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

791–800 of 944 posts

Re: OpenAI O3-Mini

#791

Post by cle »

IOUnix wrote: post #787
tomrod wrote: post #781

Earlier quoted context omitted.

As articles are increasingly published with AI, this becomes yet another example of, to coin a phrase, AI loopidity.

If writers are using ai to write articles, and readers are using ai to summarize the same article, what's the underlying inefficiency here? That writers should just be posting brief summaries of the articles in the first place? Or maybe they just need to be prompting ai to create a summary rather than a full article?

Differences in how people want to consume the article, and what information they’re looking for. Some want snippets, some want summaries, some want long form.

We have information compression machines now. Might as well raw dump the information and let the machine package it up in the format we prefer for consumption, instead of pre-packaging it. (Yeah, this is effectively what authors are doing…currently they can still do novel things that the compression machines can’t, but how long will that last?)

Re: OpenAI O3-Mini

#792

Post by ezst »

simonw wrote: post #557
Valakas_ wrote: post #554

Earlier quoted context omitted.

For those that like simpler ways (although dependent on Google) NotebookLM does all that in 2 clicks. And you can ask it questions about it, references are provided.

After you've run my hn-summary.sh script you can ask follow up questions like this: llm -c "did anyone talk about pricing?"

Back in the old ages we would get there with ctrl+f "price" and for free, but where's the fun in that.

Re: OpenAI O3-Mini

#793

Post by inciampati »

threecheese wrote: post #358
simonw wrote: post #306

I used o3-mini to summarize this thread so far. Here's the result: https://gist.github.com/simonw/09e5922be0cbb85894cf05e6d75ae... For 18,936 input, 2,905 output it cost 3.3612 cents. Here's the script I used to do it: https://til.simonwillison.net/llms/claude-hacker-news-themes...

I haven’t tried o3, but one issue I struggle with in large context analysis tasks is the LLMs are never thorough. In a task like this thread summarization, I typically need to break the document down and loop through chunks to ensure it actually “reads” everything. I might have had to recurse into individual conversations with some small max-depth and leaf count and run inference on each, and then have some aggregati…

> Is this a case of PEBKAC?

It's 2025 and every useful conversation with an LLM ends with context exhaustion. There are those who argue this is a feature and not a bug. Or that the context lengths we have are enough. I think they lack imagination. True general intelligence lies on the other side of infinite context length. Memory makes computation universal, remember? http://thinks.lol/2025/01/memory-makes-computation-universal...

Re: OpenAI O3-Mini

#794

Post by 0xFEE1DEAD »

moonsu wrote: post #770
0xFEE1DEAD wrote: post #742

After o3 was announced, with the numbers suggesting it was a major breakthrough, I have to say I’m absolutely not impressed with this version. I think o1 works significantly better, and that makes me think the timing is more than just a coincidence. Last week Nvidia lost 600 billion because of DeepSeek R1, and now OpenAI comes out with a new release which feels like it has nothing to do with the promises that were be…

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 behind closed doors?

If I put on my tinfoil hat for a moment, maybe Nvidia made a deal with OpenAI - offering a discount on computing power in exchange for a timely release.

OpenAI needs an enormous amount of computing power these days, and while Nvidia would take a financial hit by offering a discount to one of its biggest (if not the biggest) customers, that's still nowhere near as costly as losing 600 billion.

Re: OpenAI O3-Mini

#795

Post by endofreach »

jefe_ wrote: post #780
flakeoil wrote: post #757

Earlier quoted context omitted.

But then there will be no comments to summarize.

Our digital twins will write the comments. They will be us, but with none of our flaws. They will never experience the shame of posting a dumb joke, getting flamed, and then deleting it, for they will have tested all ideas to prevent such an oversight. They will never experience the satisfaction-turned-to-puzzlement of posting an expertly crafted, well-researched comment that took 2 hours of the workday to draft - on…

We all are, since a long time. The mud just doesn't shine so bright.

Re: OpenAI O3-Mini

#796

Post by scarface_74 »

ben_w wrote: post #772
scarface_74 wrote: post #748

Earlier quoted context omitted.

I’m not talking about the TTS and podcast creation. I’m talking about just asking questions where it gives you the answer with citations.

Given what it got wrong was in the LLM part, that is a distinction without a difference.

The podcast creation does a lot of small talk. Just like people do. Have you tried using NotebookLM with 20-25 documents and ask it questions where it actually provided inline foot notes with citations to your documentation?

Re: OpenAI O3-Mini

#797

Post by seunosewa »

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/

You're conflating the low price of the o3-mini medium effort model with the high performance of the o3-mini high effort model. OpenAI hasn't listed the price for the o3-mini high effort model separately on their pricing page.

Re: OpenAI O3-Mini

#798

Post by seunosewa »

dutchbookmaker wrote: post #500
sss111 wrote: post #41

Earlier quoted context omitted.

Deepseek V3 is equivalent to 4o. Deepseek R1 is equivalent to o1 (if not better) I think someone should just build an AI model comparing website at this point. Include all benchmarks and pricing

I had resubscribed to use o1 2 weeks ago and haven't even logged in this week because of R1. One thing I notice that is huge is being able to see the chain of thought lets me see when my prompt was lacking and the model is a bit confused on what I want. If I was anymore impressed with R1 I would probably start getting accused of being a CCP shill or wumao lol. With that said, I think it is very hard to compare models…

The thought-stream is very important to me as well.

Re: OpenAI O3-Mini

#799

Post by inciampati »

acka wrote: post #719
simonw wrote: post #560

Earlier quoted context omitted.

Good catch. That's likely an artifact of the way I flatten the nested JSON from the comments API. I originally did that to save on tokens but modern models have much larger input windows so I may not need to do that any more.

This is why I keep advocating that JSON should die, or at least no longer be used with LLMs. LLMs (and human brains) are simply not wired up for counting nested curly brackets across long spans of quoted text joined by colons and commas, and it is far too easy for humans to make mistakes when chunking JSON. IMO, (Strict)YAML is a very good alternative, it has even been suggested to me by multiple LLMs when I asked th…

LLMs (transformers) literally cannot balance parentheses. That's outside of their complexity class (TC0). You'd want a real UTM to count parentheses!

Re: OpenAI O3-Mini

#800

Post by ethbr1 »

s_dev wrote: post #669
simonw wrote: post #306

I used o3-mini to summarize this thread so far. Here's the result: https://gist.github.com/simonw/09e5922be0cbb85894cf05e6d75ae... For 18,936 input, 2,905 output it cost 3.3612 cents. Here's the script I used to do it: https://til.simonwillison.net/llms/claude-hacker-news-themes...

Currently on the internet people skip the article and go straight to the comments. Soon people will skip the comments and go striaght to an AI summary reading neither the original article nor the comments.

HN comments aren't valuable because of their average quality or content, but because of the feedback vote sorting.

(Aside from the tendency towards first=top. Would be nice to have time-weighted upvote decay instead of absolute count)

Post reply on HN