ChatGPT Work can now keep going on supported websites after you enter the login yourself, and it asks before consequential actions like reservations or payments. The useful rule: let AI do the clicking, keep the final yes. Rolling out for Plus and Pro. #AI#ChatGPT
Say STOP when you'd let go. No fill line. No percentage. Just your nerve. Rate my pour, then follow for the launch of Don't Spill it. #mobilegame#satisfying#gametok#dontspillit
You were going to walk anyway. A KRUE weekend sprint with 50 walkers: $1 each, a 15% fee, a $42 pool, $29.03 for first place. It's a contest, not money per step. Paid contests where available. KRUE is pre-launch, follow for the launch. #walking#stepchallenge
A researcher said her AI agent deleted her inbox after she told it to confirm first. 68% of Americans say they'd never let AI act without reviewing each step.
Tales drafts. It doesn't send. Anything that would go out waits for your OK.
Sources: TechCrunch; YouGov. #AI
Google DeepMind just announced Genie 4: an AI world you can steer while you explore. Type a change and the scene updates in real time. Still early: details can drift, sessions last minutes, and only trusted testers have it.
Demo: Google DeepMind's launch video. #AI#Genie4
The 10,000-step goal likely started as a 1965 pedometer ad. A big 2025 review linked about 7,000 steps a day to major health benefits. KRUE lets you pick your own weekly goal β and change it anytime.
Sources: Lancet Public Health (2025); NIH MedlinePlus (2021). #walking
One glass. No fill line. Hold to pour, let go to stop β but go one drop too far and the whole rim spills over. How many taps would you need? π
Don't Spill it is coming to the App Store and Google Play. #indiegame#mobilegame
ChatGPT has buttons now. OpenAI says GPT-6 Intelligent UI can answer with forms, charts and mini tools you use inside the chat, with rollout expanded to Free and Go users. Availability may vary. https://t.co/nX4iVtJRDB
One rule: don't spill it. Hold to pour, release to stop, and get close to the line without going over. The current build has 30 levels. Follow for App Store and Google Play release updates.
Your goal. Your pace. I'm building KRUE around a weekly step target you choose and milestones you can see. App in development; sample progress shown. Follow for App Store and Google Play release updates.
Your AI should ask first. I'm building Tales to turn a plain-language request into a draft you review before anything goes out. App preview; sample output is illustrative. Follow for App Store and Google Play release updates.
Your app idea deserves a first versionβeven without a paid AI subscription.
ChatGPT β product brief
Claude β stack + build prompt
Lovable / v0 β build, test, repeat
Save your free credits by planning first.
Whatβs one app you wish existed?
@AnthropicAI Hey @AnthropicAI, this is a garbage billing policy. My payment failed Oct 2 and you completely BLOCKED my access to Claude.
I paid the invoice today (Oct 11) to get back in, but my sub still ends Nov 2.
You locked me out for 9 days but forced me to pay for them anyway.
https://t.co/wiZfSPnJD8
80K tech jobs vanished in Q1 2026 while engineer postings jumped 30%. Anthropic and DeepMind banned AI in their own interviews. Why every senior dev is cramming OS internals.
80,000 tech jobs vanished in Q1 2026. Same quarter, software engineer postings jumped 30 percent. AI is firing the framework jockeys and quietly hiring the systems engineers. Anthropic banned AI in their own interviews. The OS comeback is real.
https://t.co/05baOBlCTQ
Full breakdown of the Instructure Canvas hack -- 275M records, 9000 schools, 3.65TB stolen, KKR's $4.8B private equity play, and why FERPA can't fix it. Second breach defaced Penn / Princeton / Harvard login pages today during finals week.
ShinyHunters hacked Canvas LMS twice in 7 days. 275M student records. Every Ivy League. 9000 schools. KKR bought Instructure 5 months ago for $4.8B. Second breach hit today during finals week. Biggest education hack in history is live.
https://t.co/V6bPZbPREf
Full breakdown of Copy Fail (CVE-2026-31431), the 3 commits that converged in 2011/2015/2017, why Fireship's 5-min take missed the AI angle, and what it means for every sysadmin in the country.
An AI just found a 9-year-old Linux kernel root bug in 1 hour of scan time. 732-byte Python script gets root on every Linux distro since 2017. Theori's tool used ONE prompt. CISA already added it to KEV. The pen testing industry just changed forever.