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

201–210 of 445 posts

Re: Introducing deep research

#201

Post by stonogo »

michaelgiba wrote: post #183

Gemini has had this for a month or two, also named "Deep Research" https://blog.google/products/gemini/google-gemini-deep-resea... Meta question: what's with all of the naming overlap in the AI world? Triton (Nvidia, OpenAI) and Gro{k,q} (X.ai, groq, OpenAI) all come to mind

It's a sort of unofficial trade association where they coalesce on specific redefinitions of terms to meet their sales and PR efforts. First they came for "intelligence," then "open source," then "reason," and it will continue. Any word which the PR wants but they can't achieve gets redefined -- "grok" is a perfect example, since in the original sci-fi book it meant "total understanding." The mythological Triton ruled the deeps, so the "deep learning" sales copy immediately co-opted it.

Re: Introducing deep research

#202

Post by CamperBob2 »

rvnx wrote: post #83
tmnvdb wrote: post #57

Earlier quoted context omitted.

Peak HN comment

Doesn't make it untrue. Agents that can search the internet exist for a while now and have been essentially solved and happily used in platforms like Perplexity. It's really "meh", very far from revolutionary. Keep in mind this company is trying to convince everybody they need 500B USD now (through the Stargate project).

Let us know when your Bing-bot scores over 20% on the HLE benchmark.

Re: Introducing deep research

#203

Post by CamperBob2 »

ADeerAppeared wrote: post #164
panarky wrote: post #94

Earlier quoted context omitted.

> 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.

OK, what's your next move, now that letter-counting has been solved by the current generation of frontier models?

CoT reasoning is reasoning, whether you like it or not. If you don't understand that, it means the models are already smarter than you.

Re: Introducing deep research

#204

Post by blackeyeblitzar »

figers wrote: post #27
blackeyeblitzar wrote: post #14

Earlier quoted context omitted.

I’m not sure I understand what you mean by “the button”. If you’re comparing this to DeepSeek’s copying, it’s not really the same thing right? DeepSeek essentially stole intellectual property by violating OpenAI’s terms of service. As I understand it, this is a copy of Google’s Deep Research

Didn't OpenAI steal everyone's data they could consume from the internet? Actively being sued by the NY Times and others for this...

Yes those cases will be interesting. By default a lot of copyrighted content may be legal to use for training (in the US but also many other places) under what’s called fair use. The cases you’re referring to will likely reinforce this, but it isn’t known yet. Note that it’s not just OpenAI on that side of the argument but also other (non tech) organizations that believe protecting fair use here is current law and essential.

Re: Introducing deep research

#205

Post by simonw »

azinman2 wrote: post #151
autoconfig wrote: post #89

Earlier quoted context omitted.

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.

When things are easy, you’re going to take the easy path even if it means quality goes down. It’s about trade offs. If you had to do it yourself, perhaps quality would have been higher because you had no other choice. Lots of kids don’t want to do homework. That said, previously many would because there wasn’t another choice. But now they can just ask ChatGPT for the answers they’ll write that down verbatim with zero…

"Lots of kids don’t want to do homework"

Sure, but if you're a professional you have to care about your reputation. Presenting hallucinated cases from ChatGPT didn't go very well for that lawyer: https://www.nytimes.com/2023/05/27/nyregion/avianca-airline-...

Re: Introducing deep research

#206

Post by sockaddr »

tomrod wrote: post #124
VerdisQuo5678 wrote: post #75

The accuracy of this tool does not matter. This is exclusively designed for box ticking "reports" that nobody reads and a produced for the sake of itself.

The new term for this is "AI Loopidity", highlighting the unintelligent ouroboros nature of one side using AI to generate content and then another side to consume content.

Similar to “Bullshit jobs”

All the AI commercials are designed to appeal to people that don’t produce any actual value but haven’t been detected by the system yet.

Need to send email to boss? Press magic button! Job well done, idiot.

Someone send you big scary email? Press magic button! Good job dummy!

Someone wants to go eat some Italian with you, push magic button for totally not-ad result. Enjoy your Olive Garden, moron.

Re: Introducing deep research

#207

Post by blackeyeblitzar »

therealpygon wrote: post #28
blackeyeblitzar wrote: post #14

Earlier quoted context omitted.

I’m not sure I understand what you mean by “the button”. If you’re comparing this to DeepSeek’s copying, it’s not really the same thing right? DeepSeek essentially stole intellectual property by violating OpenAI’s terms of service. As I understand it, this is a copy of Google’s Deep Research

Care to explain how something that cannot be copyrighted and was not generated by a human is “intellectual property“? Or are you just parroting a narrative?

Trade secrets are protected by the law. It doesn’t require copyright.

Re: Introducing deep research

#208

Post by simonw »

highfrequency wrote: post #196

Can it compile and run (non-Python) code as part of its tool use? Compile-run steps always seemed like they would be a huge value add during reasoning loops - it feels very silly to get output from ChatGPT, try to run it in terminal, get an error and paste the error to have ChatGPT immediately fix it. Surely it should be able to run code during the reasoning loop itself?

It sounds like it can run Python, which means it has access to Code Interpreter, which means it can run various other languages as well if you can convince it to do so.

I've used Code Interpreter to compile and run C code - https://simonwillison.net/2024/Mar/23/building-c-extensions-... - and I've managed to get it to run JavaScript (by uploading a Deno binary) and even Lua and PHP in the past as well: https://til.simonwillison.net/llms/code-interpreter-expansio...

Re: Introducing deep research

#210

Post by simonw »

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…

> Is this ability really a prerequisite to AGI and ASI?

That depends entirely on how you choose to define "AGI".

Post reply on HN