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

991–1000 of 1001 posts

Re: GPT-5 is behind schedule

#991

Post by netdevphoenix »

Biologist123 wrote: post #674
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…

But hasn’t it become quite easy to deal with this issue simply by asking for the sources of the information and then validating? I quite like using the consensus app and then asking for specific academic paper references which I can then quickly check. However this has taught me also that academic claims must also be validated…

If you need to validate the sources, you might as well go to the sources directly and bypass the LLM. The whole point of LLMs is not needing to go to the sources. The LLM consumes them for you. If you need to read and understand the sources yourself well enough to tell if the LLM is lying, the LLM is a wasteful middleman.

It's like buying supermarket food and also buying the same food from the farmers themselves.

Re: GPT-5 is behind schedule

#992

Post by Flenkno »

ConceptJunkie wrote: post #762
Flenkno wrote: post #716

Earlier quoted context omitted.

when i tried studying, i got really frustrated because i had to search for so many things and not a lot of people would explain basic math things to me in a simple way. LLMs do already a lot better job at this. A lot faster, accurate enough and easy to use. I can now study something alone which i was not able to do before.

> accurate enough Ask it something non-trivial about a subject you are an expert in and get back to me.

Accurate enough for it to explain to me details of 101, 201 and 301 university courses in math or physics.

Besides, when i ask it about things like SRE, Cloud etc. its a very good starting point.

Re: GPT-5 is behind schedule

#993

Post by dudeofea »

mrits wrote: post #787
solardev wrote: post #732

Earlier quoted context omitted.

> a revolution in knowledge-worker productivity. That's a nice euphemism for "imminent mass layoffs and a race to the bottom"...

The idea that someone should be paid by a corporation when they don't provide value is very strange to me. Doing so seems like the real race to the bottom

what about when someone provides long-term value? They would be replaced by a short-term thinking corp (namely, all of them) for providing less value than an alternative with it's value purely in the short-term.

We are accelerating by preferring short-term gains. Like a fire becoming an explosion, that's modern society. Corps now throw the future under the bus for a slight boost in short-term value.

Re: GPT-5 is behind schedule

#994

Post by xanderlewis »

SahAssar wrote: post #299
xanderlewis wrote: post #206

Earlier quoted context omitted.

It’s somehow funny to hear a British company being described as ‘in Europe’, but I suppose you’re technically correct…

The UK is part of Europe. It's technically, geographically, politically, historically, lingustially, tectonically and socially correct. In what ways is it not?

I don’t know — I’m not claiming to. I’m simply claiming that it’s a commonly-held belief.

Re: GPT-5 is behind schedule

#995

Post by phyalow »

pphysch wrote: post #959
phyalow wrote: post #940

Earlier quoted context omitted.

Um, augmentation (i.e. the generation of synthetic data) is a very very well known technique for improving learning. Also whats with the hate for MBA’s? Your comment is off kilter with the rules here.

Synthetic data is being proposed here as a solution to extrapolate ML scaling. Augmentation, interpolation, smoothing are different concepts.

I think you're drawing an artificial distinction here. Synthetic data generation is fundamentally an extension of augmentation. When OpenAI uses expert generated examples and curriculum based approaches, that's literally textbook augmentation methodology. The goal of augmentation has always been to improve model fit, and scaling is just one aspect of that.

Your concern about extrapolation is interesting but misses something key when we generate synthetic data through expert demonstration or guided curriculum, we're not trying to magically create capabilities beyond the training distribution. Instead, we're trying to better sample the actual distribution of problemsolving approaches humans use. This isn't extrapolation rather, better sampling of an existing, complex distribution!

i.e. if you think about the manifold hypothesis then we know real data lives on a lowerdimensional manifold, and good synthetic data helps fill those gaps. This naturally leads to better extrapolation, it's pretty well established at this point.

TBH I think you are characterizing this as some kind of blind data multiplication scheme, but it's much closer to curriculum learning you start with basic synthetic examples and gradually ramp up complexity. So it isn't whether synthetic data is "real" or not, but if it effectively helps map the underlying distribution and reasoning patterns.

Funny enough, your oil analogy actually supports the case for synthetic data refined petroleum is more useful than crude for specific purposes, just like well designed synthetic data can be more effective than raw internet text for certain learning objectives.

Re: GPT-5 is behind schedule

#996

Post by tiffanyh »

solardev wrote: post #732
LASR wrote: post #610

So the team I lead does a lot of research around all the “plumbing” around LLMs. Both technical and from a product-market perspectives. What I’ve learned is that for the most part that AI revolution is not going to be because of PHD-level LLMs. It will be because people are better equipped to use the high-schooler level LLMs to do their work more efficiently. We have some knowledge graph experiments where LLMs contin…

> a revolution in knowledge-worker productivity. That's a nice euphemism for "imminent mass layoffs and a race to the bottom"...

It’s that sang, “radiologist aren’t losing their jobs due to AI .. only radiologist who don’t use AI are losing their jobs”.

Re: GPT-5 is behind schedule

#997

Post by fennecfoxy »

kevinsync wrote: post #797
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…

I actually find LLMs lacking true expertise to be a feature, not a bug. Most of the time I'm starting from a place of no knowledge on a topic that's novel to me, I ask some questions, it replies with summaries, keywords, names of things, basic concepts. I enter with the assumption that it's really no different than googling phrases and sifting through results (except I don't know what phrases I'm supposed to be googl…

TBF they're only truly useful when hooked up to RAG imo. I'm honestly surprised that we haven't yet built a digital seal of authenticity for truth that can be used by AI agents + RAG to conceivably give the most accurate answer possible.

Scientists should be writing papers sealed digitally once they're peer reviewed and considered "truth", same thing with journalist/news articles - sealed once confirmed true or backed up by a solid source in the same way we trust root certificates.

But then again, especially when it comes to journalism, cropping photos, chopping quotes, etc all to misrepresent etc. Turns out we're all the bad actors; it's in our DNA. And tbf, many people when presented with hard evidence to the contrary of the opinion that they cling onto like a babe to a breast, just plug their ears and cover their eyes.

Okay so maybe there's no point seeking truth/factual correctness, our species doesn't want it 99% of the time, unless it affects them directly (eg people that shoot down public healthcare until they have an expensive illness themselves).

Re: GPT-5 is behind schedule

#998

Post by fennecbutt »

EVa5I7bHFq9mnYK wrote: post #890
ammanley wrote: post #841

Earlier quoted context omitted.

Im looking forward to the life experience that is content I want to read badly enough to endure a rectal exam.

It's not that bad ...

Not sure why you're being downvoted. Watching str8 bois react with shock and horror at the idea of anything near their butt is hilarious.

Prostate and rectal cancer is real, boys. Grow tf up about it.

Re: GPT-5 is behind schedule

#999

Post by fennecbutt »

scottLobster wrote: post #902
RationPhantoms wrote: post #848

Earlier quoted context omitted.

Would you trust a ML self-driving algorithm trained on a "digital twin" of a city? I would. I view synthetic training data like a digital twin in which it can provider further control or specified noise to understand from.

No, because right now I'm working closely with some EEs to troubleshoot electrical issues on some prototype boards (I wrote the firmware). They're prototypes precisely because we know the limits of our models and simulations and need real world boards to test our electronics design and firmware on. You're suggesting the new, untested models in a new, untested technological field are sufficient for deployment in real…

Hey, let's shut down humanity because human behaviour can't be perfectly simulated.

Re: GPT-5 is behind schedule

#1000

Post by fennecbutt »

atoav wrote: post #738
solardev wrote: post #732

Earlier quoted context omitted.

> a revolution in knowledge-worker productivity. That's a nice euphemism for "imminent mass layoffs and a race to the bottom"...

These productivity gains won't be shared with the employees. I think some people underestimate what a violent populus can do to them if they squeeze out even more Yacht money from the people.

Psssh, y'all been letting the billionaires and trillionaires do this forever now. Products only get more subpar and profit margins only grow and we're all too busy hating each other for sex, skin colour, sexuality, etc because we're just animals.

Ain't gonna change unless we genetically engineer our dumbass evolutionary history out of ourselves.

Post reply on HN