Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

DeepMind and OpenAI win gold at ICPC

codeforces.com

Post reply on HN
Print view View original

241–250 of 255 posts

Re: DeepMind and OpenAI win gold at ICPC

#241

Post by notemap »

seanmmward wrote: post #234
JohnKemeny wrote: post #7

I went to ICPC's web pages, downloaded the first problem (problem A) and gave it to GPT-5, asking it for code to solve it (stating it was a problem from a recent competitive programming contest). It thought for 7m 53s and gave as reply # placeholder # (No solution provided)

If you want to see some example solutions (not yet validated via official judging): https://github.com/iGentAI/icpc-maestro-solutions-2025

Don't trust it. Problem B required an algorithm to run for inputs up to N ~ 200, and a clever graph theory lemma, before succumbing to pattern matching / law of larger numbers. Claiming that there's a pattern for N EDIT: Just submitted it, WA. Yeah.

Problems are hard enough where consumer models can't solve all 12 problems.

Re: DeepMind and OpenAI win gold at ICPC

#242

Post by j_timberlake »

asboans wrote: post #104
amluto wrote: post #59

I've contemplated this a bit, and I think I have a bit of an unconventional take: First, this is really impressive. Second, with that out of the way, these models are not playing the same game as the human contestants, in at least two major regards. First, and quite obviously, they have massive amounts of compute power, which is kind of like giving a human team a week instead of five hours. But the models that are co…

Firstly, automobiles are really impressive. Second, with that out the way, these cars are not playing the same game as horses… first, and quite obviously they have massive amounts of horsepower, which is kind of like giving a team of horses… many more horses. But also cars have an absolutely massive fuel capacity. Petrol is such an efficient store of chemical energy compared to hay and cars can store gallons of it. I…

This response is good but the more general problem is that people are in "It doesn't look like anything to me" mode like Westworld robots seeing advanced technology. If there's a way to snap people out of that, I've never seen it.

Re: DeepMind and OpenAI win gold at ICPC

#243

Post by modeless »

LunaSea wrote: post #214
modeless wrote: post #74

Earlier quoted context omitted.

Human brains and cloud instances are not remotely equivalent. What you can compare on an equivalent basis is cost.

Or energy usage.

Cost is inclusive of energy and a lot more. It's a better metric.

Re: DeepMind and OpenAI win gold at ICPC

#244

Post by pedrosorio »

LunaSea wrote: post #227
matheusd wrote: post #220

Earlier quoted context omitted.

> Humans are more efficient watt for watt than any AI ever invented. Indeed they are. For now. The long term trend is not in our favor.

I disagree, the long term trend is that we do not have the available electricity and water for this to continue for much longer.

Regarding electricity, it depends on what you mean by “we”, I guess

https://www.voronoiapp.com/energy/-China-Generated-More-Elec...

Re: DeepMind and OpenAI win gold at ICPC

#245

Post by in-silico »

thrwaway55 wrote: post #181
in-silico wrote: post #156

Earlier quoted context omitted.

The point is that up until now, humans were the best at these competitions, just like horses were the best at racing up until cars came around. The other commenter is pointing out how ridiculous it would be for someone to downplay the performance of cars because they did it differently from horses. It doesn't matter if they did it using different methods, that fact that the final outcome was better had world-changing…

I feel the main difference is cars can't compress time in the way an array of computers can. I could win this competition with an infinitely parallel array of random characters typed by infinite monkeys on infinite typewriters instantly since one of them would be perfectly right given infinite submissions. When I make my tweet I would pick a single monkey cus I need infinite money to feed my infinite workforce and th…

Considering that OpenAI's model got a higher score than any of the world's best collegiate programming teams, I'd guess that a mechanical turk would not do better (even if you gave them quite a bit of time).

Re: DeepMind and OpenAI win gold at ICPC

#246

Post by fullparens »

I briefly looked at a few of Gemini's solutions https://github.com/google-deepmind/gemini_icpc2025. What struck me was how Gemini finds clean ways to express an idea - perhaps because it knows a large set of tricks for each kind of sub-algorithm (within the larger algorithm). I am a former competitor and managed to reach world finals at Google CodeJam and Topcoder Open. It took me a lot of work to get there but I will gladly concede that Gemini is way better than my peak. I haven't competed in 15 years and have forgotten a lot of tricks but Gemini's code reminded me how quickly algorithms can get complicated sometimes without a bag of tricks.

There are parallels to tactics in chess - humans might miss them but a machine will not. And that can be a huge difference in a game or even in a software project.

EDIT: minor correction.

Re: DeepMind and OpenAI win gold at ICPC

#247

Post by AndrewKemendo »

smokel wrote: post #224
NoahZuniga wrote: post #143

Earlier quoted context omitted.

Actually collegiate means that the contestants are in college.

Ha, shows what I know :) As a non-native speaker I had always assumed it referred to working with colleagues. Etymologically related, but apparently not the same.

Clearly strongly held Hallucinations are problematic no matter what agent produces them

Re: DeepMind and OpenAI win gold at ICPC

#248

Post by chpatrick »

raspasov wrote: post #239
chpatrick wrote: post #222

Earlier quoted context omitted.

It's not new scientific ground but a machine beating a challenging computer science problem unassisted is a big deal. If they can do that then there are a lot of other challenging things they can do.

Like what exactly? As far as I can tell, the drug discovery is fizzling out, so it's not talked about much. Toxicity, for one, is a big problem, and the AI is not going to tell you whether the new drug it just concocted is suitable for humans or not. Small model solves an easy problem; big model solves a challenging problem. I wouldn't call those problems; they are more like invented puzzles. Perfect match for the AI…

Whatever you say.

Re: DeepMind and OpenAI win gold at ICPC

#249

Post by JohnKemeny »

HarHarVeryFunny wrote: post #57

ICPC = The International Collegiate Programming Contest. These are college level programmers, not elite competitive programmers. Apparently Gemini solved one problem (running on who knows what kind of cluster) by burning 30 min of "thinking" time on it, and at a cost that Google have declined to provide. According to one prior competition paricipant, writing in the comments section of this ArsClasica coverage, each y…

> ICPC = The International Collegiate Programming Contest. These are college level programmers, not elite competitive programmers.

ICPC finalists are very much in the world elite of competitive programmers.

Re: DeepMind and OpenAI win gold at ICPC

#250

Post by JohnKemeny »

chpatrick wrote: post #137
bgwalter wrote: post #122

Earlier quoted context omitted.

"Database" was not meant in a literal sense. Clearly a lot of knowledge from similar problems is encoded in the model, that is why you can use models as a kind of fuzzy encyclopedia. It is like an open book exam for humans where they also can lookup similar problems. The current top comment makes the same point, but in a more diplomatic and sophisticated manner.

I mean strong human contestants would also know a lot of similar problems, I'm not seeing how it's fundamentally different or not a meaningful achievement.

Do you also think that IBM Watson won Jeopardy primarily because of its ability to reason?

Or due to its massive internal database combined with superhuman button pressing skills?

Post reply on HN