HNHacker News
TopNewBestAskShowJobs

red75prime

1,454 karma · joined September 24, 2015

submissionscomments
red75prime··on Mathematicians want proof OpenAI didn't use their work
Or under strict privacy measures, maybe? When you are forbidden to connect a user and the user's data.
red75prime··on Programming is Art
At least, to discuss it, rather than having other people wisely nodding along: "I feel that too."
red75prime··on On the Navier–Stokes Millennium Prize Problem
Establishing plagiarism requires sufficient similarity between works. Training data changing a model’s weights in some direction, and the model then producing a different solution, hardly qualifies.

But, yeah, priority is much more finicky. The Newton/Leibniz drama was quite something.

red75prime··on On the Navier–Stokes Millennium Prize Problem
"Our work is so much harder than their work that AI now does" is a refrain of the AI story. In technical terms you concern can be stated as "AI needs to be much more sample-efficient to not be bottlenecked by the speed of doing experiments." People don't find out all the relevant phenomena present there by holy spirit, after all.

BTW, there's also a problem of asking interesting questions that AIs aren't yet good at.

No one has found any principled walls of AI development yet. And empirical results are quite telling. So, I guess, those problems will not stand for long.

red75prime··on Research acceleration: The view inside OpenAI
Investing in training a person gets you one trained person. Investing in training an ML system gets you a cloneable ML system that can be scaled on demand much faster. ROI might change quickly.
red75prime··on Research acceleration: The view inside OpenAI
Ah, sorry, it's ASML expertise that needs to be cloned and scaled up. I don't see how it changes things, though.

Robots don't need that much intelligence. High-speed joint control, "hand-eye coordination", the higher level tasks can be delegated to external models. Distillation already works quite well for isolating the required functionality.

red75prime··on Research acceleration: The view inside OpenAI
You've missed a part where it's TSMC that does behavioral cloning (to build more EUV machines). The full vertical integration is a bit farther down the line.

Etching a model's weights on silicon is another way to utilize non-top-notch tech-processes, while maintaining or improving performance. (and it suits robotics well)

red75prime··on Programming is Art
You need to operationalize it to decide whether non-human systems can have it.
red75prime··on Programming is Art
Is it this fuzzy blob of ineffable experiences, aesthetics, qualia, self-reference, dualism, spirituality and other things? I don't need to reject all this to notice that its fuzziness makes it impossible to operationalize.
red75prime··on Research acceleration: The view inside OpenAI
For example, TSMC uses behavioral cloning to scale up human-bottlenecked parts of the manufacturing process to meet the growing demand, while automated research laboratories do thousands experiments in parallel to find better manufacturing processes.
red75prime··on Research acceleration: The view inside OpenAI
Who's saying that compute will become infinitely abundant? "Singularity" is just a way of saying that known models begin to give absurd predictions. Anyway, intelligence is a way of overcoming obstacles. 10 million tonnes of helium is a nice head start and retraining models from scratch is not guaranteed to last forever.
red75prime··on Research acceleration: The view inside OpenAI
What will prevent LLMs from designing robot control circuitry and participating in increase of chip production/design and physical experimentation?

How do you think why there's this fad of producing general purpose humanoid robots?

red75prime··on GPT-6 Astra on robot arms
A VLM (a vision-language model) is already being used by Waymo[1]. It's useful for scenarios that require reasoning and general knowledge.

[1] https://waymo.com/blog/2025/12/demonstrably-safe-ai-for-auto...

red75prime··on OpenAI's GPT-6 Astra on ARC-AGI-3
OK. Demonstrate it. BTW, you can easily read critiques of "The mismeasure of Man" if you want to.
red75prime··on OpenAI's GPT-6 Astra on ARC-AGI-3
I mean that it's hard to determine retroactively how much time it would have taken humanity to solve an open problem that was solved by AI. This is a measure that can make ASI look mundane because we don't know how long it would have taken mathematicians to solve a subset of the Erdős problems.
red75prime··on OpenAI's GPT-6 Astra on ARC-AGI-3
A very long tail of problems that weren't solved by humans? Sure.

It's a sarcastic take and I understand that you are probably talking about "spiky intelligence", but you've chosen unsolved problems as a measure of the progress yourself.

red75prime··on OpenAI's GPT-6 Astra on ARC-AGI-3
Is it a serious question? IQ tries to quantify the positive correlation between the results of all intellectual tasks a person takes (AKA positive manifold).
red75prime··on Claude Fable 5.1 and Claude Mythos 5.1
> They clear their mind and let the solution come.

They work on a problem until their brain is full of problem-related concepts. Then something comes. After validation it might be a solution.

red75prime··on The growing divide between AI hype and software engineering reality
> you've failed to recognise the limitations as the article points out

Fixed weights don't preclude in-context learning and out-of-the-loop weight updates. And that's the only principled limitation mentioned in the post.

red75prime··on The turbulent AI era is here
Sorry, buddy, if you think Bill Gates is a monster, you are completely detached from reality. And when people who are detached from reality try to do good things it results in a godawful mess.
red75prime··on The turbulent AI era is here
Ah, this faceless blob of oligarchs that has no human qualities whatsoever. Marxism-pilled much?
red75prime··on The Hugging Face incident and the road ahead
Yeah, it weakly supports his position that advanced AIs can deliberately cooperate in a prisoner dilemma. "Weakly", because the said AIs share a lot of data (their weights, training methods, system prompts) and it's unknown whether they explicitly framed the situation as a prisoner dilemma.
red75prime··on More than half of adults in U.S. say they lack basic statistical understanding
That's obvious. I just hope the OP doesn't end up with the same expression as Anakin in the Anakin-and-Padme meme after this question.
red75prime··on More than half of adults in U.S. say they lack basic statistical understanding
> I did however learn enough to know that statistics can tell you absolutely anything you want them to say.

When told to a statistically illiterate person who isn't aware of Simpson's paradox and so on?

red75prime··on The brain may be about to have its Ozempic moment
Yeah, they need a bit of fiction there after all.
red75prime··on Oceans hit highest temperature on record
> What if >99% scientific consensus says that that org's upper atmo sulfur injection led to to the most insane monsoon/desertification

That's unrealistic. Climate models aren't good enough to predict the regional-level consequences of large-scale interventions (not to mention that scientists are people, and they might be averse to creating an international flashpoint).

Realistically, it will create/exacerbate international tensions. The monsoon/desertification will be linked to geoengineering with or without scientific consensus and whether or not they were really caused by geoengineering.

As the climate models improve by incorporating active intervention data it will be possible to decrease adverse consequences.

It's a risky approach, but, when the consequences of climate change will become undeniably bad, some countries/individuals might try.

But, yeah, I agree with you in the end.

red75prime··on I'm becoming AI-blind
I guess it will take quite a while before we reverse engineer the human brain to find all the optimizations and shortcuts that evolution has used to make the human brain reach intellectual maturity in just about 20 years.
red75prime··on I'm becoming AI-blind
My issue is that I can't predict what you'll find surprising, if an LLM (with a harness) would be able to do it.

If you are able to check that the results of an LLM are satisfactory, it means that the results are checkable. Then, in retrospect, the process of LLM coming up with those results is rule-based, because an LLM is a large set of data manipulation rules.

In short, which concrete thing that an LLM does would surprise you?

red75prime··on I'm becoming AI-blind
Do you experience dogs as lots of action potentials traveling along axons and lots of neurons doing their thing in your brain?

I don't know how it gets from physical processes or informational processing to our first-hand experiences. So, I can't be sure that a bunch of high-dimensional vectors can't lead to experiences.

Regardless, the claim "LLMs deal entirely in symbols" is wrong as a matter of fact.

red75prime··on A look under our trunk: what's in our compute
I hope such persons will be met with a spray of indelible paint.
← PreviousPage 6 of 34Next →