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

GPT-5 is behind schedule

wsj.com

Post reply on HN
Print view View original

691–700 of 1001 posts

Re: GPT-5 is behind schedule

#691

Post by Hugsun »

atonse wrote: post #457
SamPatt wrote: post #415

I'm sure the debate over the definition of AGI is important and will continue for a while, but... I can't care about it anymore. Between Perplexity searching and summarizing, Claude explaining, and qwen (and other tools) coding, I'm already as happy as can be with whatever you want to call this level of intelligence. Just today I used a completely local AI research tool, based on Ollama. It worked great. Maybe it won…

Same here. The ability to “talk to an expert” about any topic I’m curious about and ask very specific questions has been invaluable to me. It reminds me of being a kid and asking my grandpa a million questions, like how light bulbs worked, or what was inside his radio, or how do we have day and night. And before anyone talks about accuracy or hallucinations, these conversations usually are treated as starting off poi…

I mostly feel sorry for grandpa, he'll receive much less of these questions, if any. This is partially because I expect to become this grandpa and already suspect that some people aren't asking me questions they would be, if they had no access to chatgpt.

Re: GPT-5 is behind schedule

#692

Post by Fargren »

TeMPOraL wrote: post #572
Fargren wrote: post #555

Earlier quoted context omitted.

A history book is written by someone who knows the topic, and then reviewed by more people who also know the topic, and then it's out there where people can read it and criticize it if it's wrong about the topic. A question asked to an AI is not reviewed by anyone, and it's ephemeral. The AI can answer "yes" today, and "no" tomorrow, so it's not possible to build a consensus on whether it answers specific questions c…

> A question asked to an AI is not reviewed by anyone, and it's ephemeral. The AI can answer "yes" today, and "no" tomorrow, so it's not possible to build a consensus on whether it answers specific questions correctly. It's even more so with humans! Most of our conversations are, and has always been, ephemeral and unverifiable (and there's plenty of people who want to undo the little of permanence and verifiability w…

The context of my comment was what is the difference between an AI and a history book. Or going back to the top comment, between an AI and an expert.

If you want to compare AI with ephemeral unverifiable conversations with uninformed people, go ahead. But that doesn't make them sound very valuable. I believe they are more valuable than that for sure, but how much, I'm not sure.

Re: GPT-5 is behind schedule

#693

Post by karel-3d »

I have to say I finally "caved in" to LLMs last month.

While I still think Copilot is useless, I recently had a very complex code that did a lot of crazy bit-flipping and xoring, and I had no idea what is it doing, so I threw it to ChatGPT.. and it knew what it was doing.

I also needed to rewrite this code to PHP (for... reasons) while I know very little PHP. And it did that! It was a bit wrong, I needed to correct a bunch of stuff (based on domain knowledge), but... it helped me a ton.

I still can't image using it daily for domain and language I already know (that's why I never used copilot). But it actually helped me in measurable ways when it's something new.

Re: GPT-5 is behind schedule

#694

Post by FrustratedMonky »

gizmo wrote: post #686
netdevphoenix wrote: post #640

Earlier quoted context omitted.

> The ability to “talk to an expert” about any topic I’m curious about and ask very specific questions has been invaluable to me. It is dangerous to assume that LLMs are experts on any topic. With or without quotes. You are getting a super fast journalist intern with a huge memory but inability to reason critically, lacking understanding about anything and huge unreliability when it comes to answering questions (you…

People who are experts (PhD and 20 years of experience) often have very dumb opinions in their field of expertise. Experts make amateur mistakes too. Look at the books written by expert economists, expert psychologists, expert historians, expert philosophers, expert software engineers. Most books are not worth the paper they're written on, despite the authors being experts with decades of experience in their respecti…

Right, but, then what? If you throw away all of the books from experts, what do you do, go out in your backyard and start running experiments to re-create all of science? Or start googling? What, some random person on the internet is going to be a better 'expert' than someone that wrote a book?

  Books might not be great, but they are at least some minimum bar to reach.  You had to do some study and analysis.
Seems like any critic of books, if you scratch the surface is just the whole anti-science/anti-education tropes again and again. What is the option? Don't like peer review science, fine, it has flaws, propose an option.

Re: GPT-5 is behind schedule

#695

Post by bilsbie »

Angostura wrote: post #688
bilsbie wrote: post #681

Earlier quoted context omitted.

IMO it’s dangerous to call experts experts as well. Possibly more dangerous.

No. Expertise isn’t a synonym for ‘infallible’ it denotes someone whose lived experience, learned knowledge and skill means that you should listen to their opinion in their area of expertise - and defer to it, unless you have direct and evidence-based reasons for thinking they are wrong.

By that definition an expert would be trustworthy. (Usually they want you to look at credentials instead.)

However that still ignores human nature to use that trust for personal gain.

Nothing about expertise makes someone a saint.

Re: GPT-5 is behind schedule

#696

Post by ionwake »

coffeebeqn wrote: post #561
throwaway314155 wrote: post #270

Earlier quoted context omitted.

They hyped them like crazy and haven't discussed them once since then. I agree that the inability to change the model is pretty absurd when the whole point was to "supercharge" specific tasks. There was even talk of some sort of profit sharing with creators which clearly never happened. I just think the premise is too confusing for many and can still be served by using a custom system prompt via the API.

Was it hyped? I tried a few of them and they seemed absolutely useless. Like I could install a “custom GPT” that just appends something to my prompt? How great..

No the whole Point is you make a gpt for yourself and upload all your related documents to it and then query that. It performs 10x better than a generic query without attaching every single doc that could be relevant.

I am unsure if the answer is to use “projects” maybe this has superseded myGpts?

I am perplexed why HN isn’t focusing on this issue as all the Llm gains I’ve ever had were wit highly customised personal myGpts.

I can understand OpenAI and Sam’s having access to their own models may not even know what the best way to use the released stuff is

Ps - typing on my phone hence typos

Re: GPT-5 is behind schedule

#697

Post by jstummbillig »

jacobolus wrote: post #573
jstummbillig wrote: post #562

Earlier quoted context omitted.

A pop sci fi book can be written by someone who knows the topic and reviewed by people who know the topic — and a history book can also not. LLM generated answers are more comparable to ad-hoc human expert's answers and not to written books. But it's much simpler to statistically evaluate and correct them. That is how we can know that, on average, LLMs are improving and are outperforming human experts on an increasin…

In my experience LLM generated answers are more comparable to an ad-hoc answer by a human with no special expertise, moderate google skills, but good bullshitting skills spending a few minutes searching the web, reading what they find and synthesizing it, waiting long enough for the details to get kind of hazy, and then writing up an answer off the top of their head based on that, filling in any missing material by j…

> I've never gotten an answer from an LLM to a tricky or obscure question about a subject I already know anything about that seemed remotely competent.

Certainly not my experience with the current SOTA. Without being more specific, it's hard to discuss. Feel free to name something that can be looked at.

Re: GPT-5 is behind schedule

#698

Post by FrustratedMonky »

ithadtobe119 wrote: post #638
atonse wrote: post #457

Earlier quoted context omitted.

Same here. The ability to “talk to an expert” about any topic I’m curious about and ask very specific questions has been invaluable to me. It reminds me of being a kid and asking my grandpa a million questions, like how light bulbs worked, or what was inside his radio, or how do we have day and night. And before anyone talks about accuracy or hallucinations, these conversations usually are treated as starting off poi…

(throwaway account because of what I'm about to say, but it needs to be said) While my main use case for LLMs is coding just like most people here, there are lots of areas that are being ignored. Did you know llama 3.X models have been trained as psychotherapists? It's been invaluable to dump and discuss feelings with it in ways I wouldn't trust any regular person. When real therapists also cost more than what people…

This seems to be part of a side plot in Blade Runner 2049.

The movie was about replicants of course, but in the background, the technology shown with the AI being a companion, it was a huge corporate hit, a big seller. In the background you see ad's for it, and they reference it as their most popular product. And, as you allude to, in the movie it was both for loneliness AND sexual. They interacted like a relationship with talking and hooking up.

I don't doubt that with current AI, something similar could be done. We're just missing the holograms.

And as you say, I'm sure the porn industry will catch on.

Kind of crazy how Porn isn't leading this tech wave like past ones. Maybe because people are scared of tracking?

Re: GPT-5 is behind schedule

#699

Post by tomjen3 »

croes wrote: post #479
atonse wrote: post #457

Earlier quoted context omitted.

Same here. The ability to “talk to an expert” about any topic I’m curious about and ask very specific questions has been invaluable to me. It reminds me of being a kid and asking my grandpa a million questions, like how light bulbs worked, or what was inside his radio, or how do we have day and night. And before anyone talks about accuracy or hallucinations, these conversations usually are treated as starting off poi…

How do you know the answers are correct? More than once I got eloquent answer that are completely wrong.

You enable the search functionality.

Re: GPT-5 is behind schedule

#700

Post by sekai »

rwyinuse wrote: post #591
concerndc1tizen wrote: post #571

Earlier quoted context omitted.

20w for 20 years to answer questions slowly and error-prone at the level of a 30B model. An additional 10 years with highly trained supervision and the brain might start contributing original work.

And yet that 20w brain can make me a sandwich and bring it to me, while state of the art AI models will fail that task. Until we get major advances in robotics and models designed to control them, true AGI will be nowhere near.

> Until we get major advances in robotics and models designed to control them, true AGI will be nowhere near.

AGI has nothing to do with robotics, if AGI is achieved it will help push robotics and every single scientific field further with progression never seen before, imagine a million AGIs running in parallel focused on a single field.

Post reply on HN