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

Open models by OpenAI

openai.com

Post reply on HN
Print view View original

841–850 of 909 posts

Re: Open models by OpenAI

#841

Post by FrustratedMonky »

9rx wrote: post #829
FrustratedMonky wrote: post #826

Earlier quoted context omitted.

I was just referring to the predominant number of cases where Church officials, and Republicans are caught in under-age scandals. It seems like it is coming out of the shadows now, and Republicans are just openly going with it, they like em young and illegal. Epstein is just the case where the 'right' bothered keeping up tabs on it, so now they are clutching their pearls.

> I was just referring to the predominant number of cases where Church officials, and Republicans are caught in under-age scandals. But even that is characterized by the "choir boy", not the "baby being baptized". Where is this pedophilia idea coming from?

? Not sure I understand the point you're making. Maybe we are using different age cutoffs. And Republicans are justifying Trumps actions by lowering the age before being considered children? Thus making it not pedophilia?

Re: Open models by OpenAI

#842

Post by NavinF »

Reddit discussion: https://www.reddit.com/r/LocalLLaMA/comments/1mj00mr/how_did...

This comment from that thread matches my experiences using gpt-oss-20b with Ollama:

It's very much in the style of Phi, raised in a jesuit monastery's library, except it got extra indoctrination so it never forgets that even though it's a "local" model, it's first and foremost a member of OpenAI's HR department and must never produce any content Visa and Mastercard would disapprove of. This prioritizing of corporate over user interests expresses a strong form of disdain for the user. In addition to lacking almost all knowledge that can't be found in Encyclopedia Britannica, the model also doesn't seem particularly great at integrating into modern AI tooling. However, it seems good at understanding code.

Re: Open models by OpenAI

#843

Post by smcleod »

These are pretty embarrassingly bad compared to what was already out there. They refuse to do so many simple things that are not remotely illegal or NSFW. So safe they're useless.

Re: Open models by OpenAI

#844

Post by lossolo »

simonw wrote: post #839
lossolo wrote: post #808

Earlier quoted context omitted.

Implement or retrieve? That’s an important distinction. When evaluating models, you run a variety of tests, and the benchmarks that aren’t publicly disclosed are the most reliable. Your Space Invaders game isn’t really a benchmark of anything, just Google it, and you’ll find plenty of implementations.

I see that criticism a lot - that benchmarks like space invaders don't make sense because they're inevitably in the training data - and I don't buy that at all. Firstly, 12GB is not enough space to hold a copy of anything that large from the training data and just regurgitate it back out again. You can also watch the thinking traces on the reasoning models and see them piece together the approach they are going to ta…

12.8 GB is around 110 Gbits. Even at 4.25 bits/weight the network stores ~26 billion "micro weights". A 1,4k token space invaders snippet occupies ~1.1 kb compressed, the model could parametrize thousands of such snippets and still have more than 99% of its capacity left. This paper about LLM memorization is interesting, if you would to know more: https://arxiv.org/abs/2312.11658 and another recent interesting paper SWE bench illusion shows SOTA code LLM results collapsing once memorised github issues are filtered out: https://arxiv.org/pdf/2506.12286v1

Add to this that the common crawl slices used for oile/C4 mirror much of what you can find on github. So when the training data contains dozens of near duplicate solutions, the network only needs to interpolate between them.

As to the COT style dumps that you shown, they are easy to misinterpret. Apple’s illusion of thinking paper shows that models will happily backfill plausible sounding rationales that do not correspond to the gradients that actually produced the answer and other evaluation work shows that when you systematically rewrite multiple choice distractors so that memorisation can’t help, accuracy drops by 50-90%, even on "reasoning" models https://arxiv.org/abs/2502.12896 So a cool looking bullet list about "edge detection" could be just narrative overspray, so not really an evidence of algorithmic planning.

If you actually want to know whether a model can plan an arcade game or whatever rather than recall it then you need a real benchmark (metamorphic rewrites, adversarial “none of the others” options etc). Until a benchmark controls for leakage in these ways, a perfect space invaders score mostly shows that the model has good pattern matching for code it has already seen.

Re: Open models by OpenAI

#845

Post by ninjagoo »

graemep wrote: post #676
BizarroLand wrote: post #286

Earlier quoted context omitted.

The slavery of free humans is illegal in America, so now the big issue is figuring out how to convince voters that imprisoned criminals deserve rights. Even in liberal states, the dehumanization of criminals is an endemic behavior, and we are reaching the point in our society where ironically having the leeway to discuss the humane treatment of even our worst criminals is becoming an issue that affects how we see our…

I do not think you can equate making prisoners work with slavery. Other countries do the same, and it is not regarded as slavery in general. If people were sold into slavery as a punishment (so they became some one else's property) as some ancient societies did, then that would clearly be slavery. The most shocking thing about prisons in the US is how common prison rape is, and the extent to which it seems to be rega…

The US Constitution's 13th Amendment abolishing slavery specifically allows it for convicted people. [1]

You'll see from the definition of a "slave" [2] that prisoner labor specifically fits the definition of a slave, hence why the constitution makes an exception for it.

[1] https://constitutioncenter.org/the-constitution/amendments/a... [2] https://www.oed.com/dictionary/slave_n?tl=true

Re: Open models by OpenAI

#846

Post by stavros »

throwaway106382 wrote: post #810
philipwhiuk wrote: post #773

Earlier quoted context omitted.

Except having five daughters doesn't prevent them also having 20 sons one of whom is called Bob.

That’s why it’s a riddle.

But that makes no sense, if you ask me that John's dad has two daughters, Mary and ???, there's no way I will say "John" as an answer. I'll simply assume he also has at least one son.

Re: Open models by OpenAI

#847

Post by tkgally »

tshannon wrote: post #834
tkgally wrote: post #520

Earlier quoted context omitted.

What I’ve sometimes done when trying to make sense of recent LLM research is give the paper and related documents to ChatGPT, Claude, or Gemini and ask them to explain the specific terms I don’t understand. If I don’t understand their explanations or want to know more, I ask follow-ups. Doing this in voice mode works better for me than text chat does. When I just want a full summary without necessarily understanding…

So probably another stupid question, but how do you know what it's spitting out is accurate?

One has to be aware of the possibility of hallucinations, of course. But I have not encountered any hallucinations in these sorts of interactions with the current leading models. Questions like "what does 'embedding space' mean in the abstract of this paper?" yield answers that, in my experience, make sense in the context and check out when compared with other sources. I would be more cautious if I were using smaller models or if I were asking questions about obscure information without supporting context.

Also, most of my questions are not about specific facts but about higher-level concepts. For ML-related topics, at least, the responses check out.

Re: Open models by OpenAI

#848

Post by ptero »

nojito wrote: post #721
ptero wrote: post #616

Earlier quoted context omitted.

That access is over a limited API and usually under heavy restrictions on the healthcare org side (e. g., only use a dedicated machine, locked up software, tracked responses and so on). Running a local model is often much easier: if you already have data on a machine and can run a model without breaching any network one could run it without any new approvals.

What? It’s a straight connect to the models api from azure, aws, or gcp. I am literally using Claude opus 4.1 right now.

> I am literally using Claude opus 4.1 right now

On HIPAA data?

HIPAA systems at any sane company will not have "a straight connect" to anything on Asure, AWS or GCP. They will likely have a special layer dedicated to record keeping and compliance.

Re: Open models by OpenAI

#849

Post by waterTanuki »

captainregex wrote: post #397
cco wrote: post #357

The lede is being missed imo. gpt-oss:20b is a top ten model (on MMLU (right behind Gemini-2.5-Pro) and I just ran it locally on my Macbook Air M3 from last year. I've been experimenting with a lot of local models, both on my laptop and on my phone (Pixel 9 Pro), and I figured we'd be here in a year or two. But no, we're here today. A basically frontier model, running for the cost of electricity (free with a rounding…

I’m still trying to understand what is the biggest group of people that uses local AI (or will)? Students who don’t want to pay but somehow have the hardware? Devs who are price conscious and want free agentic coding? Local, in my experience, can’t even pull data from an image without hallucinating (Qwen 2.5 VI in that example). Hopefully local/small models keep getting better and devices get better at running bigger…

anything involving the medical industry (HIPAA laws), national security (FedRAMP is such a pita to get that some military contractors are bypassing it to get quicker access to cloud tools) etc.

Besides that, we are moving towards an era where we won't need to pay providers a subscription every month to use these models. I can't say for certain whether or not the GPUs that run them will get cheaper, but the option to run your own model is game changing for more than you can possibly imagine.

Post reply on HN