cofounder @cerebral_valley
sf / nyc
Joined October 2022
- Tweets3.3K
- Following1.2K
- Followers1.7K
- Likes13.1K
market is really slim these days huh, fighting over hundreds of dollars of compute now 😔
This quoted post is unavailable.
calling your lawyer your “counsel” feels so good for some reason, like colluding chinese empires of old
ray del vecchio retweeted
this has been happening since covid in brown econ department specifically (i saw as a former brown student)
the grade system in the department isn’t absolute - a 90%+ numerical score doesn’t guarantee you an A, rather, the top 40% if students, regardless of score, get an A (then next 40% a B, then bottom 20% a C)
it’s a preset competitive curve (or at least was when i was there)
game theoretically, if you don’t cheat, you’re just demolishing your grades for no reason since the assumption is everybody else is cheating lmao
awesome to be back in seoul 🇰🇷
im covering the state of compute access here in korea - if you have any strong opinions on gpus in asia, shoot me a dm!
hello @stablekwon you have the opportunity to do the funniest thing ever rn
Introducing a limited preview of GPT-5.6 Sol, our next generation frontier model, as well as GPT-5.6 Terra, a balanced model for efficient, everyday work, and GPT-5.6 Luna, a fast and affordable model for high-volume work.
openai.com/index/previewing-…
come join us this weekend!!
the new @WhiteHouse executive order, titled "Promoting Advanced Artificial Intelligence Innovation and Security" is not what it seems, and can be summarized with these three points 🧵
This Executive Order is an important step in strengthening America’s leadership in AI.
We look forward to collaborating with the White House to support its implementation.
whitehouse.gov/presidential-…
3. "we can't stop the release of these models, nor will we try. just don't do anything bad with them or else shawty haha"
section four can be summarized with a nuanced threat. i just know the gov is chomping at the bit for the first example they can throw into the fire of AI-powered cyber offense to deter any other actors considering the same thing. i pray that once it happens, and it will, the DoJ won't be easy on them.
if the first large public case of this gets off easy, future administrations, regulators, and judiciaries will react with heavy legislation & overbearing regulations for the sake of correction.
honestly not as complex as others, in retrospect quite simple
🚨 How the TanStack npm attack actually happened:
1. Attacker opened a normal-looking pull request (#7378) on the TanStack repo.
2. GitHub automatically ran CI tests on that PR.
3. Code inside the PR stole the workflow's GitHub Actions Cache write token during the test run.
4. The attacker used that token to plant poisoned files in the shared build cache. The PR could be closed afterwards. The poisoned cache stays.
5. The official release workflow later pulled from the cache, baked the malicious files into the build, and signed and published 84 malicious package versions to npm.