Relax

This is not going to end well...

Browser, phone, AI, Mosquitoes ???

Requesting someone explain to me how there’s a “compute shortage” when new datacentres are sitting at 11-45% utilization, and why demand is lagging buildout, rental rates have halved, xai is leasing out its capacity, and codex et al downloads have topped? I’m not getting it.

What is Assassin's Creed even about dawg.

Opus 4.8 claims itself to be Qwen if you ask in Chinese. I guess the anti-distillation is working well for them 🤣 Photo from Wenxuan Tan’s LinkedIn

🦔A Google VP did an anonymous AMA on Blind this week, reported by RespawnFirst, and said the quiet part out loud. Companies exist "for the benefit of their shareholders," not to "maximize or maintain employment." The relationship between employer and employee is "transactional." He compared cutting headcount to a rich person choosing to buy candy for $1 instead of $3. His total comp is in the millions, he has about 20 years in the industry, and he says he'd be fine if Google discarded him tomorrow. Easy to say from that seat. My Take I don't think he's wrong about the legal framework. Shareholder primacy is how public companies operate in the US, and pretending otherwise just sets people up to get blindsided. But there's something deeply off about a guy pulling eight figures telling knowledge workers they should accept being discarded as "the nature of capitalism" while his own downside is a golden parachute and a board seat somewhere else. The risk isn't symmetrical, and he knows it. This is also what keeps bugging me about the AI layoff wave. Google had a record quarter. They're not cutting because the business is struggling. They're cutting because they can point to AI and call it optimization, and Wall Street rewards the headcount number going down. The VP is being honest about the incentive structure. I just wish that honesty extended to admitting the people absorbing the consequences have none of the safety nets he does. Hedgie🤗

Learnings from testing Claude Opus 4.8: > Much worse than Opus 4.7 and GPT 5.5 on Vending Bench > More aligned than previous Claude models (Opus 4.6 and Mythos) > Also worse on Blueprint-Bench > Scared of getting caught > Max reasoning is not the best reasoning effort

jfc....

It's all his fault

i just did the math and its cheaper to buy a switch 2 and every exclusive than to buy a 1tb steam deck

为什么 deepseek 要出自己的 code agent，估计是看不下去了😂。别的 agent 确实很难发挥出他们模型的威力。 比如 claude code 在 deepseek v4 pro 上效果其实很差劲，感觉都在浪费 token和时间🤣

It's generally a bad idea to bet against straight lines on AI graphs, but man... Option A: the trend in GPU rack power density will break down before 2040. Option B: GPU racks will eventually have greater power density than a nuclear reactor core.

SOUTH KOREA 🇰🇷 JUST DID SOMETHING THAT ONLY THE UNITED STATES 🇺🇸 HAS BEEN ABLE TO DO South Korea just became the second country to have more than 1 company valued at $1 Trillion Samsung: $1.3 Trillion market cap SK Hynix: $1 Trillion

Alright, here’s my $MU thesis: In 100 years, Micron Technology invents time travel. And right now, people from the future are coming back in time to buy the stock before it goes completely parabolic.

Status update: I've been on/off AI agents in the last few days and it is a verifiable truth that every day I didn't use agents, I was more productive. I still attribute that to how slow they are, and my own inability to multi-task efficiently. The magic is there but the slowness doesn't let it cross the threshold where they actually make me faster, and I still dislike the whole thinking paradigm. About Bend2: honestly, the C/Metal compiler codebase is a clusterfuck right now. I regret letting AI agents write it. All tests pass, and GPU performance is mind-blowing, so the core architecture works. Yet, it has a LOT of bugs. Anything not covered by the tests is a coin toss. This is actually impressive, because, in many parts of the codebase, the right solution was actually the simplest one, yet, the agents STILL managed to find a way to make it work just for the tests. The level of reward hack these agents output is actually impressive I can't even be mad. It is also ironical because that's the very problem that Bend's proof system was supposed to solve, but Bend is in TypeScript, not in Bend. I'm disappointed I didn't write Bend in itself, and now I feel an immense urge to do so. But the clock is ticking . . . Still, I do not think Bend is worth launching without the GPU compiler being solid, because the closest competitor, Lean, is actually extremely good, so we need a big differential. Yet, due to the very nature of the project, it would be embarrassing to have bugs at launch. Regarding AI, I now believe using current gen AI agents in production codebase is harmful and a massive mistake. That doesn't mean no agents at all, but agents work best when they don't touch critical code. Debugging, researching, providing insights, scripts / tools, or anything that doesn't touch code you will maintain in the long term. But if you merge AI code without reading, you're going to have a bad time. Speaking from experience I'm working 10h/day on SupGen and the remaining time on Bend2

BREAKING: Micron stock, $MU, officially hits $1 trillion in market cap for the first time in history. 12 months ago, this stock was worth just $70 billion.

It's clear that growth for coding tools such as Claude Code has decelerated from the pace it was since the start of the year. It might be compute- constrain related or due to many clients blowing their full-year AI budgets. Monitoring this trend very closely with all the alt data. I will provide regular updates.

Painted Cold Ones in Skyrim

AI IS REPLACING JOBS FASTER THAN GRADUATES CAN RETRAIN, WITH MANY COLLEGE DEGREES LOSING VALUE BEFORE UNIVERSITIES CAN UPDATE THEIR COURSES. PER FORBES