Voted “Most Likely To Appear on Wikipedia’s ‘List of Largest Trading Losses’” among graduating class
- Tweets 72,957
- Following 518
- Followers 98,547
- Likes 236
This article has to bury the lede behind 20,000 words of carefully marshaled evidence because if it actually lead with its accusation it would be instantly dismissed as a deranged conspiracy:
14
17
726
89,232
Hey guys check out my new startup, Wolf. We provide a cutting edge model to perform verifiable and deterministic output categorization for structured input data with a non-Transformer architecture at incredible speed. Best of all, output tokens are totally free!
3
1
68
7,129
Quantіan retweeted
Explaining to TradFi guys I meet at conferences that insulting Dear Leader is to North Koreans what garlic is to a vampire is an endlessly gleeful factoid to me
That a group could be amongst the most sophisticated threat actors on the planet, yet have such a simple weakness, is something most people just couldn't conceive of on their own. Their reactions to learning about it are an endless source of entertainment
Thinking about how I’d explain to my 18 year old self how our libertarian streak somehow results in North Korea becoming the most omnipresent occupational risk that we face on a day-to-day basis
73
324
7,560
1,618,998
I had my first real “feel the AGI” moment today with Fable 5.1—it casually mentioned a trivial but niche fact about a firm employee it deduced from an unrelated internal database that it didn’t need to query for the problem it was solving, in a way a human definitely wouldn’t.
13
6
609
58,512
You know how a market crash is a great time to unwind a Ponzi scheme because you can go “oops we lost all your money trading just like everybody else”? Now is a great time to unwind your AI startup scam with “oops we had a security incident and are pausing like everybody else”
ssi are delayed after a catastrophic security incident discovered after seeing what happened with openai, order of magnitude more damage caused.
12
34
664
41,053
Quantіan retweeted
He’s saying AI labs are Goodhart-ing mathematics but you midwits are too stupid to understand.
To put it in simple terms: millennium prize problems are *metrics* for mathematics progress. AI labs are *optimizing* the metrics and thereby making them less useful.
Get it now?
104
258
3,197
206,204
Yeah I already have non sentient life forms I can interface with via my computer to get tough restaurant reservations, they’re called brokers
An aspect of the AI discussion I feel very disengaged from is how much people seemed focused on AI being able to book restaurant reservations, something I simply never do
6
4
300
17,506
Quantіan retweeted
Mathematics has now fully morphed into biology:
Project costs millions $ ✅
Put out paper that nobody has read ✅
Massive paper supplement nobody will read ✅
Advertise with pretty art ✅
Vicious authorship fight ✅
Say you're curing cancer ✅
45
638
7,874
229,674
Get me borrow on every share Anthropic has ever issued this shitco fraud is going to zero
It's fun to make predictions.
Here's a new one: Anthropic has cured a major disease
And I'll be even more specific.
Claude has cured Alzheimer's.
It is out for expert review.
And to give myself a hard deadline, they will announce it before the IPO.
6
5
336
52,499
Oh ok, well if Sam Altman says he was intimately involved with the entire project and personally vouches that they acted with integrity then that settles things. Case closed.
I spent much of the weekend talking with the team who did this work. Seb--and everyone else--acted with integrity and generosity throughout.
Initially we believed the other team had also solved the problem. We wanted to collaborate and do a joint release.
When we learned that they had Euler but not Navier-Stokes, we offered to let them go first, to suggest that they should be the ones to get the prize, and optionally for Tristan to be the lead author on a rewrite of the OpenAI proof. We felt it was challenging to offer the same to Levent (an Anthropic employee), who was not willing to talk or coordinate with us anyway. We were open to other solutions.
We would have greatly preferred coordination. We did not rush to publish even though the other team wasn't communicating with us. The team threatened us with unfounded accusations of plagarism.
Now that we can see their work, the approaches appear to be different. It is also worth noting that our latest model can solve many, many other math problems.
It is true that we tried this because there were rumors on the internet last week that Anthropic's models had solved a millennium problem and we were curious if ours could do it too.
17
69
2,622
52,915
We like complaining about Anthropic on here but it’s important to keep in mind that OpenAI is a snake pit of lying scumbags that everyone is tripping over themselves to leave for a reason, and this kind of behavior is why:
25
136
1,834
73,914
Sam would strangle a child on national television if he could get to claim Astra solved a Milennium problem afterward, playing peek-a-boo with one free user’s reasoning traces is table stakes especially because it’s literally in the TOS already
fwiw I think it is _extremely_ unlikely that user data had any influence here - there is no way OAI would pull user transcripts for this, or knowingly train on it in a way that would've influenced this.
I think its pretty important people don't run away with 'your user data isn't safe in codex' - because it surely is (based on everything I can assume from the outside)
5
11
392
17,420
The current OAI party line is explicit denial anyone saw the chats, followed by a careful non-denial denial about training, so their argument is “We checked and we did RL on those chats, but you can’t prove it wasn’t the $10mm+ of compute that cracked it”
Two things to distinguish:
Did any human or agent look at user data as part of the Navier Stokes effort? No.
Do we use user feedback and de-identified data to improve ChatGPT and Codex in a holistic way? Yes. And so does every LLM company.
3
4
95
5,835
