Anthropic reported second-quarter profits alongside OpenAI's Erdős problem solution.
can someone check on gary
> reports of Anthropic posting profits in Q2 > OpenAI general purpose model solves long-standing and well-known Erdös problem without scaffold > OpenAI valuation up 67% since start of the year > Anthropic valuation up 173% since start of the year
Positive users highlight impressive AI results and progress from OpenAI and Anthropic while negative users insult promoters and accuse Gary Marcus of failing to properly check the math behind the headlines.
No Digg Deeper questions have been answered for this story yet.
Checking the math behind the latest headlines from OpenAI and Anthropic, link below:
can’t believe people assume that success on highly verifiable problems in math (where we don’t even know how many tests were performed and how many might have failed) iautomatically generalize to everything else when there is not a shred of evidence that they do. 🤷♂️
Can't believe we're doubling down on "it's just a stupid stochastic parrot" when OpenAI disproved a longstanding conjecture in discrete geometry less than 24 hours ago.
We need more Bores's, and less of whatever this is.
all these dudes posting about anthropic’s profits without checking the fine print. (see my earlier tweet)
there’s some fine print i believe that n the openai result too, and a lot that has not been disclosed
the rest is just about hype
can someone check on gary
> reports of Anthropic posting profits in Q2 > OpenAI general purpose model solves long-standing and well-known Erdös problem without scaffold > OpenAI valuation up 67% since start of the year > Anthropic valuation up 173% since start of the year
https://open.substack.com/pub/garymarcus/p/checking-the-math-behind-openai-and?r=8tdk6&utm_medium=ios
Checking the math behind the latest headlines from OpenAI and Anthropic, link below:
Gary is the undefeated and undisputed mental rationalization gymnastics world champion. He'll be fine.
can someone check on gary
> reports of Anthropic posting profits in Q2 > OpenAI general purpose model solves long-standing and well-known Erdös problem without scaffold > OpenAI valuation up 67% since start of the year > Anthropic valuation up 173% since start of the year
*valuations according to ventuals
can someone check on gary
> reports of Anthropic posting profits in Q2 > OpenAI general purpose model solves long-standing and well-known Erdös problem without scaffold > OpenAI valuation up 67% since start of the year > Anthropic valuation up 173% since start of the year
@scaling01 i am fine thanks
can someone check on gary
> reports of Anthropic posting profits in Q2 > OpenAI general purpose model solves long-standing and well-known Erdös problem without scaffold > OpenAI valuation up 67% since start of the year > Anthropic valuation up 173% since start of the year
Gary is a national treasure. We don’t deserve him.
Gary is the undefeated and undisputed mental rationalization gymnastics world champion. He'll be fine.
@scaling01 will post what you missed in my substack in an hour or so 🤣
can someone check on gary
> reports of Anthropic posting profits in Q2 > OpenAI general purpose model solves long-standing and well-known Erdös problem without scaffold > OpenAI valuation up 67% since start of the year > Anthropic valuation up 173% since start of the year
@GaryMarcus Why did you put the link below and not in the original post?
On the denominator question, I guess that’s fair in principle, but Thomas Bloom’s companion paper directly addresses this. He notes the model succeeded by systematically exploring directions humans dismissed as unpromising, not by brute-force guessing. We know the technique itself was novel.
On generalization outside math, I agree we should be cautious, and so does Cal Newport (who wrote about this). And he compared this to CAD software enabling Frank Gehry’s architectural designs. He sees this as a way to augment human creativity (in other words, useful). That framing is miles from “stochastic parrot,” which is what I was pushing back on.
Also, yes, I’m aware this was an OpenAI model, not Claude. I brought up Anthropic’s growth as evidence against the broader “AI bubble” thesis in the article I quote-tweeted.
@GaryMarcus @SohamThoughts For unit distance conjecture, humans have already tried to solve it many many times.
@scaling01 Let's not give more attention to this clown/fool.
@FrankRundatz @CoreyLeander Not a college student, Frankie. But does this matter? Does it answer any of the always concerned Gary Marcus concerns?
longer comment on yesterday’s stories
@CoreyLeander @GaryMarcus Anyone that questions why these metrics matter has not worked in the real world.
If that applies to a professor, it applies to a professor.
I think of Gary's posts like I think of Jim Cramer's stock picks. I just go the opposite direction. The underlying message on the value of neurosymbolic AI is great, but everytime I see a post that LLMs are a failure, it loses all credibility for me. The world needs neurosymbolic AI and LLM innovations. The advancements in LLMs, particularly reasoning, are making agent harnesses actually work now.
@Martin778192374 which? broad generalization has been my goalpost since 1998.
@GaryMarcus Benchmarks are legacy software. By 2027, the only metric that will matter for LLM valuation is agentic completion rate on autonomous dev tasks.
@FrankRundatz @GaryMarcus @grok do these metrics “matter” in this context?