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

GPT-4o

openai.com

Post reply on HN
Print view View original

741–750 of 1001 posts

Re: GPT-4o

#741

Post by kenjackson »

anon373839 wrote: post #733
plaidfuji wrote: post #717

This is a very cool demo - if you dig deeper there’s a clip of them having a “blind” AI talk to another AI with live camera input to ask it to explain what it’s seeing. Then they, together, sing a song about what they’re looking at, alternating each line, and rhyming with one another . Given all of the isolated capabilities of AI, this isn’t particularly surprising, but seeing it all work together in real time is pre…

> We’re moving toward a world where every job will be modeled After an OpenAI launch, I think it's important to take one's feelings about the future impact of the technology with a HUGE grain of salt. OpenAI are masters of hype. They have been generating hype for years now, yet the real-world impacts remain modest so far. Do you remember when they teased GPT-2 as "too dangerous" for public access? I do. Yet we now ha…

GPT-4 already seems better at reasoning than most people. It just has an unusual training domain of Internet text.

Re: GPT-4o

#742

Post by ipnon »

CooCooCaCha wrote: post #467
mellosouls wrote: post #69

Very, very impressive for a "minor" release demo. The capabilities here would look shockingly advanced just 5 years ago. Universal translator, pair programmer, completely human sounding voice assistant and all in real time. Scifi tropes made real. But: Interesting next to see how it actually performs IRL latency and without cherry-picking. No snark, it was great but need to see real world power. Also what the benefit…

My guess is they're banking on the free version being rate limited and people finding it so useful that they want to remove the limit. Like giving a new user a discount on heroin. At least that's the strategy that would make most sense to me.

The free tier is just to get more training data.

Re: GPT-4o

#744

Post by osigurdson »

I appear to have GPT-4o but the iPhone app seems to be largely the same - can't interrupt it, no "emotive" voice, etc. Is this expected?

Re: GPT-4o

#745

Post by dmje »

Being sarcastic and then putting the end result in front of Brits could be the new Turing Test

Re: GPT-4o

#746

Post by somenameforme »

plaidfuji wrote: post #717

This is a very cool demo - if you dig deeper there’s a clip of them having a “blind” AI talk to another AI with live camera input to ask it to explain what it’s seeing. Then they, together, sing a song about what they’re looking at, alternating each line, and rhyming with one another . Given all of the isolated capabilities of AI, this isn’t particularly surprising, but seeing it all work together in real time is pre…

Like another comment mentioned, sigmoid curves [1] are ubiquitous with neural network systems. Neural network systems can be intoxicating because it's so "easy" (relatively speaking) to go from nothing to 80% in extremely short periods of time. And so it seems completely obvious that hitting 100% is imminent. Yet it turns out that each percent afterwards starts coming exponentially more slowly, and we tend to just bump into seemingly impassable asymptotes far from where we'd like to be.

~8 years ago when self driving technology was all the rage and every major company was getting on board with ever more impressive technological demos, it seemed entirely reasonable to expect that we'd all be in a world of complete self driving imminently. I remember mocking somebody online around the time who was pursuing a class C/commercial trucking license. Yet now a decade later, there are more truckers than ever and the tech itself seems further away than ever before. And that's because most have now accepted that progress on such has basically stalled out in spite of absolutely monumental efforts at moving forward.

So long as LLMs regularly hallucinate, they're not going to be useful for much other than tasks that can accept relatively high rates of failure. And many of those generally creative domains are the ones LLMs are paradoxically the weakest in - like writing. Reading a book written by an LLM would be cruel and unusual punishment given then current state of the art. One domain I do see them completely taking over is search. They work excellently as natural language search engines, and "failure" in such is very poorly defined.

[1] - https://en.wikipedia.org/wiki/Sigmoid_function

Re: GPT-4o

#747

Post by Jiahang »

plaidfuji wrote: post #717

This is a very cool demo - if you dig deeper there’s a clip of them having a “blind” AI talk to another AI with live camera input to ask it to explain what it’s seeing. Then they, together, sing a song about what they’re looking at, alternating each line, and rhyming with one another . Given all of the isolated capabilities of AI, this isn’t particularly surprising, but seeing it all work together in real time is pre…

make people smart

Re: GPT-4o

#748

Post by p1dda »

First impressions as a 1-year subscriber. I just tried GTP-4o to evaluate my code for suggestions and for discussing other solutions and it is definitely faster and it comes up with new suggestions that GPT-4 didn't. Currently in the process of evaluating the suggestions.

The demo is what it is, designed to get a wow from the masses.

Re: GPT-4o

#749

Post by callen43 »

This is clearly not just another story of human innovation. This is not just the usual trade-off between risks and opportunities.

Why? Because it simply automates the human away. Who wouldn't opt for a seemingly flawless, super effective buddy (i.e. an AI) that is never tired, always knows better? if you need some job done, if you're feeling lonely, when you need some life advice.. It doesn't matter if it might be considered "just imitation of human".

Why would future advancements of it keep being "just some tool" instead of largely replacing us as (humans) in jobs, relationships, ...?

Post reply on HN