Catch every conversation that could improve your business. Free live chat, AI replies & customizable contact pages for startups, SMBs and freelancers. ↓
Make it easier for users to reach you.
Add live chat to your website, app, or Contact Page — and reply directly from Telegram.
Free to start. No credit card required.
Try Knocket: https://t.co/BLs9T8guHr
#Knocket
Note what Airbnb deliberately did NOT automate: safety tickets. And the rule that a human has to be able to explain the PR, even when AI wrote it.
The version that survives contact with real users in one place so nobody has to re-explain themselves.
FREE: https://t.co/gAWi4uGi63
60% of Airbnb's code is now AI-authored.
80% more features. 1.6x PR throughput.
The rule they had to write down: every engineer must explain what they built, even if AI wrote the PR.
Next: on-call agents that close your alert themselves.
Apple just changed who gets to read your whole Mac.
Not a law. A Friday developer post.
Apple's line: Full Disk Access "largely sidesteps" privacy controls. It was built for backup apps.
New rule: "explicit user action." No date, no API.
Apple is now the gatekeeper. #AI
Most websites lose the sale at the question. Mine answers it live — from my phone. 10 minutes to set up. Zero backend.
Save this 🚀 https://t.co/hb9dSq8Y8o
#buildinpublic#founders
Google shipped its best model this week.
Paying subscribers can't have it yet.
A handpicked list of cyber defenders can — with the guardrails off.
Benchmarks: state of the art.
Bloomberg: internally it's uneven at real coding.
Whoever holds the list holds the model. #AI
Introducing Gemini 4 Argon – our new frontier model.
It’s built for complex workflows across coding, enterprise knowledge work, and cybersecurity defense – rolling out today to a set of trusted testers through our Fairwind Program.
@ns123abc The number I can't shake is 6% — reportedly that's how much of a sampled week's research compute went to safety, at the lab that warns loudest.
If they can't spend more than that, nobody can. So don't build assuming the model will always know when to stop.
OpenAI killed its flagship Monday for failing alignment tests.
Tuesday it shipped always-on agents that work for hours unsupervised.
Both are true. Same company. One day apart.
Dots: own cloud computer, 4,000+ apps, Slack and Teams.
$500/mo buys speed. Plus gets nothing. #AI
AMD paid $8.2B for a two-year-old lab whose best product makes a 3D room from photos.
Official line: know the workloads, shape the chip roadmap.
Translation: they can't guess what to build.
Nvidia ships Cosmos. AMD had text and video.
So they bought the godmother of AI. #AI
We are excited to announce that World Labs is joining @AMD.
The research and technical breakthroughs we have achieved since our founding in 2024 have given us a clear vision for AI’s potential to solve problems in the spatial and physical world.
Accelerating the future of spatial and physical intelligence requires scaling our efforts, scaling our reach, and getting closer to the hardware.
We are excited to announce that World Labs is joining @AMD.
The research and technical breakthroughs we have achieved since our founding in 2024 have given us a clear vision for AI’s potential to solve problems in the spatial and physical world.
Accelerating the future of spatial and physical intelligence requires scaling our efforts, scaling our reach, and getting closer to the hardware.
Most iframes I see in the wild are missing at least three things: a title for screen readers, loading="lazy", and a sane sandbox.
So I built a checker. Paste an iframe, get a 0–100 score, plain reasons, and a corrected version you can copy.
Runs in your browser ⬇️🔗
Knocket — a free contact layer for makers who ship.
Add a web widget, mobile SDK, contact page, and unified inbox to your product.
One script tag. Every channel in one place.
100% free · no credit card required · forever.
Learn more: https://t.co/BLs9T8guHr