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

OpenAI O3-Mini

openai.com

Post reply on HN
Print view View original

851–860 of 944 posts

Re: OpenAI O3-Mini

#851

Post by ZeroTalent »

0xFEE1DEAD wrote: post #794
moonsu wrote: post #770

Earlier quoted context omitted.

This is the mini version which is not as good as o1 and I don’t think they demoed in the o3 announcement. I’m hoping the full release will be impressive

I know this isn't the full o3 release, but I find it odd that they're branding it as o3 when it feels more like an update to o1 mini. Yes, reasoning has improved, but the overall results haven't advanced as much as one would expect from a major version update. It's highly unusual for OpenAI to release a milestone version like this - it feels more like a marketing move than a genuine upgrade. Who knows what's going on…

o3 is absolutely fantastic compared to r1 what are these people talking about

Re: OpenAI O3-Mini

#852

Post by ZeroTalent »

anotherpaulg wrote: post #503

For AI coding, o3-mini scored similarly to o1 at 10X less cost on the aider polyglot benchmark [0]. This comparison was with both models using high reasoning effort. o3-mini with medium effort scored in between R1 and Sonnet. 62% $186 o1 high 60% $18 o3-mini high 57% $5 DeepSeek R1 54% $9 o3-mini medium 52% $14 Sonnet 48% $0 DeepSeek V3 [0] https://aider.chat/docs/leaderboards/

Also Gemini API is free for coding.

Re: OpenAI O3-Mini

#853

Post by JohnnyMarcone »

0xFEE1DEAD wrote: post #794
moonsu wrote: post #770

Earlier quoted context omitted.

This is the mini version which is not as good as o1 and I don’t think they demoed in the o3 announcement. I’m hoping the full release will be impressive

I know this isn't the full o3 release, but I find it odd that they're branding it as o3 when it feels more like an update to o1 mini. Yes, reasoning has improved, but the overall results haven't advanced as much as one would expect from a major version update. It's highly unusual for OpenAI to release a milestone version like this - it feels more like a marketing move than a genuine upgrade. Who knows what's going on…

This release was planned well before the Deepseek hype. When they demoed o3, they said they were planning on releasing o3-mini at the end of January.

Re: OpenAI O3-Mini

#854

Post by martin-t »

WhitneyLand wrote: post #847
martin-t wrote: post #824

Earlier quoted context omitted.

Thank you, this is a perfect argument why LLMs are not AI but just statistical models. The original is so overrepresented in the training data that even though they notice this riddle is different, they regress to the statistically more likely solution over the course of generating the response. For example, I tried the first one with Claude and in its 4th step, it said: > This is safe because the wolf won't eat the…

The problem with claims like these that models are not doing “actual reasoning” is that they are often hot takes and not thought through very well. For example, since reasoning doesn’t yet have any consensus definition that can be applied as a yes/no test - you have to explain what you specifically mean by it, or else the claim is hollow. Clarify your definition, give a concrete example under that definition of somet…

I am not gonna pretend to have one true definition but there are hopefully components of reasoning we can agree on:

- The ability to hold a collection of beliefs/statements and perform operations on them such as noticing one contradicts the others. Claude in my example clearly failed this.

- The ability to make logical inferences. Since logical steps can be represented by text, current LLMs seem to simulate this on the token stream and seem to have decent results most of the time. True inferences would be representing statements in some abstract form and being able to perform operations on them according to rules. The key here is that once it has a collection of statements and is given ("decides to use") an operation to perform, then it can do this correctly 100% of the time (ignoring hardware errors).

Re: OpenAI O3-Mini

#855

Post by ruibiks »

maddmann wrote: post #774
ruibiks wrote: post #733

Earlier quoted context omitted.

You are a maddmann! No bot there, ask me a question a bot would not figure out bro ;)

How many r’s in strawberry sherbet?

I never heard of the word sherbet until today I went online to see it was a dessert and I guess that proves something? ohh sh*t...bots can go online too... am I a bot? Could I be AGI?

Re: OpenAI O3-Mini

#856

Post by yodot555 »

martin-t wrote: post #824
sixhobbits wrote: post #803

For years I've been asking all the models this mixed up version of the classic riddle and they 99% of the time get it wrong and insist on taking the goat across first. Even the other reasoning models would reason about how it was wrong, figure out the answer, and then still conclude goat. o3-mini is the first one to get it right for me. Transcript: Me: I have a wolf, a goat, and a cabbage and a boat. I want to get th…

Thank you, this is a perfect argument why LLMs are not AI but just statistical models. The original is so overrepresented in the training data that even though they notice this riddle is different, they regress to the statistically more likely solution over the course of generating the response. For example, I tried the first one with Claude and in its 4th step, it said: > This is safe because the wolf won't eat the…

This is a dumb argument. Humans frequently fall for the same tricks, are they not "intelligent"? All intelligence is ultimately based on some sort of statistical models, some represented in neurons, some represented in matrices.

Re: OpenAI O3-Mini

#857

Post by jumploops »

inciampati wrote: post #807
jumploops wrote: post #541

Earlier quoted context omitted.

LLMs are the new compilers. As a student, you should continue to focus on fundamentals, but also adapt LLMs into your workflow where you can. Skip writing the assembly (now curly braces and semicolons), and focus on what the software you’re building actually does, who it serves, and how it works. Programming is both changing a lot, and not at all. The mechanics may look different, but the purpose is still the same: e…

LLMs are actually the new computers. Compilation is only one program they can run.

LLMs are the way computers were always supposed to work!

Re: OpenAI O3-Mini

#859

Post by waynecochran »

smallerize wrote: post #615
waynecochran wrote: post #576

I just had it convert Swift code to Kotlin and was surprised at how the comment was translated. It "knew" the author of the paper and what is was doing!? That is wild. Swift: // // Double Reflection Algorithm from Table I (page 7) // in Section 4 of https://tinyurl.com/yft2674p // for i in 1 .. Kotlin: // Use the Double Reflection Algorithm (from Wang et al.) to compute subsequent frames. for (i in 1 until N) { val X…

Wow, haven't seen a viglink in a while.

Not sure I meant that to happen. I am sure I just went to tinyurl to make the link -- I don't know how it ended up this way.

Re: OpenAI O3-Mini

#860

Post by waynecochran »

Sharlin wrote: post #696
waynecochran wrote: post #576

I just had it convert Swift code to Kotlin and was surprised at how the comment was translated. It "knew" the author of the paper and what is was doing!? That is wild. Swift: // // Double Reflection Algorithm from Table I (page 7) // in Section 4 of https://tinyurl.com/yft2674p // for i in 1 .. Kotlin: // Use the Double Reflection Algorithm (from Wang et al.) to compute subsequent frames. for (i in 1 until N) { val X…

Well, of course it knew the author. I'm sure you can ask just about any LLM who the author of the DRA is and it will answer Wang et al. without even having to google or follow the tinyurl link. And certainly it would also know that the algorithm is supposed to compute rotation minimizing frames.

I don't know if you are being sarcastic or not, but that is pretty cool if true.
Post reply on HN