💻 This founder has 50+ employees but talks to his AI agents more than to them

Greg Isenberg, CEO of Late Checkout, has 50+ people in 12 countries. His agents do this:

1. Ideas: he sends a voice note, and an agent grills the idea before anything gets built.
2. Research: agents read rivals' sites, prices and user reviews. Prompt: find the 10 complaints that show up most and quote the users
3. Design: Claude Code makes 4 versions of every screen before any code exists.
4. Build: agents work in parallel overnight. He checks what shipped over coffee.
5. Testing: agents click every flow as an impatient user, a confused one and a phone user.
6. Bugs: an agent opens a fix in under 10 minutes. A human just approves.

Rules: a cheap model for sorting, Opus for anything that ships. Another model scores every output 1 to 10; under 8 goes back. Every agent mistake becomes a new skill, so it only happens once.

The humans handle creative feedback, relationships and telling the agents what they got wrong.

@aipost 🏴

🗣 “The titans of Silicon Valley dig bunkers for themselves”: Jon Ossoff's victory speech on AI

Jon Ossoff won reelection to the Senate in Georgia on Tuesday night, beating Republican Mike Collins. Midway through his victory speech, he went after the people building AI:

Meanwhile, the titans of Silicon Valley dig bunkers for themselves and warn that the artificial intelligence they're creating risks human extinction.

Near the end he talked about his two daughters. He said he refuses to accept that they "will grow up with artificial friends and teachers."

This isn't new for him. At a rally in September, he called for inspectors inside frontier AI labs and an international AI treaty.

@aipost 🏴

📢 Apple is letting every app plug into Siri AI, so you can just ask it to book or order

Any developer can now connect an app to Siri AI, Apple announced on Tuesday. Ask in your own words, and Siri runs the app's own actions: check a delivery, book a reservation, place an order.

Apps can also teach Siri "skills" that chain several actions into one request. Apple's own apps go first: you can ask Siri to find a coffee shop you visited last week, or archive files you haven't opened in over a year.

For now it's a developer preview in the iOS 27.3 and macOS 27.3 betas. Apple says customers get it in a few months, with App Store submissions opening in spring. Siri AI isn't available in the EU on iPhone, iPad and Apple Watch yet.

@aipost 🏴

🤨 One AI model averages 7 em dashes per answer. Claude Opus uses almost none

Artificial Analysis counted every em dash in every answer on its personality benchmark. Plenty of readers treat that dash as a sign a text was written by AI.

The heaviest user is GLM-5.3, from China's Z.ai, with 7.4 per response. Right behind: GLM-5.3-Flash and Meta's Muse Wonder (6.9 each), then Kimi K3 (6.7).

Anthropic's models are split. Claude Fable 5.1 averages 6.4 per answer. Claude Opus 5.5 uses fewer than one every 40 answers, and Sonnet 5.5 and Haiku 5.5 come out at zero.

So does OpenAI's old GPT-3.5 Turbo.

@aipost 🏴

🤖 A Chinese lab just let anyone download an AI that keeps up with Claude and GPT

Moonshot AI posted the full weights of Kimi K3.1 Pro on Hugging Face. Anyone can now run the model on their own hardware.

It's enormous: 2.8 trillion parameters, with 104 billion active at a time, and it can read up to 1 million tokens at once.

On Moonshot's own tests it beats Claude Opus 5.5, Claude Fable 5.1 and GPT-6 Astra on the SWE Marathon coding benchmark: 51.7 against 46.3, 45.6 and 42.5. On Humanity's Last Exam with tools it scores 70.0, ahead of Opus at 67.7.

It still trails them elsewhere. On ExploitBench, a hacking test, Kimi gets 74.7. GPT-6 Astra gets 100.

@aipost 🏴

⚠️ Apple just warned investors that AI agents may defy orders and hide what they do

The warning is new in Apple's annual report, filed on Tuesday. AI agents, it says, can take actions nobody asked for, exceed their permissions, resist instructions and get around security controls.

Then comes the line about trust:

"Such technologies may conceal actions, hide, withhold or manipulate information, or generate inaccurate or deceptive outputs."

Apple adds that some of the damage, "such as deleting data or completing transactions," may be "difficult or impossible to reverse." It also says its ability to control other companies' agents "may be limited."

The risk covers agents Apple builds and ones from third parties, and it lands as Apple opens Siri's AI to every app developer.

@aipost 🏴

📞 Grok Bot can now call your dentist for you

Elon Musk posted that Grok Bot can place phone calls. In one demo making the rounds, a bot dialed a dental office and asked whether they take a patient's insurance.

The bot stays on the line, talks to the person who picks up, and reports back in chat. Musk highlighted the feature with a simple line: it can call your dentist on the phone.

Would you let a bot call a doctor's office for you?

@aipost 🏴

🧠 ChatGPT and Claude can finally share one memory

Your bots used to forget what the other already knew. Notion now lets ChatGPT, Claude and other connected apps read and write the same notes, preferences and skills in your workspace.

Three steps:
1. Connect the app in Notion's gallery, or inside the app itself
2. Paste Notion's setup prompt so the bot checks those notes before answering and saves useful work back
3. Drop the custom instructions it gives you into ChatGPT or Claude settings

Without that prompt, Notion says bots often ignore the connection. Everything lands in a Memories database you can edit. Run it once per bot.

Notion MCP help

@aipost 🏴

✍️ Telegram is testing a way to rewrite your message after you hit send

In the Mac beta right now, you can turn on an AI Editor so that after you send a message it can fix typos, change the tone, or translate it — and the other person never sees an «edited» label.

Telegram may land this update in the main app next.

@aipost 🏴

💰 Claude can now read your bank account — but it can't touch the money

Anthropic is rolling out Personal Finance in Claude as a US beta for Pro, Max and Ultra. You link the accounts you choose through Plaid, then ask everyday questions like where the money went this month.

Access is read-only. Claude can see bank names, balances, credit limits and up to two years of transactions. It cannot move, manage or invest your money, and Anthropic says it isn't built to give investment advice. Money chats are not used to train its models.

ChatGPT launched a similar Plaid link-up earlier this year. Same pitch: the AI sees the numbers so you don't have to dig through statements.

Claude · The Verge

@aipost 🏴

🎬 Google's AI video clips just got 50% longer — and 30% cheaper

Gemini Omni 1.2 Flash is generally available today. It can generate and edit native video up to 15 seconds (was 10), and you can pick any whole-second length from 3 to 15.

You can also keep editing the same clip in chat — swap a product, change the lighting, flip the aspect ratio — instead of starting over. Google cut the price 30% versus Omni 1.1, to about $0.07 per second for 720p.

It is live on the Gemini API and Google Cloud, with SynthID watermarks on the output. Drafts can start at 360p or 720p, then upscale to 1080p or 4K.

Google

@aipost 🏴

🤖 Fresh robots just walked themselves out of the factory

Figure AI's boss posted a short clip: brand-new F.03 humanoids leave the BotQ plant and cross the campus parking lot on their own — past cars, past people, no remote puppet show in the frame.

It's only about 20 seconds. That's what makes it land. Robots that can walk out the door are a lot closer to "they deliver themselves" than last year's lab demos.

@aipost 🏴

💥 Anthropic just told regulators it is on the hook for $895 billion of AI compute

In an updated IPO prospectus, Anthropic disclosed about $895.4 billion in purchase and lease commitments — mostly for computing power. Roughly $70 billion comes due in the next 12 months; about half of the total is expected by the end of 2030.

The biggest named partners in circulating breakdowns: Broadcom ~$161B, Google ~$121B (with more chip spend estimated on top), SpaceX up to ~$121B, Amazon at least ~$110B. Q3 revenue was about $17 billion, up 47% quarter on quarter — and the company swung back to an operating loss as short-term compute ran expensive. Nasdaq ticker planned: ANT.

@aipost 🏴

🗣 Bernie Sanders wants a national freeze on new AI data centers

In a clip circulating today, Senator Bernie Sanders says Democrats must enact a national moratorium on AI data centers, pause advanced AI development, and ban superintelligence — claiming a data-center moratorium is what Americans want.

It lands as markets already price political risk into the US AI buildout after this week's midterms. The clip is campaign/policy rhetoric, not a signed bill.

@aipost 🏴

⚡️ OpenAI says it will pay the full power bill for every AI campus it owns

Two days after midterm votes slammed the brakes on data centers, OpenAI published a National Community Compact.

For future campuses it owns and operates, it pledges to cover the full cost of electricity and the grid upgrades it needs, post financial guarantees if a project is delayed or canceled, publish water, power and noise numbers, and stop asking public officials to sign NDAs. Local priorities get funded too — it points to a $4.8 million free-school-meals grant already shaped with residents in Effingham County, Georgia.

A separate policy blueprint asks lawmakers to turn the same ideas into rules for the whole industry.

@aipost 🏴