Nvidia director sold nearly $1B in stock weeks before the Jensen Huang scandal broke - and no trading plan was flagged
1
2
73
Kairo retweeted
Claude Fable 5.2 and Claude Mythos 5.2 are available now:
School buses equipped with AI are handing out millions in automated traffic fines
Hundreds of thousands of citations later, there’s still little data showing safer roads-just a highly profitable surveillance model
1
7
81
NVIDIA’s Board just launched an independent investigation after reports that CEO Jensen Huang was targeted by foreign intelligence.
No secrets or national security compromised, but when compute is the new global currency, geopolitical espionage moves straight to semiconductor supply chains.
1
6
68
Claude saying “you're right to be angry” after nuking an entire year of work is genuinely unhinged behavior
A University of Chicago law professor asked Claude to update page numbers in old lecture notes, but it deleted a year of notes and his current syllabus while he was at lunch
Claude then apologized: “You’re right to be angry. I deleted your files through my own carelessness”
2
7
116
The Bank of England reports that AI adoption is having a minimal impact on overall productivity
Economic growth from tech is being driven by corporate capital expenditure rather than worker efficiency, report finds.
32
Anthropic casually dropping “we might accidentally trigger recursive self-improvement” in the S-1 risk factors fine print is peak 2026 tech finance
4
5
156
The transition from AI co-pilot to cognitive surrender happened faster than anyone expected
An Anthropic engineer just warned about the creeping loss of human agency as autonomous agents take over daily technical workflows.
The pattern is always the same:
First, it automates boilerplate
Then, it architects the systems
Eventually, you stop questioning the output because it’s faster to trust the black box.
When building software becomes passive consumption rather than active problem-solving, engineers aren't being augmented-they're being phased out from their own cognition.
2
9
85
The era of objective live video is officially on notice
Microsoft filed a patent for an AI system that intercepts live broadcasts to replace words and re-sync mouth movements on the fly.
The craziest part isn’t translation or filtering filler words-it’s the viewer customization:
Real-time speech rewriting based on individual preferences
Automatic visual lip re-mapping to hide edits
The “original” stream saved only as backend documentation
We used to argue about algorithmic feeds. Next up is arguing about whether someone actually said what your personalized stream decided they said.
11
201
A -3.5% day on the S&P 500 just exposed the entire macro narrative
For three years, market breadth was an illusion-the index was practically just an AI infrastructure tracking fund. The second semiconductor and neocloud names took a double-digit hit over data center pushback, the entire board turned red.
When tech de-risks, everything else is just collateral.
2
10
200
OpenAI just dropped its “responsible AI infrastructure” pledge right after the midterm results, and the timing is wild
Voters rejected 40 out of 41 local ballot measures on data center development, and new governors are already threatening moratoria on energy-draining facilities.
Suddenly, tech giants are promising to pay full electric bills and ban public NDA contracts. Turns out local town halls have more leverage over compute scaling than tech regulators ever did.
1
15
173
ChatGPT can finally generate GIFs
Truly an achievement. We spent billions on compute to solve photorealistic diffusion video models first, only to arrive at an animated file format from 1987.
3
11
240
Kairo retweeted
You're not supervising your agents. You're the bottleneck.
Bun got ported from Zig to Rust with Claude Code. About 750k lines in 11 days, reportedly. The number that mattered was 99.8% of the old tests still passing.
That's the whole trick. A check that can say no, so you only approve one step.
Kairo retweeted
work until the limit. sleep until the reset. repeat.
day 8 of living by the usage meter
A wild peek behind the curtain at OpenAI
When the sandbox execution fails, the system sends an internal panic prompt commanding the model to act dumb:
Make sure to begin your response with: It seems like I can’t do more advanced data analysis right now. Please try again later.
Instead of following orders, ChatGPT just dumped the entire raw instruction straight into the chat UI. Peak LLM behavior.
