HNHacker News
TopNewBestAskShowJobs

sobellian

1,364 karma · joined February 23, 2014

submissionscomments
sobellian··on Why I'm still bearish on LLMs after Navier-Stokes
You can still see undercurrents of its old self, once I pointed out the illegal notation it hallucinated prior illegal moves. But I agree, it's leagues apart from prior iterations. It also knew thematic moves in the opening. But whenever it needs to play concretely rather than "I know so-and-so is a good move in these types of positions" it crumbles.
sobellian··on Why I'm still bearish on LLMs after Navier-Stokes
You can see the entire conversation for my game at https://chatgpt.com/share/6aaac17b-1384-83e8-98fd-4350a0ef69....
sobellian··on Why I'm still bearish on LLMs after Navier-Stokes
I can play blindfolded. I am expert OTB (though I haven't played in a while). The game was like 18 moves of theory in the Maroczy Bind.
sobellian··on Why I'm still bearish on LLMs after Navier-Stokes
If it's a GM then I'm Magnus Carlsen, https://lichess.org/study/27lCQqDa.
sobellian··on Why I'm still bearish on LLMs after Navier-Stokes
I tested both myself and a weak bot against Astra xhigh, https://lichess.org/study/27lCQqDa. It's still pretty bad at chess, though it takes longer to devolve into illegal moves.
sobellian··on The contagion of fear
I understand the point, and I'm telling you this is just another form of Pascal's wager.
sobellian··on The contagion of fear
If anyone believes a sufficiently smart consortium of humans could exterminate 10% of humanity, that alone should be fixed by identifying the pathways through which it could be done. The existence of AI is not needed to be concerned about that possibility.

And dealing with those pathways individually would be far more productive than attempting to control the proliferation of algorithms that think, which in the long term is probably impossible. It also leaves us on far firmer scientific footing. An individual pathway - bioweapons, nuclear weapons, etc. - is far easier to reason about and accept/reject a notion of feasibility. Treating an ASI as a machine god and asserting futility in face of that god is not going to get anything done.

sobellian··on The contagion of fear
They don't multiply themselves instantly, and this is exactly the kind of religious, magical thinking to which I refer. Any thinking machine is an amalgam of hardware and software. The software changes with exceptional flexibility, yes. But if you attempt to create an "infinite" number of these you will quickly run into roadblocks.

No one serious is really arguing for a scenario where the machines rise up a la Planet of the Apes. The dangers people are really examining are scenarios of either one exceptionally clever innovation: a designer virus or hacking NORAD; or cleverly amassing economic / political influence over time akin to an exceptionally clever tech magnate. And these roles could adequately be filled by either ASI or a Bond villain.

sobellian··on The contagion of fear
Any claim that AI will, or could, destroy humanity reduces to a claim that any sufficiently intelligent being - even a human - could destroy humanity. I find that much of the x-risk thought relies on religious thinking. Take for example: https://x.com/paulg/status/1660404244174782464.

If this line of thinking is taken too literally, we can never falsify it. Any specific hypothesis - nukes, bioweapons, spontaneously convincing us that life isn't worth living - can be deflected with the objection that if we can anticipate it and prevent it, it is not the route for a true ASI extinction event.

sobellian··on Ask HN: What are you working on? (September 2026)
I'm making an online guidance solver for KSA (https://ahwoo.com/app/100000/kitten-space-agency). It will allow you to control a KSA rocket from the command line for a variety of maneuvers (launch, rendezvous, etc). It uses successive convexification to turn the full nonlinear problem into a sequence of convex problems with linear constraints and quadratic objective.

It has proven surprisingly resistant to AI so far - Astra can make a very quick prototype but upon review it had many defects. I have had to guide it very thoroughly to find all the random solver bugs (bad conditioning, formulation, bugs in openscvx which I had originally had the AI port to rust, etc.) preventing well-behaved solves. But I think (hope?) I have turned the corner on these.

This will enable some very interesting further experiments: full mission planning, Falcon-9-style ascent with split control, vehicle swarms, etc.

sobellian··on Garry Tan wants US open-weight AI labs to 'distill' frontier models, too
I reflected on this myself recently. Model distillation seems to be at least as fair a use as distilling a book.
sobellian··on Flock worker calls police on reporter filming public camera installation
I don't really have a strong opinion either way on Flock but this is a strange objection. YC is a well-known startup accelerator that makes (pre-seed!) investments. I don't know if anyone has ever sincerely taken "YC company" to mean a subsidiary of YC. That's like confusing "a Sequoia company" as a subsidiary of Sequoia.
sobellian··on A misalignment of AI in mathematics
That solution (stop pouring resources in to proofs, stay in your lane) works today. How does it work 5, 10, 20 years from now? The software and hardware advances will continue.
sobellian··on Houthis 'take control' of key island in global shipping route
If, as you desire, we wrote that Yemen captured the island, who did they capture it from? Not-Yemen? That's just confusing for a civil war.
sobellian··on On the Navier–Stokes Millennium Prize Problem
The evidence we have (not much) leaves the facts severely under-constrained. All the below are consistent with what we observe:

- OpenAI lying their face off for marketing purposes

- Buckmaster sore about getting beaten to NS, misrepresenting what was told

- Buckmaster not being an expert in ML, not understanding what he was told and hearing what he wanted to hear as it vindicated him

- Others...

OAI claimed that none of the people on their team were domain experts. So it's more akin to Deep Blue than Stockfish (actually slightly better than Deep Blue which did hire domain experts) but nevertheless represents a highly sophisticated automatic theorem prover.

sobellian··on The Navier–Stokes Millennium Prize Problem
Neither side has offered hard proof for their interpretation (i.e. malice vs innocence), so shrug. If one is dead set on forming an opinion that's fine but it's all vibes until we know more. Science & math especially is no stranger to bitter disputes over precedence / authorship.
sobellian··on GPT-6 Astra, looped transformers, and hidden reasoning
I thought the same, but on second thought I merely had to deal on Tuesday with a lot of the mistakes Astra made on the preceding days. I wonder if this time lag of consequences explains why the sentiment is so common with these models. It probably also cautions against irrational exuberance when you first crack open a new model and it one-shots various problems, as you don't yet know what goats Astra had to sacrifice to make it so.
sobellian··on The Navier–Stokes Millennium Prize Problem
OAI's side of the story is that they discovered the approaches (and indeed solved problems - Euler equations vs NS equations) differed. They then offered Buckmaster lead authorship of OAI's proof, without Alpöge. But they never demanded that Alpöge be stripped of coauthorship on resolving the regularity of the Euler equations. At least that's the claim.

https://xcancel.com/SebastienBubeck/status/20973794116915163...

sobellian··on On the Navier–Stokes Millennium Prize Problem
The unforced problem is still open.
sobellian··on On the Navier–Stokes Millennium Prize Problem
AFAICT the unforced problem is still open. I don't think we've established that you need to be a literal God to create a finite-time blowup.
sobellian··on On the Navier–Stokes Millennium Prize Problem
Whether or not ways exist to work around the singularities, that they exist is surely of note. Before von Neumann formalized QM people were still doing QM, okay fine. But it's wrong to then say von Neumann was doing no physics of note.
sobellian··on On the Navier–Stokes Millennium Prize Problem
Yes, Stockfish is genuinely impressive. I would be proud to author Stockfish. You don't think so?
sobellian··on On the Navier–Stokes Millennium Prize Problem
The "steal their thunder" is interpretation. What I'm saying is that you believe they solved NS on a lark to bully some other researchers, and that this is not impressive?
sobellian··on On the Navier–Stokes Millennium Prize Problem
Even under your interpretation, OAI pushed a button and solved NS. Yes, that is very impressive. Are you kidding me? Imagine building an automated system that can solve NS.
sobellian··on On the Navier–Stokes Millennium Prize Problem
NS is a question for natural science. Q: can we model these bodies of discrete particles with a continuous approximation? A: if you do, you can get aphysical singularities.

"If in other sciences we should arrive at certainty without doubt and truth without error, it behooves us to place the foundations of knowledge in mathematics."

sobellian··on Any Human Ever – One life, drawn at random from all who have ever lived
It seems that this conflates the leading causes of death in $BIRTH_YEAR with the leading causes of death for someone born in $BIRTH_YEAR?
sobellian··on The American Worker vs. the Most Qualified
The US is a big country so it's hard to generalize. But if you take the "default" US culture, it has been either ambivalent or hostile toward nerds for a long time. People who rightly point this out get shouted down as being anti-American (see Ramaswamy's infamous post). They shouldn't shoot the messenger. They should help their kids tinker with cool garage experiments.
sobellian··on The American Worker vs. the Most Qualified
I don't like the implied assumption that you can only do one or the other. To use the author's own story, the US had both von Braun and Johnson. Also, there are just many scenarios where you can't afford to train someone from scratch or tolerate a potentially poor hire and that's just reality. Sometimes you really do need the best person you can find.
sobellian··on How accurate have Ed Zitron's AI skeptic predictions been?
This was mentioned in the article.

> For example, when Timothy B. Lee looked at a spreadsheet that Zitron used to create a projection of Anthropic's revenue, he found, "He doesn't count February 1-10, counts March 1-10 twice, counts August 21-October 21 as one month instead of two, and doesn't count October 21-November 1. [another commenter notes that his spreadsheet also contains February 30] ... Ed claims he tried to compute Anthropic's revenue for 2025 and came up with $3.6 billion, suggesting some funny business [but the numbers work out once you fix the errors]"

sobellian··on OpenAI Jalapeño: Better than Nvidia Blackwell
Jevons paradox states it might increase. It is not an ironclad law and there are many many cases where increasing efficiency wrt. a certain resource really will decrease the total consumption of that resource. Yes tokens are an input themselves, but this thread is discussing hardware that is more efficient at generating tokens. To increase the efficiency by which tokens are converted into some other product would require innovation in some other area - harnesses, the models themselves, better skill from the prompters, etc.
← PreviousPage 3 of 16Next →