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

211–220 of 1001 posts

Re: GPT-4o

#211

Post by gabiruh »

cube2222 wrote: post #154
msoad wrote: post #58

They are admitting[1] that the new model is the gpt2-chatbot that we have seen before[2]. As many highlighted there, the model is not an improvement like GPT3->GPT4. I tested a bunch of programming stuff and it was not that much better. It's interesting that OpenAI is highlighting the Elo score instead of showing results for many many benchmarks that all models are stuck at 50-70% success. [1] https://twitter.com/Lia…

I think the live demo that happened on the livestream is best to get a feel for this model[0]. I don't really care whether it's stronger than gpt-4-turbo or not. The direct real-time video and audio capabilities are absolutely magical and stunning . The responses in voice mode are now instantaneous, you can interrupt the model, you can talk to it while showing it a video, and it understands (and uses) intonation and…

It's weird that the "airplane mode" seems to be ON on the phone during the entire presentation.

Re: GPT-4o

#212

Post by fvdessen »

cube2222 wrote: post #154
msoad wrote: post #58

They are admitting[1] that the new model is the gpt2-chatbot that we have seen before[2]. As many highlighted there, the model is not an improvement like GPT3->GPT4. I tested a bunch of programming stuff and it was not that much better. It's interesting that OpenAI is highlighting the Elo score instead of showing results for many many benchmarks that all models are stuck at 50-70% success. [1] https://twitter.com/Lia…

I think the live demo that happened on the livestream is best to get a feel for this model[0]. I don't really care whether it's stronger than gpt-4-turbo or not. The direct real-time video and audio capabilities are absolutely magical and stunning . The responses in voice mode are now instantaneous, you can interrupt the model, you can talk to it while showing it a video, and it understands (and uses) intonation and…

The demo is impressive but personally, as a commercial user, for my practical use cases, the only thing I care about is how smart it is, how accurate are its answers and how vast is its knowledge. These haven’t changed much since GPT-4, yet they should, as IMHO it is still borderline in its abilities to be really that useful

Re: GPT-4o

#213

Post by mike_hearn »

atgctg wrote: post #19

Tiktoken added support for GPT-4o: https://github.com/openai/tiktoken/commit/9d01e5670ff50eb74c... It has an increased vocab size of 200k.

Does that imply they retrained the foundation model from scratch? I thought changing the tokenization was something you couldn't really retrofit to an existing model. I mean sure they might have initialized the weights from the prior GPT-4 model but it'd still require a lot of retraining.

Re: GPT-4o

#214

Post by aaroninsf »

cube2222 wrote: post #154
msoad wrote: post #58

They are admitting[1] that the new model is the gpt2-chatbot that we have seen before[2]. As many highlighted there, the model is not an improvement like GPT3->GPT4. I tested a bunch of programming stuff and it was not that much better. It's interesting that OpenAI is highlighting the Elo score instead of showing results for many many benchmarks that all models are stuck at 50-70% success. [1] https://twitter.com/Lia…

I think the live demo that happened on the livestream is best to get a feel for this model[0]. I don't really care whether it's stronger than gpt-4-turbo or not. The direct real-time video and audio capabilities are absolutely magical and stunning . The responses in voice mode are now instantaneous, you can interrupt the model, you can talk to it while showing it a video, and it understands (and uses) intonation and…

Absolutely agree.

This model isn't about basemark chasing or being a better code generator; it's entirely explicitly focused on pushing prior results into the frame of multi-modal interaction.

It's still a WIP, most of the videos show awkwardness where its capacity to understand the "flow" of human speech is still vestigial. It doesn't understand how humans pause and give one another space for such pauses yet.

But it has some indeed magic ability to share a deictic frame of reference.

I have been waiting for this specific advance, because it is going to significantly quiet the "stochastic parrot" line of wilfully-myopic criticism.

It is very hard to make blustery claims about "glorified Markov token generation" when using language in a way that requires both a shared world model and an understanding of interlocutor intent, focus, etc.

This is edging closer to the moment when it becomes very hard to argue that system does not have some form of self-model and a world model within which self, other, and other objects and environments exist with inferred and explicit relationships.

This is just the beginning. It will be very interesting to see how strong its current abilities are in this domain; it's one thing to have object classification—another thing entirely to infer "scripts plans goals..." and things like intent, and, deixis. E.g. how well does it now understand "us" and "them" and "this" vs "that"?

Exciting times. Scary times. Yee hawwwww.

Re: GPT-4o

#215

Post by Jensson »

MVissers wrote: post #200
summerlight wrote: post #87

This is really impressive engineering. I thought real time agents would completely change the way we're going to interact with large models but it would take 1~2 more years. I wonder what kind of new techs are developed to enable this, but OpenAI is fairly secretive so we won't be able to know their sauce. On the other hand, this also feels like a signal that reasoning capability has probably already been plateaued a…

Why would reasoning have plateau’d? I think reasoning ability is not the largest bottleneck for improvement in usefulness right now. Cost is a bigger one IMO. Running these models as agents is hella expensive, and agents or agent-like recurrent reasoning (like humans do) is the key to improved performance if you look at any type of human intelligence. Single-shot performance only gets you so far. For example- If it c…

> Running these models as agents is hella expensive

Because they are dumb so you need to over compute so many things to get anything useful. Smarter models would solve this problem. Making the current model cheaper is like trying to solve Go by scaling up Deep Blue, it doesn't work to just hardcode dumb pieces together, the model needs to get smarter.

Re: GPT-4o

#216

Post by bredren »

simonw wrote: post #140
Jensson wrote: post #24

The most impressive part is that the voice uses the right feelings and tonal language during the presentation. I'm not sure how much of that was that they had tested this over and over, but it is really hard to get that right so if they didn't fake it in some way I'd say that is revolutionary.

That was very impressive, but it doesn't surprise me much given how good the voice mode is in the ChatGPT iPhone app is already. The new voice mode sounds better, but the current voice mode did also have inflection that made it feel much more natural than most computer voices I've heard before.

The voice mode was quite good but the latency and start / stop has been encumbering.

Re: GPT-4o

#217

Post by mrkramer »

MP_1729 wrote: post #105

This thing continues to stress my skepticism for AI scaling laws and the broad AI semiconductor capex spending. 1- OpenAI is still working in GPT-4-level models. More than 14 months after the launch of GPT-4 and after more than $10B in capital raised. 2- The rhythm that token prices are collapsing is bizarre. Now a (bit) better model for 50% of the price. How people seriously expect these foundational model companies…

>This thing continues to stress my skepticism for AI scaling laws and the broad AI semiconductor capex spending.

Imagine you are in 1970s and saying computers suck, they are expensive, there is not that many use cases....fast forward to 90s and you are using Windows 95 with GUI and chip astronomically more powerful that we had in 70s and you can use productivity apps , play video games and surf Internet.

Give AI time, it will fulfill its true protentional sooner or later.

Re: GPT-4o

#218

Post by bamboozled »

bbconn wrote: post #179
syntaxing wrote: post #32

I admit I drink the koolaid and love LLMs and their applications. But damn, the way it’s responds in the demo gave me goosebumps in a bad way. Like an uncanny valley instincts kicks in.

Yeah it made me realize that I actually don't want a human-like conversational bot (I have actual humans for that). Just teach me javascript like a robot.

Maybe it’s the geek in me, but I don’t want a talking computer.

I have enough talking people to deal with already .

Re: GPT-4o

#219

Post by moffkalast »

wslack wrote: post #184
syntaxing wrote: post #32

I admit I drink the koolaid and love LLMs and their applications. But damn, the way it’s responds in the demo gave me goosebumps in a bad way. Like an uncanny valley instincts kicks in.

It should do that, because it's still not actually an intelligence. It's a tool that is figuring out what to say in response that sounds intelligent - and will often succeed!

That kind of is an inteligence though. Chinese room meets solipsism and all that.

It is interesting how insanely close their demo is to the OSes in the movie "Her", it's basically a complete real life reproduction.

Re: GPT-4o

#220

Post by layer8 »

cs702 wrote: post #115

The usual critics will quickly point out that LLMs like GPT-4o still have a lot of failure modes and suffer from issues that remain unresolved. They will point out that we're reaping diminishing returns from Transformers. They will question the absence of a "GPT-5" model. And so on -- blah, blah, blah, stochastic parrots, blah, blah, blah. Ignore the critics. Watch the demos. Play with it. This stuff feels magical .…

> HAL's unemotional monotone in Kubrick's movie, "Space Odyssey," feels... primitive by comparison.

I’d strongly prefer that though, along with HAL’s reasoning abilities.

Post reply on HN