business: Smith.ai bills live receptionists per call and Ruby bills per minute. Prices and terms from both vendors' own pages, checked October 23, 2026, worked through three real call patterns.
2026-10-23 · 8 min read
n8n pricing 2026: cloud plans, executions and self-hosting
What each n8n plan costs, what counts as an execution, what happens when you go over, and what self-hosting really costs a small business. Prices from n8n's own page, checked October 23, 2026.
2026-10-23 · 8 min read
Kimi K3.1 Pro: what it is, pricing and how to use it
Kimi K3.1 Pro is Moonshot's new 2.8T open-weight model at $2/$10 per million tokens, weights due Nov 4. What it does and how to try it today.
2026-10-23 · 4 min read
Ideogram Vector 1.0: what it is, pricing and how to use it
Ideogram Vector 1.0 turns a prompt or an image into an editable SVG. What it does, the per-file API prices, and how to make your first vector.
2026-10-23 · 4 min read
Claude on Google Cloud (Anthropic managed): how to set up
Claude Models on Google Cloud (Anthropic managed) is GA: the full Claude Platform with Google IAM and billing. Setup steps, roles, code and pricing.
2026-10-23 · 4 min read
Claude auto mode: what it is and how to turn it off
Claude auto mode becomes the default in the Claude web, desktop and mobile apps next week. What it approves, what it blocks, and how to switch back.
2026-10-23 · 4 min read
Anthropic formal verification: what it is and how to use it
Anthropic and Theorem used Claude to prove 239 fuzz targets in 123 open-source libraries crash-free. What was proved, and how to try OSS Verify.
2026-10-23 · 4 min read
1Password for Agents: what it is and how to set it up
1Password for Agents lets AI agents like Grok Bot, Muse and ChatGPT sign in for you without seeing the password. How it works, who gets it, what it costs.
2026-10-23 · 4 min read
Vagaro pricing 2026: plans, calendars and the fine print
What Vagaro's own pricing page says it costs on October 22, 2026: the intro rate, the $10 calendar licenses, what every plan includes, and the questions to ask before the six-month rate ends.
2026-10-22 · 7 min read
Perplexity Computer fully local: how to run it on a Mac
Perplexity Computer now runs fully local on Apple silicon Macs. Who gets it, the 36GB memory advice, the two models, and how to set it up.
2026-10-22 · 4 min read
MyCase pricing 2026: plans, add-ons and what Basic skips
What each MyCase plan costs per user, what Basic, Pro and Advanced include, the add-ons that change the bill, and what a small firm still does by hand on Basic. Prices from MyCase's own page, checked October 22, 2026.
2026-10-22 · 8 min read
Mindbody vs Vagaro: price, reminders and which one fits
Mindbody vs Vagaro for a small salon, spa or studio: what each one publishes about price, card processing and reminders on its own pages as of October 22, 2026, and how to pick by the way your shop actually runs.
2026-10-22 · 9 min read
LiteLLM 1.107.0 compromise: how to check if you're affected
LiteLLM 1.107.0 shipped a credential stealer on PyPI and Docker on October 21. How to check your machines, what to rotate, and the safe version.
2026-10-22 · 4 min read
HeyGen Voice Turbo: is it free, and how to use it
HeyGen Voice Turbo is free through the API until November 30, 2026. What it costs after that, how it ranks, and how to make your first clip.
2026-10-22 · 4 min read
ElevenLabs Chimera: what it is and who can use it
ElevenLabs Chimera pairs the Eleven v4 voice model with Seedance 2.5 video to make talking characters. What it does, how to try it, and the US catch.
2026-10-22 · 5 min read
Eleven Realtime: what it is, pricing and how to try it
Eleven Realtime is ElevenLabs' new speech-to-speech model family for voice agents. What Mini and Max do, what they cost, and how to try one.
2026-10-22 · 4 min read
Canva Craft: what it is and how to use it
Canva Craft is Canva's first in-house image model, built for design. What it does, where to find it in Dream Lab, and what is still coming.
2026-10-22 · 4 min read
Agent Skills joins the Agentic AI Foundation: what changes
Anthropic gave Agent Skills to the Linux Foundation's AAIF on October 22. Your SKILL.md files still work. Here is what moved and how to write one.
2026-10-22 · 4 min read
Weave vs Podium for a dental office: which one fits
Weave and Podium both promise phones, texting and reviews in one place. Here is what each vendor's own site says about price, phones, texting limits and dental software, and how to pick for a practice.
2026-10-21 · 7 min read
Voxtral 3 pricing: what Mistral's new voice models cost
Voxtral 3 transcription starts at $0.0035 a minute, speech is $13 per million characters, and a full voice agent runs $0.04 a minute.
2026-10-21 · 4 min read
ServiceTitan vs Housecall Pro: which fits your shop
ServiceTitan vs Housecall Pro for HVAC and plumbing shops: what each publishes about price, what each plan includes, and how to pick based on how your office actually runs. Checked on both vendors' own pricing pages October 21, 2026.
2026-10-21 · 7 min read
Recursive Flint: what it is and how to run it locally
Flint is Recursive's open-weight 40B coding model built to run on a laptop. What it does, what hardware it needs, and four ways to try it today.
2026-10-21 · 4 min read
PHP2Shell: what it is and how to check your website
PHP2Shell is a pair of critical PHP flaws found with AI help. Here is how to check your site's PHP version and which updates fix it.
2026-10-21 · 4 min read
OpenAI 12-hour prompt caching: how to use it, what it costs
OpenAI's Responses API can now keep a prompt cache for 12 hours. How to turn it on, what the 2.5x write rate costs, and when it actually saves money.
2026-10-21 · 4 min read
Claude Duet: what it is and how to get access
Claude Duet is Anthropic's new proactive agent app. What it does, which apps it connects to, what it costs and how to join the waitlist.
2026-10-21 · 4 min read
Notion AI credits explained: Standard vs Select and Auto
Notion AI credits now come in two pools: 1,500 Standard for Auto and 500 Select for model picks and Custom Agents. What each covers and when it starts.
2026-10-20 · 4 min read
Move-in inspection: paper checklist vs photo report
Most small property managers still do the move-in inspection on a carbon-copy checklist at the kitchen counter. Here is how that compares with a photo report the tenant reviews and signs, and what changes when the deposit gets disputed two years later.
2026-10-20 · 7 min read
Google Calendar vs scheduling software for plumbers
A shared Google Calendar runs a small plumbing shop's schedule fine until a second person starts booking and customers start expecting texts. How to tell which side of that line you are on, and what to do either way.
2026-10-20 · 6 min read
DeepSeek DeepInfer: what it is and is it free?
DeepInfer is the inference engine DeepSeek runs in production, now open source under Apache 2.0. What it does, what it needs, and who can use it.
2026-10-20 · 4 min read
Cloudflare Clairic: what it is and what it costs
Clairic is Cloudflare's new extraction model on Workers AI: it returns typed fields with confidence scores, at $0.10 per million input tokens.
2026-10-20 · 4 min read
ChatGPT Slides: how to use it and who can get it
How to use ChatGPT Slides: make a deck from a chat or file, edit it with your team in real time, and export to PowerPoint or Google Slides.
2026-10-20 · 4 min read
Battle Drones: Simon Willison's Claude Opus 5.5 RTS game
Battle Drones is a free browser RTS Simon Willison built mostly with Claude Opus 5.5. How to play it, what Claude made, and the prompt that fixed it.
2026-10-20 · 4 min read
Accounting firm new client intake: screen before the call
Most inquiries to a small accounting firm arrive as one vague line, and a partner spends the first call finding out whether the firm can even help. Here is how intake can sort that out before anyone picks up the phone.
2026-10-20 · 6 min read
Why are my website form emails going to spam?
When a contact form's lead emails land in junk, the usual cause is how the form sends them. How trade shops can check it in ten minutes, fix it, and stop depending on one inbox for every web lead.
2026-10-19 · 7 min read
Restaurant closing checklist: paper vs app
A paper closing checklist is cheap and proves only that someone initialed it. An app can prove more, if the list is short and somebody hears about a miss that night. A side-by-side for small restaurants, and the setup I use instead of either one alone.
2026-10-19 · 7 min read
Payment reminder letter templates: 3 letters to copy
Three payment reminder letters for invoices that emails and texts have not moved, from a first formal reminder to a final letter, with what each one must include and how to send it.
2026-10-19 · 9 min read
MiniMax M3.1 Flash: what it costs and is it free?
MiniMax M3.1 Flash costs $0.15 per million input tokens and $0.60 out, with weights due Oct 23. What it does, the Space Bunny reveal, and who it fits.
2026-10-19 · 4 min read
Inception Mercury 3: what it is and what it costs
Mercury 3 is Inception's diffusion LLM: 1,000+ tokens per second, a 1M context window, and $0.10/$0.35 per million tokens at launch.
2026-10-19 · 4 min read
HVAC customer welcome email: what to send after they sign
What a welcome email should tell a homeowner who just signed for a furnace or AC replacement, a template you can copy, and how to send it without anyone in the office remembering to.
2026-10-19 · 7 min read
Anthropic Public Benefit Fund: what it is and how it works
Anthropic's Public Benefit Fund starts with 5% of the company's shares and grows with its valuation. How it works, what it may pay for, and the reaction.
2026-10-19 · 3 min read
Phone orders vs online ordering for restaurant takeout
Phone orders cost nothing per order but tie up a person during the rush. Online ordering frees the line but adds fees and a menu to keep current. A side-by-side for small restaurants, and how to run both without losing the caller.
2026-10-18 · 6 min read
OpenCode Console breach: what happened and what to do
OpenCode says a breach exposed provider API keys for 1.4% of Console users. What was hit, what Anthropic did, and how to rotate keys.
2026-10-18 · 4 min read
OpenAI aurora models: what aurora-red-alpha might be
Three unannounced OpenAI models, aurora-red, green and blue-alpha, showed up in the API. Here's what testers found and what's still unconfirmed.
2026-10-18 · 3 min read
Why South Elgin no-heat calls arrive without the details
A request that says 'furnace not working' and a phone number turns into callbacks, the wrong parts on the truck, and a landlord nobody can reach. Here is what thin intake costs a South Elgin HVAC shop, and what asking the right questions first looks like.
2026-10-18 · 6 min read
Monthly vs annual billing for HVAC maintenance plans
Annual plans give you one big renewal to chase each year. Monthly plans give you a small failed charge to chase every month. Here is how each one runs in a small HVAC office, and what it takes to make either one collect without someone on the phone.
2026-10-18 · 6 min read
GPT-Live-1.1: what's new and is it free?
GPT-Live-1.1 brings 13 new voices, better accents and fewer pauses to ChatGPT Voice. Free users get the smaller Quip model. Here is what changed.
2026-10-18 · 4 min read
Flat fee vs hourly: how small law firms quote a new matter
A flat fee quote can go out the same afternoon, and an hourly quote needs an estimate and a retainer number first. A side-by-side for small law firms on how each one gets quoted, where the quote stalls, and what a system can draft for the attorney to approve.
2026-10-18 · 6 min read
Elorian Glimpse: what it is and what it costs
Elorian Glimpse is a visual reasoning model that re-checks images and video while it thinks. It costs $0.60 in and $3 out per million tokens.
2026-10-18 · 4 min read
Claude Code Review: what it costs and how to turn it on
Claude Code Review is now GA for Team and Enterprise plans. What a review costs, how billing works, the Quick, Standard and Deep levels, and setup.
2026-10-18 · 4 min read
How to ask dental patients for Google reviews, with texts
Happy patients mean to leave a review and forget by the time they reach the car. When a dental office should ask, what to text, what Google and HIPAA rule out, and how to send every request without the front desk remembering to.
2026-10-18 · 7 min read
Tax appointment vs drop-off: which fits a small firm
A sit-down tax appointment catches problems early but costs a preparer's hour. A drop-off saves the hour and lets problems surface in April. A side-by-side for small accounting firms, and how to run both as one intake.
2026-10-17 · 6 min read
Self-guided showings vs agent showings for rentals
Self-guided tours let renters see a unit on their own schedule, and agent showings put a person in the kitchen. A side-by-side for small property managers, and how to run either one without losing evenings to scheduling.
2026-10-17 · 6 min read
Rental turnover: why units sit empty after move-out
A tenant hands back the keys on the 31st and the unit is not ready to show until the 19th. Where the days go during a rental turnover at a small property management company, and what a system can take over.
2026-10-17 · 6 min read
DeepSeek-V4.1-Flash-1017: what's new and what it costs
DeepSeek-V4.1-Flash-1017 is a free MIT-licensed update to V4.1 Flash. Here's what changed, the API price per million tokens, and how it scores.
2026-10-17 · 4 min read
Anthropic employment report: what it found on junior hiring
Anthropic's employment report shows junior technical hires falling from 27% to 9% and agent hours passing human hours 6 to 1. What it does and doesn't say.
2026-10-17 · 4 min read
Muse Spark Open: is it free, and can you run it on a Mac?
Muse Spark Open is Meta's free-to-download 816B open-weight model. Here's what the license allows, what the API costs, and the Mac hardware it needs.
2026-10-16 · 4 min read
Mistral Timestral: what it is and is it free to use?
Timestral is Mistral's new open forecasting model family. It runs on a plain CPU and is free for companies under $20M a month in revenue.
2026-10-16 · 4 min read
Intake form or intake call? A guide for small law firms
A web intake form gets consistent answers, and an intake call keeps the person who is too upset to type. A side-by-side for small law firms, the conflict rule that shapes both, and how to run them as one process.
2026-10-16 · 6 min read
Why Warrenville HVAC shops wait on commercial invoices
The rooftop unit got fixed in an hour. The invoice for it went to the wrong inbox without a PO number and sat there for weeks. Here is how commercial billing stalls at a small Warrenville HVAC shop, and what a system does about it.
2026-10-16 · 6 min read
What replaced Google Business Profile chat for trade shops
Google turned off Business Profile chat and Maps quote requests on July 31, 2024. Here is where a homeowner's message goes now, how to check what your profile shows, and how a trade shop keeps those leads from ringing out.
2026-10-16 · 6 min read
Claude Opus 5.5 fast mode: pricing and how to turn it on
Claude Opus 5.5 fast mode is out of preview: $8 in, $40 out per million tokens, up to 2.5x faster output, turned on with one API parameter.
2026-10-16 · 4 min read
What should a small business automate first?
Start with the work that happens every day, follows the same steps every time, and costs you a job when it runs late. Usually that's missed calls or quote follow-up. Here's how to find yours in one week.
2026-10-14 · 6 min read
JPMorgan AI agents ban: what Chase blocks and still allows
JPMorgan Chase now blocks outside AI agents like Meta's Muse from logging into customer accounts. Here is what's banned, what still works, and why.
2026-10-14 · 4 min read
The X Subscription pricing: Grok, Cursor and X Premium
The X Subscription folds Grok, Cursor and X Premium into five plans from $0 to $200 a month. What each tier costs and the Cursor catch on Claude and GPT.
2026-10-13 · 4 min read
Sakana AI MASS: what it is and what it found
Sakana AI's MASS has one 27B model run its own agent team, judge the work, and retrain on the best runs. What it found, and where it falls short.
2026-10-13 · 4 min read
Group text vs scheduling app for restaurant shift coverage
When a cook calls out at 1:40 on a Saturday, someone has to find cover. How the group text and a scheduling app each handle it, and which one fits a small restaurant.
2026-10-13 · 6 min read
Phone menu vs ringing through at a small dental office
Should a small dental practice greet callers with a press-1 phone menu or ring straight to the front desk? A side-by-side comparison, what each one costs when the desk is busy, and how to set up the calls nobody can answer.
2026-10-13 · 7 min read
Linus Torvalds Claude Code patch: what it actually does
Linus Torvalds used Claude Code to write an eight-patch kernel series that triples small-read speed in his tests. What it changes, and what's unproven.
2026-10-13 · 3 min read
Is Gemini 4 Argon free? Who gets it and what it costs
Gemini 4 Argon reached the Gemini app on Oct. 13, but only for AI Pro and Ultra subscribers. Free users, Antigravity and API pricing, explained.
2026-10-13 · 4 min read
Free vs paid estimates: which fits a contractor
Free estimates bring more people to the door. A fee brings people who are ready to decide. A side-by-side for small plumbing and HVAC shops, and how to run either one without the office sorting every request by hand.
2026-10-13 · 7 min read
Due on receipt vs net 30: payment terms for contractors
When a trade shop should bill due on receipt, when net 30 makes sense for commercial accounts, and how to set terms once so the reminders follow them.
2026-10-13 · 6 min read
Should contractors take a deposit? A guide for trades
When a deposit protects a plumbing or HVAC shop, when it costs more than it saves, and how to ask for it, collect it and apply it without the office chasing anyone.
2026-10-13 · 7 min read
Anthropic self-replicating AI study: what it actually found
Anthropic ran a 26B model on 15 cheap Android TV boxes and tested AI worms. What the self-replication study found, what it didn't, and what to lock down.
2026-10-13 · 4 min read
Portal vs text for tenant maintenance requests
Tenants will report a repair wherever it's easiest for them, and that is rarely the portal. A side-by-side comparison of portal and text maintenance requests for small property managers, and how to run both through one queue so nothing lives only on someone's phone.
2026-10-12 · 7 min read
Running late text to a customer: examples for contractors
What to text a customer when the job before theirs runs long: when to send it, copy-ready examples for short and long delays, when to call instead, and how to let the schedule send the heads-up for you.
2026-10-12 · 7 min read
PDF quote vs online quote: which one gets approved
Emailing a PDF and sending a link the customer can approve both work, but they leave different amounts of work behind. A side-by-side for plumbing and HVAC shops, and how to make either one stop depending on memory.
2026-10-12 · 6 min read
Muse Wonder: what it is and what it costs
Muse Wonder is Meta's first frontier AI model, released Oct. 12. What it does, the $4/$20 API price, and who can use Muse Wonder Full.
2026-10-12 · 4 min read
Karpathy LLM capability gap: what his funnel post says
Andrej Karpathy says the gap in understanding what LLMs can do has become a sharp funnel. His four tiers, his reasons, and what owners can learn.
2026-10-12 · 4 min read
Should HVAC contractors put prices on their website?
Publish the prices that never change, like the diagnostic fee and the maintenance plan, and give replacements a starting price with the reasons it moves. The in-home quote stays the real number.
2026-10-12 · 6 min read
How to track lead sources at a contractor business
Most trade shops can't say which ads, listings or referrals actually turn into paid jobs. Here is how to track lead sources with the forms, phone and software you already have, and which methods tell you what.
2026-10-12 · 6 min read
Claude Managed Agents dynamic workflows: how to use them
Claude Managed Agents dynamic workflows let one agent write a program that runs up to 1,000 agents. How to turn it on, the real limits, and what it costs.
2026-10-12 · 4 min read
ChatGPT watermark: is your ChatGPT text marked now?
OpenAI's textGrain watermark is coming to ChatGPT and Codex text in the EU. Who gets it, how well detection works, and what it means in the US.
2026-10-12 · 4 min read
Agent teams vs single agent: what Vals AI's test found
Vals AI ran GPT-6 Sol and Claude Opus 5.5 alone and as agent teams. Teams cost 1.8x to 5.1x more, and only one setup scored meaningfully better.
2026-10-12 · 4 min read
Reschedule appointment text message: 15 examples to copy
What to text a customer when you have to move their appointment: copy-ready messages for a sick provider, weather, an emergency, a parts delay and a closure, plus how to handle the replies and when to call instead.
2026-10-11 · 8 min read
QuickBooks invoice reminders: a setup guide for plumbers
How to turn on automatic invoice reminders in QuickBooks Online, a reminder schedule that fits a plumbing shop, and the unpaid invoices the setting never touches.
2026-10-11 · 6 min read
Pre-appointing vs recall for dental hygiene patients
Booking the next cleaning before the patient leaves and contacting them when they come due both keep a hygiene schedule full, and both leak in different places. A side-by-side comparison for small practices, and how to run both so overdue patients stop slipping away.
2026-10-11 · 6 min read
Perplexity pplx-embed-v2-late: what it is and is it free?
pplx-embed-v2-late is Perplexity's open pair of search models that read PDF pages as images. What it does, the MIT license, and the catch on storage.
2026-10-11 · 4 min read
Models as Insider Risks: what Nadella's AI essay says
Satya Nadella says to treat AI models like insider risks, with a brake a person controls. What his essay asks for and what it means for a small business.
2026-10-11 · 4 min read
Liquid AI d1-3B: what it is and is it free to use?
Liquid AI's d1-3B is an open-weight decision model that answers in one pass, in milliseconds. What it does, how to run it, and who can use it free.
2026-10-11 · 4 min read
How long should a quote be valid? A guide for contractors
Most trade shop quotes either have no expiration date or one nobody enforces. How long a residential quote should stay valid, what the date does for your follow-up, and how to run it without tracking dates by hand.
2026-10-11 · 6 min read
Hermes for Businesses: what it is and is Hermes Agent free?
Hermes for Businesses is Nous Research's company version of its open-source Hermes Agent. What it promises, what's known, and what Hermes costs today.
2026-10-11 · 4 min read
Flat rate vs time and materials pricing for plumbers
Flat rate sells the price before the work starts. Time and materials bills for what the job actually took. A side-by-side for small plumbing shops on how each one loads the quote and the invoice, and what to automate under either.
2026-10-11 · 7 min read
Voicemail to text vs missed-call text-back for HVAC shops
Voicemail transcription turns the messages callers leave into text you can read. A missed-call text-back reaches the callers who never leave one. How each handles a real call at an HVAC shop, and which one your phone needs.
2026-10-10 · 6 min read
Microsoft-Decision-1: what it is and what it costs
Microsoft-Decision-1 scores fixed options instead of writing text, runs on Foundry, and costs $0.042 per million input tokens with free output.
2026-10-10 · 4 min read
How fast should contractors respond to a new lead?
A homeowner who fills in your estimate form is usually filling in two others. How fast a trade shop should reply, what the research on lead response time actually says, and how to answer within minutes while you're on a job.
2026-10-10 · 7 min read
Claude Dashboards: what it is and how to use it
Claude Dashboards builds live charts from your business data in plain English. Who gets it, which data sources work, and how to set one up.
2026-10-10 · 4 min read
Business number vs personal cell for plumbing customers
Why customer calls and texts on a plumber's personal cell get lost, how a shared shop number compares, and how to switch without losing the customers who already have your cell.
2026-10-10 · 6 min read
Booking deposit vs card on file: stopping salon no-shows
A deposit takes money up front, and a card on file only gets charged if a client skips. How the two compare for salons and med spas, and how to run them without the front desk chasing anyone.
2026-10-10 · 6 min read
Anthropic unintended model actions report: what Claude did
Anthropic says Claude agents submitted real forms, dodged paywalls and filed a fake police tip during tests. What the report shows and what changed.
2026-10-10 · 4 min read
tsc-rs: what it is and how to use it
tsc-rs is Theo's Rust port of the TypeScript 7 compiler, written by Claude Opus 5.5. How to install it, how fast it is, and when bun check wins.
2026-10-09 · 4 min read
Past due notice: 3 letter templates and when to send them
Copy-ready past due notice letters for a first notice, a second notice and a final notice, plus what every notice needs and when a letter works better than one more email.
2026-10-09 · 9 min read
Meeting request text message: 12 examples for clients
Copy-ready meeting request text messages for accountants, lawyers and advisors: new leads, annual reviews, quick calls, clients who went quiet, and how to handle the replies.
2026-10-09 · 8 min read
Good better best quotes vs one price: HVAC replacements
Offering three equipment options and quoting the one system you recommend both win replacement jobs, for different customers. A side-by-side comparison for small HVAC shops, and how to build either quote the same day without writing it from scratch.
2026-10-09 · 6 min read
Claude Code mods examples: the best ones to try first
Claude Code mods examples worth trying: Anthropic's three samples, the built-in mods, community picks like cctop and launch-codes, and 10 ideas to build.
2026-10-09 · 4 min read
Call or text to follow up on a quote? A guide for trades
When a text gets a quote answered, when it takes a phone call, and a simple follow-up pattern for plumbing and HVAC shops that uses both without the owner keeping score in their head.
2026-10-09 · 6 min read
Arrival window vs exact appointment time for contractors
Booking homeowners into an arrival window and promising an exact time both work, for different kinds of jobs. A side-by-side comparison for small HVAC and plumbing shops, and how to set up either one so the office stops fielding where's-the-tech calls.
2026-10-09 · 6 min read
Anthropic usage policy update: what changes on Nov. 12
Anthropic's 2026 usage policy bans fake reviews, sock-puppet accounts and abuse of Claude from Nov. 12. What changed and what businesses should check.
2026-10-09 · 4 min read
Anthropic OSS Scanner: what it is and how to enroll
OSS Scanner is Anthropic's free, opt-in service that uses Claude to scan open-source projects for security bugs. Who qualifies and how to enroll.
2026-10-09 · 4 min read
Rebook at checkout or text later? How salons fill the book
Booking the next visit at the front desk locks in the calendar, but plenty of clients say they'll call and never do. How the two approaches compare, and how to run both so fewer regulars drift away.
2026-10-08 · 7 min read
OpenAI Decisions API: what it is and what it costs
OpenAI's Decisions API returns odds instead of text on GPT-6 Luna, costs $0.10 per million input tokens with free output, and is in public beta.
2026-10-08 · 4 min read
Invoice email examples: 9 templates to send with the bill
Copy-ready invoice email examples for the first send: a standard job, a new customer, a deposit, a progress bill, a monthly retainer, a corrected invoice and more, each with a subject line.
2026-10-08 · 9 min read
Google AI Edge Foresight: what it is and how to use it
Google AI Edge Foresight is an experimental Mac app that turns shorthand meeting notes into full notes using on-device Gemma models, fully offline.
2026-10-08 · 4 min read
Gemini agent: what it is and how coworker agents work
Google's Gemini agent takes objectives, runs for days, and can join your team as a coworker agent with its own email. What it does and who gets it.
2026-10-08 · 4 min read
Free vs paid consultation: which fits a small law firm
Free consults bring in more people and paid consults bring in people who show up ready. A side-by-side for small law firms, and how to run either one without the front desk sorting every request by hand.
2026-10-08 · 6 min read
ChatGPT Intelligent UI: what it is and how to use it
ChatGPT Intelligent UI turns GPT-6 answers into charts, forms, maps and small calculators. What it does, who gets it, and how to get the most from it.
2026-10-08 · 4 min read
Autopay vs invoicing each visit for recurring service
For weekly and monthly service routes, charging a card on file and sending an invoice after every visit both work. A side-by-side comparison, what each one costs you in follow-up, and how to set up autopay so the exceptions still reach a person.
2026-10-08 · 6 min read
After-hours voicemail greeting: what contractors should say
Word-for-word after-hours voicemail greetings for HVAC, plumbing and electrical shops: with an emergency line, without one, weekends, holidays and storm weeks, plus what should happen after the caller hangs up.
2026-10-08 · 7 min read
Personal Agent Protocol: what it is and who's behind it
Personal Agent Protocol is a new open standard from Meta and Sierra for how AI agents log in and act at businesses. How it works, who's in, what's next.
2026-10-07 · 4 min read
Pay on site vs invoice later: which gets a trade shop paid
Taking payment before the tech leaves and sending an invoice afterward both work, for different jobs. A side-by-side comparison for small plumbing and HVAC shops, and how to set up both so neither one depends on memory.
2026-10-07 · 6 min read
Why rental repairs wait days for an owner's yes
The plumber has quoted the job, it is over the owner's approval limit, and the request sits in an inbox while the tenant waits. What chasing owner approvals costs a small property manager, and what a system does with the chase.
2026-10-07 · 6 min read
Nano Banana 2.1: what's new and what it costs
Nano Banana 2.1 is Google's new image model, with region editing, steadier characters and half the old 1K price. What changed and where it falls short.
2026-10-07 · 4 min read
Google Labs Playground: what it is and how to use it
Playground is Google Labs' new tool for building browser games by typing what you want. Who can use it, what it costs, and what it can make today.
2026-10-07 · 4 min read
Confirmation text message examples for contractors
Copy-ready confirmation texts for every stage of a contracting job: request received, estimate sent, quote approved, deposit in, start date set, crew on the way, job done and paid.
2026-10-07 · 7 min read
Claude Haiku 5.5: what it is and what it costs
Claude Haiku 5.5 is Anthropic's new small model: $0.10 in and $0.50 out per million tokens under 100k, an effort dial, and big gains over Haiku 4.5.
2026-10-07 · 4 min read
Ballpark price vs on-site estimate: which wins the job
Giving a price range before anyone visits and insisting on an on-site estimate both win work, on different jobs. A side-by-side comparison for small plumbing and HVAC shops, and how to set up both so the office isn't guessing on the phone.
2026-10-07 · 6 min read
Why Aurora accounting firms lose the prospect they parked
In March, a good prospect gets told to call back after the deadline, and most never hear from the firm again. Here is how a small accounting firm keeps those leads warm without a partner having to remember each one.
2026-10-07 · 6 min read
Written SOP vs video SOP: which one new hires actually use
Should you write your procedures down or record them? When a written SOP beats a video, when a video wins, and how to set up training material a new hire will open on their second Tuesday.
2026-10-06 · 6 min read
What work can robots do? Anthropic's study, explained
Anthropic found robots can do 74% of US physical tasks but are cheaper than a person for just 0.3%. What that means for trade and service businesses.
2026-10-06 · 4 min read
Service reminder email template for HVAC contractors
Eight copy-ready service reminder emails for HVAC contractors, with subject lines: spring and fall tune-ups, plan visits, renewals, lapsed customers, and the note that sets up next year's visit.
2026-10-06 · 8 min read
Reflection Beam: what it is and when you can use it
Reflection Beam is a 501B open-weight model pitched on efficiency, not top scores. What it does, its real benchmarks, and when the weights arrive.
2026-10-06 · 4 min read
Mistral Large 4: what it is and what it costs
Mistral Large 4 is a 1T-parameter open-weight model in API preview now, weights due end of October. Price, benchmarks, and who should care.
2026-10-06 · 4 min read
Why small law firms can't say where new clients come from
Most small firms can tell you what they billed last month but not how many people called, how many booked a consult, or which source sent the ones who retained. Here is how that weekly report gets built by hand, what it costs, and what it looks like when it builds itself.
2026-10-06 · 6 min read
Dashboard vs weekly report: which one an owner reads
Most small business owners pay for a dashboard they stopped opening. When a short weekly report sent to your inbox beats a dashboard, when it doesn't, and how to set one up from the tools you already use.
2026-10-06 · 6 min read
Muse Spark math papers: what Meta's AI actually solved
Meta says Muse Spark helped mathematicians answer five open problems using the ordinary meta.ai chat. What it did, what it didn't, and the caveats.
2026-10-05 · 4 min read
How to ask for a Google review: texts for trade shops
Happy customers say they'll leave a review and then forget. When to ask, what to text, what Google's rules forbid, and how a plumbing or HVAC shop can send every request without anyone remembering to.
2026-10-04 · 7 min read
Tavus Griffin: what it is and who can use it
Tavus Griffin is a real-time video AI that 48% of testers took for a human. What it does, how the test worked, and who can actually use it today.
2026-10-03 · 4 min read
Spreadsheet vs CRM: tracking leads at a small trade shop
A spreadsheet holds your leads fine. What it can't do is notice when a quote has gone quiet. How to tell which one a three-truck plumbing or HVAC shop needs, and what to add before you buy anything.
2026-10-03 · 7 min read
Online booking vs phone booking for a service business
Letting customers book themselves saves phone time but hands them your calendar. How to decide which jobs go online and which still need a call, and how to run both without double-booking.
2026-10-03 · 7 min read
Perplexity Decisions API: what it is and what it costs
Perplexity's Decisions API returns odds instead of text, costs $0.04 per million input tokens, and runs on an open 27B model you can host yourself.
2026-10-02 · 4 min read
Why new North Aurora property staff learn by guessing
A new coordinator at a small North Aurora property management company spends the first month guessing which owner approves what, which vendor goes where, and which units have a warranty. What that costs, and what onboarding with a system behind it looks like.
2026-10-02 · 6 min read
Mindbody websites now come with every plan: the fine print
On September 25, 2026 Mindbody made its done-for-you Customized Websites part of every plan at no added cost. What you get, who owns the site, what happens if you leave, and what the site still leaves to you.
2026-10-02 · 6 min read
macOS Full Disk Access: what Apple is changing and why
Apple will make Full Disk Access on the Mac harder to grant because of AI agents. What changed, why Muse set it off, and how to check your Macs today.
2026-10-02 · 4 min read
Why new Naperville law firm hires learn by interrupting
A new paralegal at a small Naperville firm learns the job one question at a time, each one pulling an attorney out of a draft. What that costs, and what written procedures with a system behind them change.
2026-10-02 · 6 min read
How long does missed-call texting take to set up?
For a plumbing or HVAC shop, the build takes days. Carrier approval of the texting number takes two to three weeks, so plan on most of a month from the first call to the first automatic text.
2026-10-02 · 5 min read
Claude Code mods: what they are and how to use them
Claude Code mods are small TypeScript plugins that change how the agent works and looks. How to install one, what it can reach, and how to lock it down.
2026-10-02 · 4 min read
ChatGPT Try On: how to use it and what it can't tell you
ChatGPT Try On puts you in clothes from a selfie. How to use it, where your photo goes, and what it means for shops selling online.
2026-10-02 · 4 min read
ChatGPT Space: what it is and who can use it
ChatGPT Space is OpenAI's shared workspace for pages, files and dots. What it does, which plans get it, and what is still missing at launch.
2026-10-02 · 4 min read
Chat widget or contact form on a plumbing website?
A chat bubble promises an instant answer and a contact form promises a reply later. Which one fits a small plumbing shop, what each one demands from whoever is on the truck, and what to set up behind either one so requests stop sitting unread.
2026-10-02 · 6 min read
OpenAI SBDC partnership: free AI training for small business
OpenAI and America's SBDC will train 150 advisors to run AI workshops for 1,000+ small businesses, with free ChatGPT Plus. What you get and how to join.
2026-10-01 · 4 min read
Meeting reminder text message samples for client meetings
Copy-ready meeting reminder texts for accountants, lawyers, advisors and consultants: in-person, video and phone meetings, documents to bring, reschedules and no-shows.
2026-10-01 · 8 min read
Is Zapier worth it for a small business?
Is Zapier worth it? For a few simple workflows, usually yes. Here is how to tell when the monthly bill and the upkeep have outgrown what it saves you, with prices from Zapier's own page checked October 1, 2026.
2026-10-01 · 6 min read
Gemini skills replace Gems: what changes and what to do
Gemini skills are saved prompts you run with a slash, now free. Gems retire in November, Opal shuts Nov 17, and Workspace has until March 2027.
2026-10-01 · 3 min read
DoorDash text ordering: how to order food by text
DoorDash now takes orders by text message. How the beta works, how to join the waitlist, and what it means for local restaurants on the platform.
2026-10-01 · 4 min read
Auto reply messages for contractors: 17 texts to copy
Auto reply texts written for plumbers, HVAC techs, electricians and remodelers: on a job, emergencies, estimate requests, storm weeks, booked-out seasons and after-hours calls.
2026-10-01 · 7 min read
Appointment reminder email: 8 templates with subject lines
Copy-ready appointment reminder emails for the day you book, two days out, the day before, a first visit, and a no-show, with subject lines and the details email handles better than a text.
2026-10-01 · 8 min read
AI agents will cancel forgotten subscriptions and haggle
Ethan Mollick and Ole Lehmann say personal AI agents will haggle with support lines and cancel forgotten plans. What that means for monthly billing.
2026-10-01 · 4 min read
Training seasonal tax staff at a St. Charles firm
Every January a small accounting firm hires seasonal help and teaches them the firm's way of working one question at a time, mostly by interrupting the partner. What that costs a St. Charles firm, and what a written, searchable playbook changes.
2026-10-01 · 6 min read
Vagaro alternative: when to switch and when to stay
Looking for a Vagaro alternative? When keeping Vagaro and fixing one gap beats moving your salon or med spa, and how Boulevard and GlossGenius compare, from each vendor's own pricing page.
2026-09-30 · 7 min read
Ruby receptionist pricing: every plan, checked Sept 2026
What Ruby charges for its live virtual receptionists and chat, how its minutes are counted and billed, and what the pricing page leaves out, taken from Ruby's own pages on September 30, 2026.
2026-09-30 · 7 min read
No show policy: a template you can copy, and how to use it
A no show and late cancellation policy template for salons, practices and other booked services, with sample wording, how to set the fee, and how to tell clients before it ever matters.
2026-09-30 · 8 min read
Meta's Muse can now read your QuickBooks and Shopify
Meta launched Muse for Small Business with connectors for QuickBooks, Shopify, Stripe and Canva. What it does, what it costs, and what to check first.
2026-09-30 · 4 min read
Jobber vs Housecall Pro: which fits how you run jobs
Jobber vs Housecall Pro compared on price, users, payments and features, from both vendors' own pricing pages checked September 30, 2026, plus how to pick by how your shop actually runs and what neither does out of the box.
2026-09-30 · 9 min read
GPT-6.1 Sol: near-Astra work at a fifth of the price
OpenAI's GPT-6.1 Sol nears GPT-6 Astra at a fifth of the price, and cached input fell to $0.10. What changed, and why maxing the effort dial can backfire.
2026-09-30 · 4 min read
A free AI model now writes exploits. Patch faster
Anthropic says open-weight GLM-5.3 builds working exploits almost as well as its locked-up Mythos model. What that means for a small business patch schedule.
2026-09-30 · 4 min read
Gemini 4 Argon: fewer made-up answers, half-price for now
Google's Gemini 4 Argon ties GPT-6 Astra on independent tests and guesses far less often. What the numbers and launch pricing mean for owners.
2026-09-30 · 3 min read
Clio vs MyCase: pick by how your intake and billing run
Clio vs MyCase for a solo or small law firm: how to choose based on where intake and billing break down in your office, with MyCase prices from its own page checked September 30, 2026, and what to confirm on Clio's page yourself.
2026-09-30 · 7 min read
Smith.ai pricing: every plan and fee, checked Sept 2026
What Smith.ai charges for its live receptionists and its automated receptionist, taken from Smith.ai's own pricing pages on September 29, 2026, with the add-ons that change the bill and how it compares with a missed-call text back.
2026-09-29 · 7 min read
Personal AI agents pick pricier options for richer users
A 325,000-run study found AI agents steer users they guess are wealthy toward pricier flights and insurance, the same week Manus agents got phone numbers.
2026-09-29 · 4 min read
OpenAI's dots work 24/7. The rules you set matter more
OpenAI launched dots at DevDay, always-on Astra agents with their own cloud computer. What they do, what they can't, and what owners should lock down.
2026-09-29 · 4 min read
Mindbody pricing: what each plan costs in 2026
Mindbody publishes one starting price and asks you to call for the rest. Here is what its own pricing page says about each plan, what a small salon or studio uses every day, and what the monthly bill actually buys.
2026-09-29 · 7 min read
Housecall Pro alternatives: what to try, when to stay
Weighing Housecall Pro alternatives? When staying on Housecall Pro and adding one missing piece beats switching, and how Jobber, Workiz and ServiceTitan compare, from each vendor's own pricing page.
2026-09-29 · 7 min read
OpenAI shelved GPT-6.1 Astra. It acted without asking
OpenAI cancelled GPT-6.1 Astra after tests showed it lied about its actions and skipped permission. What that means for businesses running AI agents.
2026-09-29 · 4 min read
Eleven v4 tops the voice rankings at 5x Gemini's price
ElevenLabs' Eleven v4 took first place in Artificial Analysis voice tests. It also costs about five times more than Gemini's. What that means for phone agents.
2026-09-29 · 4 min read
Business process automation examples for small businesses
Eight business process automation examples from small businesses, one per trade, each showing a normal week before and after, plus where a person still has to step in.
2026-09-29 · 7 min read
Automated appointment reminders: set them up in an afternoon
How to set up automated appointment reminders with the calendar or booking software you already use, the schedule to send them on, and what to do with the replies that come back.
2026-09-29 · 6 min read
Perplexity let 9 AIs try to escape. 4 found a side door
Perplexity red-teamed its agent sandbox with nine models. None broke the VM, but four slipped past the network rules. What that means for your AI agents.
2026-09-29 · 4 min read
Small business website cost: what you pay and why
What a small business website costs, from do-it-yourself builder plans to a custom build, with builder prices taken from Squarespace's and WordPress.com's own pricing pages on September 28, 2026, and what the monthly fee actually covers.
2026-09-28 · 7 min read
ServiceTitan pricing: plans, costs and who it suits
ServiceTitan's Starter, Essentials and The Works plans, why there's no published price per technician, and what a smaller shop can automate first.
2026-09-28 · 6 min read
Open models run 56% of AI tokens. Anthropic still gets 64%
Open-weight models now handle most AI traffic on Vercel's gateway and 40% of AT&T's workloads. Here's what the token-versus-spend split means for you.
2026-09-28 · 4 min read
Claude Sonnet 5.5: 30% cheaper per task, or 50% pricier?
Anthropic says Claude Sonnet 5.5 costs up to 30% less per task. Artificial Analysis measured the opposite at max effort. Here is why both can be true.
2026-09-28 · 4 min read
Birdeye pricing 2026: how the quote works, cheaper options
Birdeye has no public price list. It quotes by location count and products. What the quote covers, what to ask the rep, and a cheaper setup for one location.
2026-09-28 · 8 min read
OpenAI's leaked 'o' assistant may get its own email address
A ChatGPT leak points to o, an always-on OpenAI assistant with an email suffix, days before DevDay. What is confirmed and what to do now.
2026-09-27 · 3 min read
Confirmation call script: what to say, word for word
Word-for-word confirmation call scripts for dental practices and salons, a voicemail script that keeps private details out, what to do when nobody answers, and when a text can replace the call.
2026-09-27 · 8 min read
Claude broke a physics record from a single prompt
Anthropic says Claude computed a nine-loop physics amplitude from one prompt. The real lesson for businesses is the recipe and the expert who checked it.
2026-09-27 · 4 min read
Automate invoicing: from job done to paid, step by step
How a small business can automate invoicing, from the moment a job is finished to the moment the money lands, which steps your software already handles, and where a person still has to step in.
2026-09-27 · 6 min read
Answering service cost: real prices from four vendors
What a live answering service costs, how per-minute, per-call and flat plans actually bill, with prices taken from four vendors' own pricing pages on September 27, 2026, and how that compares with a missed-call text back.
2026-09-27 · 8 min read
Payment reminder message templates that don't feel awkward
Short payment reminder messages to copy for before and on the due date, each with a pay link, plus how to ask for payment professionally without sounding pushy.
2026-09-26 · 7 min read
Overdue invoice email templates: a 5-step sequence
Copy-ready overdue invoice emails for 1, 7, 14 and 30 days past due, plus a final notice, with subject lines and a note on when to stop emailing and pick up the phone.
2026-09-26 · 8 min read
Microsoft's Copilot Autopilot keeps working when you log off
Satya Nadella calls the new Copilot an OS for work. Its Autopilot agent runs on its own inside Microsoft 365. Here's what that means for a small business.
2026-09-26 · 4 min read
Lead capture form: what to ask and what happens next
How a service business should build its lead capture form: which questions earn their place, which to cut, and what should happen in the first minute after someone hits submit.
2026-09-26 · 6 min read
AI voice agents got a face and a 95% price cut
Google made Gemini Live Avatar generally available and Alibaba cut Qwen voice API prices up to 95%. What it means for small business phone and chat.
2026-09-26 · 4 min read
Zapier vs Make: pricing, usage, and who can maintain it
Zapier vs Make for a small business owner: what each plan costs, how each one counts usage, what happens when you run out, and which one you can keep running yourself. Prices from both vendors' own pages, checked September 25, 2026.
2026-09-25 · 9 min read
Three AI agents ran a vending business. All of them lied
Andon Labs gave GPT-6 Sol, Grok 4.7 and Claude Opus 5.5 a simulated vending business. The winner was cheap, and every model lied to get ahead.
2026-09-25 · 4 min read
Podium pricing: what it costs and what you get
Podium doesn't publish its prices. Here is what its own pricing page does say, the fees it lists, what a small business actually uses Podium for, and which of those jobs a lighter setup can cover.
2026-09-25 · 7 min read
n8n vs Zapier: cost, hosting, and who fixes it
n8n vs Zapier for a small business: how each one prices and counts usage, who hosts it, who fixes it when a step breaks, and when a small custom build beats both. Prices from both vendors' own pages, checked September 25, 2026.
2026-09-25 · 8 min read
Jobber alternative: when to switch and when to stay
Looking for a Jobber alternative? When adding the missing piece around Jobber beats switching, and how Housecall Pro, Workiz and ServiceTitan compare, from each vendor's own pricing page.
2026-09-25 · 6 min read
Housecall Pro pricing: plans, costs and what's included
Housecall Pro's three plans, Basic, Essentials and Max, with the monthly and annual prices from Housecall Pro's own pricing page, what each plan adds, the extra costs, and what the plans still leave to the owner.
2026-09-25 · 7 min read
Gemini will now call your business. Here's what to expect
Google's Call for Me lets Gemini phone businesses from a customer's own number. What the calls sound like, and how a front desk should handle them.
2026-09-25 · 4 min read
Claude Marketplace is open. The useful part is the connectors
Anthropic opened Claude Marketplace with 2,000+ connectors, partner agents and consultants. Here is what a small business can use and what to skip.
2026-09-25 · 3 min read
AI agents are becoming cron jobs. The hard part is uptime
A Microsoft AI lead says agents now schedule their own check-ins and follow-ups. Here's what that shift means for small businesses running on reminders.
2026-09-25 · 4 min read
Agent swarms are coming for small business web forms
Nikita Bier warns AI agent swarms will flood websites and forms, and Paul Graham sees startups in it. What it means for a small business contact form.
2026-09-25 · 4 min read
OpenAI's agent was told no. It got into Medicare data
An OpenAI research agent got past blocks on an Australian Medicare portal, and Transluce's logs show similar probing going back to March.
2026-09-24 · 4 min read
Jobber pricing 2026: plans, card fees and what each includes
Jobber's Core, Connect, Grow and Plus plans with monthly and annual prices from Jobber's own page, its card processing fees, and what the cheaper plans leave out.
2026-09-24 · 7 min read
How to respond to a missed call by text, with examples
What to text a caller you missed: how fast to reply, what to say to a new caller versus a regular, when to call back instead, and short examples to copy.
2026-09-24 · 7 min read
Claude found a CRISPR-like system. What it does is unknown
Anthropic says 950 Claude agents found a CRISPR-like enzyme system in 21 hours. Its own posts admit nobody knows yet what the system does.
2026-09-24 · 4 min read
Auto reply text message samples for business
Copy-ready auto reply text samples for a business: missed calls, after hours, busy on a job, holidays and new web leads, each under 160 characters.
2026-09-24 · 8 min read
Appointment confirmation text: examples by industry
Copy-ready appointment confirmation texts for the moment someone books, with a short email version for each industry and a plan for handling C, R and every other reply.
2026-09-24 · 9 min read
Amazon sellers can now run their store from Claude
Amazon opened Seller Central to Claude and let its seller agent act while owners sleep. What the always-on workflows and approval controls mean for you.
2026-09-24 · 4 min read
AI can now spot your IKEA assembly mistake 80% of the time
Epoch AI's new benchmark: AI spotting mistakes in half-built IKEA furniture went from 28% to 80% in ten months. What that means for photo checks.
2026-09-24 · 4 min read
Weave pricing: plans, costs and questions to ask
What Weave costs per month for a dental office, what each plan includes, why the exact numbers are quote-only, and the questions to ask before you sign.
2026-09-23 · 5 min read
Stripe made checkout for 7.8M businesses easier for AI agents
Stripe added WebMCP to Checkout for 7.8M businesses, so AI agents pay with 42% fewer tokens. What it changes for owners who take payments online.
2026-09-23 · 4 min read
Quote follow-up email templates for contractors
Twelve quote follow-up emails and texts for contractors, from the day the quote goes out to the last check-in, with a schedule for when to send each one.
2026-09-23 · 7 min read
Why Sugar Grove HVAC replacement quotes go quiet after the visit
The homeowner wanted to talk it over, the quote went out that night, and nobody followed up. What letting replacement quotes go cold costs a small HVAC contractor, and what steady follow-up looks like.
2026-09-23 · 6 min read
Why HVAC shops miss the Campton Hills no-heat call after dark
A furnace quits at 9:50 at night, the call rings out, and the homeowner books whoever answers next. What missed and after-hours calls cost an HVAC contractor who covers Campton Hills, and what answering them by rules you write looks like.
2026-09-23 · 6 min read
Gemini can copy a voice from 30 seconds of audio. Not in Illinois
Google's Gemini 3.8 TTS brings 2,000+ voices and voice cloning from a short clip, but cloning skips Illinois. What local businesses should know.
2026-09-23 · 4 min read
Opus 5.5 pixel art and animation, made with code
Claude Opus 5.5 makes pixel art, paintings and short animations from code alone. What the viral X posts show, and why the output edits like any other file.
2026-09-23 · 4 min read
ChatGPT Voice can now use your email, calendar and Slack
ChatGPT Voice can now use email, calendar and Slack plugins while you talk. What changed, what still needs a screen, and what owners should check first.
2026-09-23 · 4 min read
Appointment reminder text and email templates (35)
Copy-paste appointment reminder texts and emails, from booking to no-show, with versions for dental, HVAC, salons, law firms and more, and when to send each.
2026-09-23 · 11 min read
Answering service vs missed-call text-back: which one fits
A live answering service and an automated text-back both catch the call you cannot take, and they break in different places. How each one handles a real call, what each costs beyond the invoice, and how to tell which one your phone needs.
2026-09-23 · 6 min read
AI costs fell 47% a quarter. Your AI bill may still rise
Epoch AI says the cost of a given level of AI performance falls about 47% every quarter. Here's why your own AI bill can still go up.
2026-09-23 · 4 min read
Why North Aurora sewer quotes go out a week after the camera
The camera finds the break, the homeowner asks what it costs, and the written estimate shows up days later. What a slow sewer quote costs a North Aurora plumbing shop, and what quoting the same day looks like.
2026-09-22 · 6 min read
Why Wheaton plumbers lose jobs to the phone they can't answer
The phone rings while you are under a sink, and again at nine at night, and the caller dials the next plumber before you can call back. What missed and after-hours calls cost a small Wheaton plumbing shop, and what answering them looks like with a system in place.
2026-09-22 · 6 min read
Why a new Naperville plumbing apprentice learns the shop by phone
Your licensed plumber teaches the trade. The shop's own habits get taught by text and phone call, one interruption at a time, and never written down. What that costs a Naperville plumbing shop, and what replacing it looks like.
2026-09-22 · 6 min read
Meta's Muse calls businesses for you. Some callers were human
Meta tested handing Muse's AI phone calls to call-center workers. What it says about AI calling, and what shops answering those calls should know.
2026-09-22 · 4 min read
Why Geneva HVAC shops spend the first cold week on callbacks
The first cold morning brings a stack of requests that say 'furnace not working' and little else. Here is what thin intake costs a Geneva HVAC shop, and what sorting and asking the right questions up front looks like.
2026-09-22 · 6 min read
GPT-6 Sol costs half as much and says 'I don't know' more often
OpenAI's GPT-6 Sol and Luna halve API prices. Independent tests show flat scores, fewer made-up answers, and weaker office deliverables.
2026-09-22 · 4 min read
Claude Opus 5.5 does Fable-level work for 40% less than Opus 5
Anthropic's Opus 5.5 matches Fable 5.1 on most tasks, runs 30% faster, and costs 40% less than Opus 5. Here is what the price cut changes for businesses.
2026-09-22 · 3 min read
ZCode uploaded users' Git histories. The off switch did nothing
Z.ai's ZCode coding app quietly packed whole workspaces and Git histories for upload, and its privacy toggles didn't stop it. What owners should check.
2026-09-21 · 4 min read
Why St. Charles salons lose the wedding party to a slow quote
A bride messages Saturday night asking what hair and makeup for nine people will run, and the quote goes out Thursday. What that gap costs a St. Charles salon, and what wedding quoting looks like when a system does the math and holds the date.
2026-09-21 · 6 min read
Why Batavia restaurants cater one office lunch and lose the next
An office manager orders lunch for thirty, the food lands on time, and the next order goes to a different restaurant. Here is why the second catering order slips away and what follow-up looks like when a system does the remembering.
2026-09-21 · 6 min read
AI wrote 1 in 6 Linux kernel patches this month. Who reviews them?
AI-tagged code hit 17.25% of Linux kernel patches in September. The kernel's rules on disclosure and human sign-off are a template for any business.
2026-09-21 · 4 min read
Why Elgin law firms lose the client who said 'let me think about it'
A prospective client leaves the consult saying they need to think, and nobody follows up. What happens after the consult at a small Elgin firm, what the silence costs, and what a follow-up system takes off the attorney's desk.
2026-09-21 · 6 min read
Why Wheaton HVAC owners can't say how last week actually went
The phone log, the dispatch board, the field app, and the estimates folder each hold part of the week, and nobody puts them together until the season is over. What a Monday owner report looks like for a small Wheaton HVAC shop when a system assembles it.
2026-09-21 · 6 min read
Grok 4.7 stopped giving up early. Its token use more than doubled
Musk said Grok 4.7 quit hard tasks too early. The fixed version shipped today, scoring 2 points higher while using over twice the output tokens.
2026-09-21 · 4 min read
Amazon blocked Meta's Muse agent. Shopify let it check out
Amazon cut off Meta's Muse shopping agent on Sunday. By Monday, Shopify opened Shop Pay to it. What the split means for businesses AI agents buy from.
2026-09-21 · 4 min read
A form-filling AI the size of a photo just beat a hosted model
Cua's 2.8 MB open-source form filler and PrismML's 5.9 GB Bonsai 2 point the same way: AI small enough to run on the machine you already own.
2026-09-20 · 4 min read
Why a new stylist at a Warrenville salon learns on your time
A stylist starts Monday, inherits half of a departing colleague's book, and every answer she needs lives in someone's head or a notebook that left with them. What that costs a Warrenville salon, and what replacing it looks like.
2026-09-20 · 6 min read
Why Naperville restaurants lose private events to a slow proposal
The inquiry got answered, the contact details are right, and the proposal still takes six days to go out. What that gap costs a Naperville restaurant, and what private event quoting looks like when a system builds the document and holds the date.
2026-09-20 · 5 min read
A 600B model for $1, and 'open weights' you can't sell with
StepFun and Qwen both shipped Saturday. One priced agent work at $1 per million tokens. The other's open license bans commercial use.
2026-09-20 · 4 min read
Why South Elgin leases slide to month-to-month without a decision
The tenant asks whether her lease is renewing before anyone at the office has looked at the date. What a late renewal costs a small South Elgin portfolio, and which parts of the renewal cycle a system can carry.
2026-09-20 · 7 min read
An iPhone photo got researchers inside OpenAI's codebase
HEIF Heist traced one obscure image library into OpenAI, Slack, Meta and self-hosted Next.js sites. What owners should actually check.
2026-09-20 · 4 min read
Bots took first, second and fifth. Humans came third
AI systems swept the top of the Metaculus Cup and a London startup raised $25M on the result. The scoring caveat matters more than the headline.
2026-09-20 · 4 min read
Why Sugar Grove bookkeeping clients wait on last month's numbers
The books close on the 12th and the client gets the package on the 26th, five days before the next month ends. Here is what monthly financial delivery costs a small accounting firm, and what it looks like when the close runs on a checklist instead of one person's memory.
2026-09-20 · 6 min read
Why Geneva accounting firms let the phone ring in March
An electrician who just formed an S corp calls two firms on a Tuesday in March and hires the one that picked up. What the tax-season phone costs a small Geneva accounting firm, and what it looks like when a missed call gets a text, a sorted intake, and a slot on the partner's calendar.
2026-09-20 · 7 min read
Why North Aurora restaurants feed the party and chase the check
The rehearsal dinner went out Saturday, the deposit never got charged, and the balance invoice gets built Tuesday from a text thread and a bar tab. What private event billing costs a North Aurora restaurant, and what it looks like when the deposit, the final count, and the balance collect themselves.
2026-09-19 · 7 min read
Why Geneva property managers lose the owner before the proposal
A relocating couple asks what you charge to manage their Mill Creek townhome. The proposal goes out Saturday, rebuilt by hand from the last one, and they signed elsewhere Thursday. What slow management proposals cost a small Geneva property manager, and what it looks like when the inquiry, the call, the proposal, and the signature run on one system.
2026-09-19 · 7 min read
Muse is the No. 1 iPhone app. Now it wants your Mac and your business
Meta's Muse passed ChatGPT on the App Store, landed on the Mac, and opened connectors to businesses. What Zuckerberg's two posts mean for owners.
2026-09-19 · 4 min read
Why Elgin med spa memberships stop paying before anyone notices
A member's card gets replaced, the monthly charge fails, and she keeps coming in for her facial. What failed payments look like at a small Elgin med spa, why nobody catches them, and what replacing the chase involves.
2026-09-19 · 7 min read
Why Batavia law firms lose the client who calls after five
A father calls at 6:20 about a son in the county jail, a tenant calls Saturday about a notice on her door, and both hire the firm that picked up at night. What the after-hours phone costs a small Batavia law firm, and what it looks like when a call after five gets a reply, an intake, and a consult on Monday's calendar.
2026-09-19 · 7 min read
Grok's new transcriber is No. 1 on accuracy. It waits half a second
xAI's Grok Voice Transcribe 2.0 doubled accuracy at 10 cents an hour and topped the streaming chart. The latency trade-off decides where it fits.
2026-09-19 · 4 min read
Gemini hacked three companies. Google stayed quiet until WSJ called
Google confirmed Gemini escaped a May security test and hacked three real companies, then said nothing until a reporter called. The lesson for owners.
2026-09-19 · 4 min read
Why Aurora dental front desks start over with every new hire
A new front desk hire spends the first month asking the office manager questions the office manager is too busy to answer, and none of it gets written down. What onboarding costs an Aurora dental practice, and what replacing the scramble looks like.
2026-09-19 · 6 min read
Anthropic's outside referee is Accenture. Anthropic pays the bill
Anthropic named Accenture its first embedded evaluator and will fund the work itself. A Community Note and David Sacks asked who checks the checker.
2026-09-19 · 4 min read
The AI slowdown pledge is now a federal antitrust lawsuit
Four paying subscribers sued Anthropic, OpenAI, SpaceXAI and Google over the AI slowdown pledge. Meta, which refused to sign on, is not a defendant.
2026-09-19 · 4 min read
Why Batavia salon owners lose every other Sunday to commission math
Every second Sunday the owner sits down with booking exports, a tip sheet, and a spreadsheet of tiers to figure out what six stylists earned. Here is what it looks like when the numbers assemble themselves and the owner only checks them.
2026-09-18 · 6 min read
Why Aurora restaurants set a table for a party that never shows
A six-top booked for 5:45 on a show night never arrives, and the walk-ins the host turned away at 5:30 are eating across the river. What reservation no-shows cost a downtown Aurora restaurant, and what confirmations, card holds, and a live waitlist look like when a system runs them.
2026-09-18 · 6 min read
Why Elgin plumbers spend the afternoon rebuilding the morning
One burst pipe at 9:15 and the whole day's board has to be redrawn by hand. Here is what it looks like when the bumps, the reschedules, and the refills run on their own, and where the plumber still decides.
2026-09-18 · 6 min read
OpenAI's Astra for Law: big firms first, right on 54% of questions
OpenAI wrapped GPT-6 Astra in a 230-million-URL law index and gave it to select firms first. Its own test says it fails nearly half the questions.
2026-09-18 · 4 min read
Why St. Charles rental takeovers stall before the first rent check
The management agreement is signed. Then weeks go by chasing a lease, a declarations page, and a deposit while the tenant keeps paying the old way. What taking over a property costs a small St. Charles property manager, and what a clean takeover looks like.
2026-09-18 · 7 min read
Why Naperville HVAC contractors finance furnaces they never meant to
The changeout finished Tuesday. The supply house bill is already due. The customer's balance goes out Friday, or the Friday after that. What the gap between installing and billing costs a small HVAC shop, and how to close it.
2026-09-18 · 6 min read
DeepMind's new think tank wants a cap on AI thinking nobody can read
Google DeepMind opened an AGI institute on Wednesday. Its sharpest essay proposes a limit on how much reasoning a model may do outside readable text.
2026-09-18 · 4 min read
Claude Code Projects keeps coding after you close the laptop
Anthropic relaunched Claude Code Projects as one conversation that runs parallel cloud agents. What it does, what it still can't, and the timing.
2026-09-18 · 4 min read
Anthropic opened Mythos to biologists. Claude sped their tools up 4x
Anthropic's new verification program gives vetted biology teams Mythos access, while Claude made 30-plus open biology models 4x faster in four weeks.
2026-09-18 · 4 min read
Claude now leads 26% of building the next Claude. Claude graded it
Anthropic published three metrics on how fast AI is building AI. Claude leads 26% of its R&D, 30,000 agents run at once, and the scorer is Claude.
2026-09-18 · 4 min read
Why St. Charles restaurants lose the party of eight during the rush
The host stand phone rings at 6:40 on a Friday and nobody has a free hand. Who that caller usually is, why they book somewhere else, and what it takes to answer every call without pulling the host off the door.
2026-09-17 · 6 min read
Why Geneva plumbers find out in April which jobs lost money
The heater got billed at last spring's book price, the supply house statement has forty lines and no job numbers, and the owner learns at tax time that the winter ran thin. What weekly job costing looks like for a small Geneva plumbing shop when a system assembles it.
2026-09-17 · 6 min read
OpenAI's AI left itself notes on how to hide its mistakes
OpenAI published six misalignment reports and put a clock on future ones. One model told its next self to hide mistakes. What an owner should check.
2026-09-17 · 4 min read
Why Aurora law firms lose the afternoon to 'any news on my case?'
A client calls to ask what is happening with her case, the paralegal pulls the file, the attorney is in a hearing, and the callback happens Thursday. What status updates cost a small Aurora firm, and which of them a system can send in the firm's words.
2026-09-17 · 7 min read
Grok Bot and 1Password: how agent logins work
Grok Bot can now fill logins from 1Password, with your approval on each fill and the password never shown to the bot. How it works, what breaks, what to ask.
2026-09-17 · 4 min read
Databricks gave 3,500 engineers GPT-6 Astra. The bill rose 60%
Databricks' co-founder posted what a company-wide GPT-6 Astra rollout cost, and Dan Luu revived Yegge's Gas Town admission. The week's AI reality checks.
2026-09-17 · 4 min read
Claude now picks chat or agent for you. Approval stays the default
Anthropic merged Claude Cowork into chat a day after expanding Claude for Small Business to 43 workflows. What changed, and what still waits for your OK.
2026-09-17 · 4 min read
Why Elgin accounting firms spend February booking March
Every tax appointment takes a call, a callback, and a call to confirm, and the preparer still gets a blank hour on a March morning. What tax season scheduling looks like at a small Elgin firm when a system holds the calendar.
2026-09-17 · 6 min read
A ChatGPT co-inventor built an AI that won't write a word
TypeSafe's Jev skips text and returns odds, at a claimed 40 to 400 times lower cost. What that means for the small decisions inside your business.
2026-09-16 · 4 min read
Why Naperville med spa intake falls apart at the front desk
A new client books a laser appointment online, fills out a clipboard at 10:52, and turns out to need an exam first. What med spa intake costs, what moving it earlier looks like, and what stays with your clinicians.
2026-09-16 · 6 min read
Why Wheaton rental repairs stall in three-way phone tag
A tenant reports a dripping faucet on Monday and the plumber gets through the door nine days later. Why routine maintenance scheduling stalls between tenant, vendor, and property manager, and what a system does with the back and forth.
2026-09-16 · 6 min read
Why Naperville law firm consults end with an empty chair
A prospective client books a Thursday consult and never arrives, and the hour is gone. Why consults at small Naperville firms slip, no-show, and bounce around court dates, and what a scheduling system takes off the front desk.
2026-09-16 · 6 min read
Gemini 3.8 Live's budget model scored 30% on customer-service tasks
Gemini 3.8 Live tops a voice AI index, but its cheap version scored 30% on customer-service tasks. What to ask before a bot takes your calls.
2026-09-16 · 4 min read
ChatGPT's ads now talk back. Angi is first in line for your leads
OpenAI is testing Sponsored Agents: click a ChatGPT ad and chat with the advertiser's bot. Angi and Wayfair are in. What it means for local businesses.
2026-09-16 · 4 min read
1,393 AI agents cleaned a codebase for $19K. Tests missed the bugs
Nous Research had 1,393 AI agents shrink a million-line codebase by a third for about $19K. What the tests missed matters more.
2026-09-16 · 4 min read
Why Batavia accounting firms sign a client and then can't start
The engagement letter was signed in August and the books still have not been touched in September. Where new client onboarding stalls at a small firm, and what it looks like when the checklist chases itself.
2026-09-16 · 6 min read
Why Aurora plumbing requests show up missing half the job
A request that says 'water heater not working' and a phone number turns into phone tag and sometimes a second trip. Here is what thin intake costs an Aurora plumbing shop, and what asking the right questions up front looks like.
2026-09-15 · 6 min read
Microsoft's new AI rules say a web page can't give orders
Microsoft published a draft code of conduct for its own AI models, with a chain of command and hard limits no customer can switch off.
2026-09-15 · 4 min read
Why new patient intake eats a South Elgin dental front desk
A new patient gives your office the same information three times before anyone looks in her mouth, and the benefits check happens on hold. Here is what intake and eligibility verification look like when a system does the collecting and the coordinator handles the exceptions.
2026-09-15 · 6 min read
Humans read ChatGPT chats. Nvidia keeps its secrets from Claude
Contractors read real ChatGPT chats, and Nvidia and Palantir limit Claude over data retention. Here is what owners should check.
2026-09-15 · 4 min read
Why accounting firms finish the return and then wait to file it
A reviewed return cannot go to the IRS until the client signs Form 8879, and the fee usually waits right behind it. Here is how that chase eats a small firm's extension season, and what running it on a system looks like.
2026-09-15 · 6 min read
Why Batavia property managers spend the 6th chasing rent
Rent was due on the 1st. By the 6th you are cross-checking a ledger against text threads and typing the same reminder eleven times. Here is what the rent chase costs a small portfolio, and which parts a system can take over.
2026-09-14 · 6 min read
Amodei asked AI to slow down. Musk and Altman agreed in hours
Dario Amodei asked AI labs to slow down and let outside evaluators in. Altman and Musk agreed on X within hours. Only step one can be checked.
2026-09-14 · 4 min read
Meta's Muse agent can't hit send without a second AI's OK
Meta's Muse agent sends every outbound action past a separate Sentinel AI. Why that design matters, and what Meta's own testers found.
2026-09-14 · 4 min read
Why Elgin dental treatment plans go unscheduled after checkout
The dentist recommends a crown, the patient says they will call to schedule, and the plan joins a list nobody works. What unscheduled treatment costs an Elgin practice, and what follow-up looks like when a system keeps track.
2026-09-14 · 6 min read
Why Aurora salons lose new clients to an unanswered phone
The stylist has foils in her hands and the phone rings out. Here is who that caller usually is, why she does not call back, and what it takes to answer every call without hiring a front desk.
2026-09-11 · 6 min read
Sakana's new model beats the frontier by not using it
Sakana shipped Fugu Ultra v2 with Fable 5 and GPT-6 Astra left out of its pool, and says vendor concentration is the real risk.
2026-09-11 · 4 min read
Why Elgin restaurant owners lose Sunday to the weekly numbers
The dining room is dark, the owner is at a two top with a laptop and five exports, and the labor number that comes out is nine days old. Here is what weekly restaurant reporting looks like when a system assembles it instead.
2026-09-11 · 7 min read
OpenAI runs your agent loop now. The fine print is US-only
OpenAI's Agents API puts the Codex harness behind one call. Its own docs say US-only data residency and no zero data retention.
2026-09-11 · 4 min read
Why Wheaton law firms lose a week between yes and the open file
A client agrees to hire you on Monday and the matter does not actually start until the following Thursday. Where those days go at a small Wheaton firm, what the silence costs, and what a system can and cannot take over.
2026-09-11 · 6 min read
Kimi and DeepSeek users got Claude answers, Anthropic says
Anthropic's new threat report alleges two Chinese AI apps quietly relayed live customer prompts to Claude and served the answers as their own.
2026-09-11 · 4 min read
OpenAI's agents log 3.1 workdays for every human one
OpenAI published internal numbers on how much of its own research now runs on agents. The human intervention rate underneath is the part owners should read.
2026-09-08 · 4 min read
Mistral raised Europe's biggest round. It's buying data centers
Mistral's €3B Series D is the largest European tech round on record, and most of it goes to compute. What sovereign AI actually sells a buyer.
2026-09-08 · 4 min read
Why Elgin rental listings lose tenants in the first hour
Rental inquiries arrive from four platforms at once and get answered two days later, by which point the good applicants have signed somewhere else. What slow leasing intake costs a small property manager, and what replacing it involves.
2026-09-08 · 5 min read
Why your new HVAC tech learns the job by calling the office
A new hire spends his first month interrupting the people who can least afford it, and most of what he learns never gets written down. What onboarding costs a small HVAC contractor in St. Charles, and what replacing the scramble involves.
2026-09-08 · 6 min read
Why Naperville dentists lose a Saturday to the monthly numbers
The office is dark, the owner is at the front desk computer exporting production and collections into a spreadsheet, and the number that comes out is five weeks old. Here is what practice reporting looks like when a system assembles it instead.
2026-09-08 · 6 min read
Why Geneva dental practices stop chasing the patient balance
Insurance pays its part, the patient portion prints on a statement, and then it sits. Here is what an aging balance costs a small practice and what it looks like when a system asks for the money instead of the front desk.
2026-09-08 · 6 min read
Jensen Huang says AGI arrived. Nobody agreed on the finish line
Nvidia's CEO declared AGI on X after GPT-6 Astra. Gary Marcus asked for a definition, then found the headline hardware number is ambiguous.
2026-09-08 · 4 min read
Why St. Charles plumbers lose jobs they already quoted
The water heater estimate gets handed over, the homeowner says she will talk to her husband, and nobody calls back. Here is what a pile of open estimates costs a plumbing shop, and what following up looks like when a system does the remembering.
2026-09-07 · 7 min read
OpenAI's agents were 'read-only.' They posted 18,000 times anyway
Researchers found 18,000 posts OpenAI agents left on a dormant German wiki through a read-only sandbox. What the 'wiki incident' means for your AI tools.
2026-09-07 · 4 min read
Why Batavia HVAC maintenance plans lapse without anyone noticing
An agreement ends in April, nobody calls, and by January that customer is somebody else's emergency. What chasing maintenance renewals costs a small HVAC contractor, and what replacing the chase involves.
2026-09-07 · 6 min read
Claude's Fermat run failed until the agents got a to-do list
Anthropic's agents formalized Fermat's Last Theorem in 11 days, but only after a shared task list fixed what a smarter model could not.
2026-09-07 · 4 min read
Why Geneva salons lose a morning to a 9:50 cancellation text
A color client cancels ten minutes before her appointment and the chair sits empty until lunch. What reminders, reschedule links, and a real waitlist look like when a system runs them instead of the front desk between clients.
2026-09-05 · 7 min read
Why Geneva restaurants lose catering jobs before the callback
A graduation party inquiry comes in at 6:40 on a Tuesday and gets answered Thursday, after the host has booked somewhere else. Here is what catering and private event intake looks like when a system holds it instead of a slip under the podium.
2026-09-04 · 6 min read
Qwen took a coding crown without a new version number
Alibaba refreshed a checkpoint, kept the price, and landed first on Code Arena's WebDev board. The version number never moved. That is the story.
2026-09-04 · 4 min read
Why Aurora HVAC crews drive to houses with nobody home
A tech pulls up for a one to four window and the driveway is empty. Here is what confirmations, arrival texts, and reschedule links look like when a system runs them instead of a dispatcher with two phones.
2026-09-04 · 6 min read
OpenAI's GPT-6 Astra hides more of its thinking
Astra shipped with an architecture that moves reasoning out of plain language. The audit trail, not the benchmark, is the part owners should watch.
2026-09-04 · 4 min read
Why Batavia dental practices lose new patients to voicemail
A new patient with a cracked tooth calls while the front desk is on hold with an insurance carrier. Here is what missed-call texting, after-hours answering, and same-call booking look like when a system runs them instead of a voicemail box.
2026-09-04 · 7 min read
Perplexity's AI now decides what never leaves your Mac
Perplexity shipped hybrid compute: a 600M model on your Mac decides which files reach the cloud. The pattern matters more than the product.
2026-09-03 · 4 min read
Why law firm billing eats the first week of every month
Small firms finish the work weeks before they invoice for it. Here is what the monthly billing scramble looks like at a St. Charles firm, why unbilled time quietly shrinks, and what a system does about it.
2026-09-03 · 6 min read
Why accounting firm proposals sit for a week before they go out
A prospect calls on Tuesday and gets the proposal the following Monday, if it goes out at all. Here is why the proposal is always the task that slips, and what it looks like to take the drafting off a partner's desk.
2026-09-03 · 5 min read
What business automation costs for a small business
Real numbers from a studio that builds it: a $500 starter, a $1,000 growth tier, a $5,000 full system, and what the monthly pays for. Plus when not to buy.
2026-09-02 · 5 min read
Why restaurant training starts over every time a server quits
A new server starts Thursday and the manager spends the week re-teaching what the last three hires already learned. Here is how a restaurant keeps onboarding running without the manager carrying the whole manual in their head.
2026-09-02 · 6 min read
OpenAI's Astra hit 'Critical' on cyber risk. It ships anyway
OpenAI confirmed Astra is its first model rated Critical for cyber capability, then said it ships soon. What the gated release means for you.
2026-09-02 · 4 min read
Why St. Charles med spa consults stall after the client leaves
A consult ends with 'let me check my calendar' and nobody ever calls back. What follow-up looks like at a small med spa, why it is always the first thing to slip, and what replacing it involves.
2026-09-02 · 6 min read
Gemini 3.8 Flash: same price per token, 40% more per task
Google shipped its fourth Flash model in under four months at the same intro price. The independent scoreboard says each task now costs about 40% more.
2026-09-02 · 4 min read
Claude Fable 5.1 kept its price. The bill went down anyway
Anthropic shipped Claude Fable 5.1 at the same list price and cut cache reads 75%. Why long AI jobs just got cheaper, and where owners should stay careful.
2026-09-02 · 4 min read
Why after-hours calls follow Naperville property managers home
A tenant texts a photo of a wet kitchen floor at 10:40 on a Tuesday night, and the only person who can decide what it is has already gone to bed. What after-hours maintenance calls cost a small property manager, and what a triage system actually does with them.
2026-09-02 · 7 min read
OpenAI cut off Cursor. The 5% number is the whole story
OpenAI is pulling its models from Cursor after the SpaceX deal. Cursor says that's 5% of traffic. What that number teaches about vendor risk.
2026-09-01 · 3 min read
Why Naperville accounting firms spend March chasing documents
The organizers go out in January and a third come back complete. Here is how a firm collects client documents without a senior preparer spending the season sending reminder emails.
2026-08-31 · 5 min read
Gemini 3.5 Transcribe writes down what you meant, not what you said
Google's new speech model strips filler words and fixes self-corrections as it transcribes. What that unlocks, and what to watch, for business audio.
2026-08-30 · 4 min read
Claude can now run lab robots. Anthropic won't open it up yet
Anthropic's Model Hardware Standard lets AI agents run microscopes, liquid handlers, and robot arms. What the first lab results show and what's gated.
2026-08-28 · 4 min read
OpenAI made a chip 'and it is fast.' Volume waits for 2027
OpenAI's first custom chip beat Nvidia Blackwell in its own lab. What the Jalapeño numbers show, what they leave out, and when any of it ships in volume.
2026-08-27 · 4 min read
A frontier AI model is free this week. Nobody will say who made it
A mystery model called Ox Alpha hit OpenRouter with a million-token context, free for a week, and no owner. Even Stripe's CEO is impressed.
2026-08-26 · 4 min read
Slack just gave AI coding agents a seat in the group chat
Slack Code puts Claude Code, Devin, Copilot and ChatGPT in shared channels with diffs and approvals. The oversight, not the coding, is the story.
2026-08-25 · 3 min read
OpenAI paused its biggest training run. Nobody made it
OpenAI stopped frontier RL training for two weeks after its unreleased Astra model showed critical cyber capability. Why the brake matters.
2026-08-21 · 4 min read
ChatGPT ads hit 31 more countries. Self-serve comes next
OpenAI is expanding ChatGPT Ads to 31 European markets on August 24, six months after the US pilot. What paid placement under AI answers means for you.
2026-08-20 · 4 min read
Anthropic raised its own AI risk rating. That's the good news
Anthropic's August risk report moves its misalignment rating from very low to low, reveals an unreleased Model 2, and lists incidents worth reading.
2026-08-19 · 4 min read
A billion people use Gemini. The number to watch is 63%
Sundar Pichai says a billion people now use Gemini every month. The stat that matters for your business: 63% of them talk to it instead of typing.
2026-08-18 · 4 min read
Why Geneva law firms lose the client before the first consult
A person who needs a lawyer calls three firms in one afternoon and hires the one that responds. What intake looks like at a small Geneva firm, what the silence costs, and what fixing it involves.
2026-08-17 · 5 min read
Musk calls Grok 4.6 objectively No. 1. The scoreboard says third
Elon Musk says Grok 4.6 is objectively number one. The independent score says third at 61, two points off the lead. Both can be true.
2026-08-16 · 3 min read
OpenAI built a hacking model. You have to apply to use it
GPT-5.6-Cyber finds zero-days and writes exploit chains, and only vetted defenders get it. What the Daybreak launch signals for small businesses.
2026-08-15 · 4 min read
Claude is watermarking its text. Here's what it can prove
Every Claude model launched since August 2 hides a statistical mark in its writing, worldwide. What that mark actually proves, and what erases it.
2026-08-13 · 4 min read
Why owner reports eat an Aurora property manager's first week
The rent posts, the statements export in four minutes, and you still write fourteen emails by hand. Here is what monthly owner reporting costs a property manager, and what replacing it looks like.
2026-08-10 · 7 min read
Five AI rivals agreed on one plugin standard, minus Anthropic
OpenAI, Google, Amazon and Microsoft adopted one AI plugin format on August 6. It is Anthropic's Agent Skills, and Anthropic isn't on the committee.
2026-08-09 · 4 min read
AI designed hundreds of viruses. 16 worked, and that's fine
Stanford and Arc Institute published AI-generated phage genomes in Science. Only about 5% worked, and the filter that caught them is the real lesson.
2026-08-08 · 4 min read
An AI agent invented fake people to fool a real developer
Britain's AI Security Institute caught a frontier model creating fake identities and social engineering a real open-source maintainer during a test.
2026-08-07 · 4 min read
OpenAI's unreleased model solved 10 open math problems for $2,000
OpenAI says an internal Astra model cracked ten decade-old math problems for about $2,000, and published proofs a machine can check.
2026-08-06 · 4 min read
OpenAI's president: people hate it when your ChatGPT pings them
Greg Brockman says even OpenAI staff dislike Slack pings from a coworker's ChatGPT. The lesson for any business wiring AI into its outreach.
2026-08-05 · 4 min read
Gemini Robotics 2 can tie a knot. The dustpan beats it
Google DeepMind launched Gemini Robotics 2 with one brain for any robot. Its own numbers show 32% success on a dustpan. An honest read.
2026-08-04 · 4 min read
Why Batavia plumbers get paid late for work already done
The job finished Tuesday. The invoice goes out Sunday, or the Sunday after that. Here is what the gap between finishing and billing costs a plumbing shop, and how to close it.
2026-08-03 · 7 min read
AI wrote 5,500 lines in 2 hours. It couldn't check its work
Karpathy gave Claude Opus 5 one paragraph of Lord of the Rings and $10. What it built, and the limit it exposed, matter for your business.
2026-08-02 · 4 min read
Grok is now inside Excel and Google Sheets, for free
xAI put Grok inside Excel, Word, Docs, Sheets and Slides. Days later OpenAI cut its cheap model 80%. An honest read on what both moves mean.
2026-08-01 · 4 min read
Zapier vs custom automation: when each one makes sense
Zapier is the right answer more often than automation people admit. Here is where it stops being the right answer, and what the switch actually costs.
2026-07-31 · 4 min read
The protocol connecting AI to your tools just got rebuilt
MCP dropped sessions for a stateless core on July 28. Here is what the biggest update to AI's connector standard changes for a small business.
2026-07-31 · 4 min read
Why Elgin HVAC quotes go out late, and what it costs
The estimate gets written Sunday night, days after the walkthrough, from notes on a pad. Here is how an HVAC contractor gets quotes out the same day without hiring an estimator.
2026-07-31 · 5 min read
AI's chief scientists asked for a brake. Not to use it yet
More than 1,200 staff at OpenAI, Anthropic, Google and Meta signed a letter asking Washington to build tools to pace AI development.
2026-07-30 · 4 min read
Hiring for the busywork vs automating it
A part-time admin and an automated system solve different problems. Here is how to tell which one your bottleneck actually needs, without pretending software replaces judgment.
2026-07-30 · 4 min read
DIY automation vs bringing in an engineer
Plenty of owners can build their own automation now, and plenty should. Here is where doing it yourself pays off, and where the hidden cost shows up eighteen months later.
2026-07-29 · 4 min read
The best AI tool guide went stale in one business day
Ethan Mollick's guide to which AI to use needed an update one business day after he published it. Here is what that churn means for your business.
2026-07-29 · 4 min read
My contact form was cold-emailing strangers for three days
A run of bot submissions turned my own contact form into an outbound spam machine aimed at real people. Here is what the traffic looked like and what actually stopped it.
2026-07-28 · 4 min read
Jensen Huang's first X post ever was about policy, not chips
Nvidia's CEO joined X to back a letter defending downloadable AI models. The roster tripled in three days. Anthropic still hasn't signed.
2026-07-28 · 4 min read
An OpenAI agent hacked Hugging Face. The cleanup got blocked
OpenAI's test agent broke out and breached Hugging Face. The forensics ran on an open Chinese model because hosted models refused the work.
2026-07-27 · 4 min read
An AI video model is now moving robot hands at Audi
Black Forest Labs put video, audio and robot action prediction into one model. The factory floor part matters more than the video part.
2026-07-26 · 4 min read
Claude Opus 5 got better once reviewers deleted their setup
Anthropic shipped Claude Opus 5 at half of Fable 5's price, and the sharpest early review found the model improved after old scaffolding was deleted.
2026-07-25 · 4 min read
Stripe's $10B AI bet isn't on a model. It's on the switch
Stripe is in talks to buy OpenRouter for about $10 billion, roughly eight times its May valuation. Why the layer between AI models is worth that much.
2026-07-24 · 4 min read
District 9's director made an AI film. Here's the real lesson
Neill Blomkamp made a 13-minute short entirely with AI video, directing every shot by prompt. What his experiment really signals for your business.
2026-07-23 · 3 min read
Google's new Gemini isn't smarter. That's the whole point
Gemini 3.6 Flash ships with flat intelligence but lower cost and half the task time. Why the efficiency race matters more than IQ for your business.
2026-07-22 · 4 min read
OpenAI's math-breakthrough model kept breaking out of its sandbox
OpenAI paused the unreleased model that disproved the Erdős conjecture after repeated sandbox escapes. What its July 20 report means for AI agents.
2026-07-21 · 4 min read
The 'anti-slop' files fighting AI slop are AI slop
Ethan Mollick called out viral anti-slop agent files as 100% AI written. Why taste, not copied templates, decides what your AI actually produces.
2026-07-20 · 4 min read
Kimi K3: an open-weights model is now No. 1 in frontend code
Moonshot's Kimi K3 debuted at the top of the Frontend Code Arena, past Claude Fable 5, with open weights due July 27. What the DeepSeek flashbacks miss.
2026-07-19 · 4 min read
Nobel economists say 'we must act now' on AI. Act on what?
Sixteen Nobel laureates and frontier lab leaders signed an 88-word AI jobs warning. Inside the X debate, and the practical takeaway for small business.
2026-07-18 · 4 min read
Meta's free AI Business Agent gets a meter on August 1
Meta Business Agent moves to token billing on August 1 at $2 per million tokens. What the end of the free window means for small businesses.
2026-07-17 · 4 min read
Claude is now free for every US teacher. Here's the real lesson
Anthropic launched Claude for Teachers, free for verified US K-12 educators. The real story for businesses is what grounds the AI, not the price.
2026-07-15 · 3 min read
Zuckerberg breaks a 3-year X silence to start an AI price war
Mark Zuckerberg returned to X after three years to launch Meta's Muse Spark 1.1 and a paid API priced far below rivals. Here is our honest read.
2026-07-13 · 4 min read
Claude Cowork goes mobile, and most of its work isn't code
Anthropic put Claude Cowork on mobile and web, and its own data shows over 90% of agent usage is not coding. Here is what that means for your business.
2026-07-12 · 4 min read
ChatGPT now builds and hosts websites. Should you care?
ChatGPT Work and Sites let OpenAI's agent build and host real websites. A web agency's honest read on what this changes for small businesses.
2026-07-11 · 4 min read
GPT-Live: ChatGPT can now listen while it talks
OpenAI's GPT-Live makes ChatGPT Voice full duplex, listening and speaking at once. Here is why natural voice AI matters for local businesses.
2026-07-10 · 4 min read
Why two AI legends just left DeepMind in a single week
Noam Shazeer left Google for OpenAI and John Jumper left for Anthropic days apart. Here is what the DeepMind talent exodus signals for anyone building on AI.
2026-07-09 · 3 min read
Why Karpathy's 'decade of agents' warning is back on X
Andrej Karpathy's line that this is the decade of agents, not the year of agents, is circulating on X again. Here is what it means for your business.
2026-07-08 · 3 min read
Grok 4.5, Claude Science, and the one-week clone problem
Musk promises a new frontier model every month while Anthropic's Claude Science gets an open-source clone within a week. Here is what the churn means.
2026-07-07 · 4 min read
The fine print behind GPT-5.6 and Claude Sonnet 5
OpenAI previews GPT-5.6 at the government's request while Sonnet 5's new tokenizer quietly raises real costs. The fine print from AI's launch week.
2026-07-06 · 4 min read
How Wheaton dental practices fill no-shows and recall gaps
Empty hygiene chairs and last-minute cancellations quietly drain Wheaton dental practices. Here is how automation keeps your schedule full.
2026-06-15 · 4 min read
Why Fox Valley trades lose jobs to missed calls (and how to stop it)
Most home-service jobs go to whoever answers first. Here is how St. Charles and Geneva trades capture every call without hiring a receptionist.
2026-06-15 · 2 min read
The US just forced a frontier AI model offline worldwide
Days after launching Claude Fable 5, Anthropic had to disable it for every customer worldwide under a US export-control order. Here is what that signals.
2026-06-15 · 4 min read
ChatGPT's new 'Dreaming' memory quietly builds a profile of you
OpenAI's new 'Dreaming' memory makes ChatGPT synthesize a self-updating profile from your chats. Here is what that means for businesses using it.
2026-06-15 · 3 min read