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

Introducing deep research

openai.com

Post reply on HN
Print view View original

161–170 of 445 posts

Re: Introducing deep research

#161

Post by handfuloflight »

hnisoss wrote: post #144
handfuloflight wrote: post #48

Earlier quoted context omitted.

Then consent is granted by transitive property because these AI are yielded by humans.

Yea but guy paying closedai to get "insights" that basically copy-pasted content from my blog is definitely violating my blogs copyright, and in the end no coin comes to me either. What about that?

Could you provide an example where OpenAI outputting verbatim quotes actually constitutes the copyright violation? Because mechanically retrieving relevant quotes seems analogous to grep/search - the copyright status would depend on how downstream users transform and use that content. Like how quoting your blog in a technical analysis or critique is fair use, but wholesale republishing isn't. This suggests the violation occurs at usage time, not retrieval time.

Re: Introducing deep research

#162

Post by numba888 »

observationist wrote: post #153
spyckie2 wrote: post #133

Is this ability really a prerequisite to AGI and ASI? Reasoning, problem solving, research validation - at the fundamental outset it is all refinement thinking. Research is one of those areas where I remain skeptical it is that important because the only valid proof is in the execution outcome, not the compiled answer. For instance you can research all you want about the best vacuum on the internet but until you try…

It's a direction in a vast landscape, not a feature of itself - being better at different tasks, like search generally, and research in conjunction with reasoning, gets the model closer to AGI. An AGI will be able to do these tasks - so the point of the research is to have more Venn diagrams of capabilities like these to help narrow down the view on things that might actually be fundamental mechanisms involved in AGI…

> hings like feeling pain and pleasure

can machine feel? without that there is no AGI according to definition above.

and the second question: are animals "GI"? they don't have language and don't solve math problems, never heard of np-complete.

Re: Introducing deep research

#163

Post by OutOfHere »

prng2021 wrote: post #103

"Deep research was trained using end-to-end reinforcement learning" Does this mean they skipped supervised fine tuning like DeepSeek did with R1?

No, it just suggests that RL was used over a base SFT model, and moreover that RL here was tuned to this research task. Personally I don't think that RL is strictly necessary for this task at all, but perhaps it helps.

Re: Introducing deep research

#164

Post by ADeerAppeared »

panarky wrote: post #94
ADeerAppeared wrote: post #49

I'm sorry but what the fuck is this product pitch? Anyone who's done any kind of substantial document research knows that it's a NIGHTMARE of chasing loose ends & citogenesis. Trusting an LLM to critically evaluate every source and to be deeply suspect of any unproven claim is a ridiculous thing to do. These are not hard reasoning systems, they are probabilistic language models.

> they are probabilistic language models This is like arguing an Airbus cannot possibly fly because it is 165 tonnes of aluminum, steel and plastic. The proof is in the fact that it flies, not what it is constructed from.

> The proof is in the fact that it flies, not what it is constructed from.

And LLMs do not.

> "But it looks like reasoning to me"

My condolences. You should go see a doctor about your inability to count the number of 'R's in a word.

Re: Introducing deep research

#166

Post by ADeerAppeared »

autoconfig wrote: post #89
hi_hi wrote: post #56

This is terrifying. Even though they acknowledge the issues with hallucinations/errors, that is going to be completely overlooked by everyone using this, and then injecting the outputs into their own powerpoints. Management Consulting was bad enough before the ability to mass produce these graphs and stats on a whim. At least there was some understanding behind the scenes of where the numbers came from, and sources w…

Either you care about being correct or you don't. If you don't care then it doesn't matter whether you made it up or the AI did. If you care then you'll fact check before publishing. I don't see why this changes.

> If you care then you'll fact check before publishing.

Doing a proper fact check is as much work as doing the entire research by hand, and therefore, this system is useless to anyone who cares about the result being correct.

> I don't see why this changes.

And because of the above this system should not exist.

Re: Introducing deep research

#168

Post by InkCanon »

bonoboTP wrote: post #79
reader9274 wrote: post #36

I think we're all reaching AI fatigue. Fewer and fewer people care anymore

Sure if you're viewing this as some kind of spectator thing, or entertainment, maybe it's less interesting. But it doesn't really matter whether "people care". What matters is whether it's useful and has impact. It's enough if the small number of people use it for whom it is useful. It doesn't matter if the average Joe on the street is excited by it. Few people care or even know about various advances in various spec…

Not sure if it's just me, but it looks like all SOTA companies are doubling down to chase the new benchmark, which beyond hype, doesn't seem to translate into real world uses. Why don't these companies just plug it into a popular git repo and say, hey our AI fixed these 100 issues! Or something real? The only people who seem to be doing something real is DeepMind.

Re: Introducing deep research

#169

Post by febin »

febin wrote: post #18

Is this "deep research" tool exploiting open knowledge creators, using their work without compensation?

I see many are offended, but I am genuinely asking a question.

I want to understand does this mean it's ethical for anyone to create a research AI tool that will go through arXiv and related GitHub repo and use it to solve problems, implement ideas like cursor.

Re: Introducing deep research

#170

Post by random_cynic »

YmiYugy wrote: post #24

If I understood the graphs correctly, it only achieves 20% pass rate on their internal tests. So I have to wait 30min and pay a lot of money just to sift through walls of most likely incorrect text? Unless the possibility of hallucinations is negligible, this is just way too much content to review at once. The process probably needs to be a lot more iterative.

The difference is that it takes few minutes to an hour at most so it can be run multiple times a day, using the results of previous runs to further refine the search and reasoning process to get better outcomes. Pretty much how any human research works but much faster and with potentially vastly more world-knowledge and reasoning capability than average humans. And these capabilities will rapidly improve with further RL.
Post reply on HN