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

First Impressions with GPT-4V(ision)

blog.roboflow.com

Post reply on HN
Print view View original

311–320 of 345 posts

Re: First Impressions with GPT-4V(ision)

#311

Post by sexy_seedbox »

wokwokwok wrote: post #26

I’m impressed, technically, but this seems niche. Who holds their phone up and takes a photo then wants to know it was a photo of? That’s weird. If you don’t know what it is, wtf did you take photo? The obvious use here is natural language improvement / photo editing for photos, but this is just a stepping stone to that, and bluntly, as it stands… the examples really don’t shine… Great for the vision impaired. …not s…

> If you don’t know what it is, wtf did you take photo?

Currently traveling in foreign country, this is exactly my use case seeing weird flowers and fruits I've never seen before.

Re: First Impressions with GPT-4V(ision)

#312

Post by cambaceres »

cs702 wrote: post #10

Sure, there are a few edge-case failures and mistakes here and there, but I can't help but be in awe . AWE. Let me state the obvious, in case anyone here isn't clear about the implications: If the rate of improvement of these AI models continues at the current pace, they will become a superior user interface to almost every thing you want to do on your mobile phone, your tablet, your desktop computer, your car, your…

I read this comment in the worst of moments, because I am a frontend freelancer that all of a sudden can't find a gig plus I just bought an expensive flat for my family of three kids.

Can someone please relieve my anxiety?

Re: First Impressions with GPT-4V(ision)

#313

Post by jug »

sexy_seedbox wrote: post #280
Nfthero wrote: post #162

Earlier quoted context omitted.

I got Tendonitis and very very curious how soon I can really start to do an interaction like in star trek. Walking around and thinking out loud with the computer.

Pretty sure somebody has made a Starfleet Badge before tied to Alexa or Google Home, just need to change it to ChatGPT and mentioning "computer" activates it.

And for less perfect designs you’ll have ChatGPT rolling out voice chat on smartphones pretty much now to Plus subscribers.

And as for third party solutions, you have watchGPT for Apple Watch with voice commands and support for adding it as a watch complication.

Re: First Impressions with GPT-4V(ision)

#314

Post by TrapLord_Rhodo »

Well... that's it. It can officially do image recognition better than i can.

I had no idea those were zloty, and i've been to poland. One of them looked like a Euro with the gold rim, and i thought the other two were state quarters.

It got way closer on the nvidia joke than some of my non-technical friends would have.

Re: First Impressions with GPT-4V(ision)

#315

Post by fdeage »

zamadatix wrote: post #204
fdeage wrote: post #194

Could someone with access tell me what GPT-4V has to say of this image? http://karpathy.github.io/assets/obamafunny.jpg Andrej Karpathy used it in 2012 as an example of an image he thought would be extremely hard for a model to interpret. I'm wondering how this holds 11 years later...

Prompt: "What do you have to say about this image?" Response 1: "The image depicts a candid moment in what appears to be a hallway or corridor setting. Several individuals are present, engaging in various activities. The prominent features include a man on the left, who seems to be peering into a room, and another man next to him standing on a weighing scale and taking some notes. The individuals in the background ar…

Thank you. The quality of the answer it gave you is on par with what I was expecting, except I am very surprised it didn't recognize Obama.

Also, do you know why this response differ so much from ( and is far less accurate than) this response? https://news.ycombinator.com/item?id=37674968

In the previous link, GPT4V seems to have understood everything there was to understand about the picture (suspiciously, btw. As someone else said it, the picture and its text are almost certainly in the training data).

Re: First Impressions with GPT-4V(ision)

#316

Post by zarzavat »

nimithryn wrote: post #223
Michelangelo11 wrote: post #195

Earlier quoted context omitted.

So ... in this example, your job is continually feeding information to the AI from various sources. Why would the AI not be automatically hooked up to all those sources? Building a system that can do that is essentially trivial with the OpenAI API.

Have you tried to build a system like this? It’s not trivial

We have 32k contexts now, how big do you want to bet the context will be in 10 years?

That’s before you add any summarization, fine tuning or other tricks.

The thing that computers have always done much better than humans is deal with much larger volumes of information. The thing that humans have always done much better than computers is reason on that information better. Now the computers are coming for that too.

Re: First Impressions with GPT-4V(ision)

#317

Post by troupo »

Workaccount2 wrote: post #197
troupo wrote: post #181

Earlier quoted context omitted.

> It will be able to do it even faster, better and more cheaply than a human can. Take what you did in the past year. Write down every product decision taken, every interaction with other teams figuring out APIs you had, all the infra where your code is running and how it was setup and changed, all the design iterations and changes that had to be implemented (especially if you have external partners demanding it). Ye…

We'll have jobs, but they sure as shit won't be worth $150k anymore. Any grunt can feed meeting notes into an AI. And frankly, and AI can parse an audio recording on a meeting.

Can any grunt then check that the AI actually produce what was actually required?

And yes, how can we forget that any audio of a meeting has just the correct and final specifications, and not meandering discussions about anything and everything. Can't wait to see a canhazcheeseburger in a financial app because people in the meeting had cats on camera, and people demanded to see them.

I mentioned a "Revert Norway tax code" elsewhere https://news.ycombinator.com/item?id=37680060 It was a bit tongue in cheek, but a similar requirement ended up with:

- six months of discussions involving almost 20 people

- 4 new BigTable tables

- Deployment of 4 new Dataflow jobs, and fixes to two other Dataflow jobs

- Several complex test runs across the entire system including a few recreations of last year's full data to test that nothing broke

Not a grunt job, definitely. And I'm 100% sure that people doing that would still have their jobs 10 years from now, even with AI.

It amazes me, really, that people who would otherwise boast about how rational they are, and how they follow logic etc. completely replace all their knowledge and expertise with child-like belief in magic when it comes to anything AI-related.

Re: First Impressions with GPT-4V(ision)

#318

Post by troupo »

Michelangelo11 wrote: post #195
troupo wrote: post #181

Earlier quoted context omitted.

> It will be able to do it even faster, better and more cheaply than a human can. Take what you did in the past year. Write down every product decision taken, every interaction with other teams figuring out APIs you had, all the infra where your code is running and how it was setup and changed, all the design iterations and changes that had to be implemented (especially if you have external partners demanding it). Ye…

So ... in this example, your job is continually feeding information to the AI from various sources. Why would the AI not be automatically hooked up to all those sources? Building a system that can do that is essentially trivial with the OpenAI API.

Yup. "Just" "trivially" plug it in to those sources.

https://news.ycombinator.com/item?id=37688169

Re: First Impressions with GPT-4V(ision)

#319

Post by troupo »

astrange wrote: post #214
troupo wrote: post #168

Earlier quoted context omitted.

> The concept of self-driving vehicles even a decade ago seemed absurd (or even AI for that matter), but now it all seems like a reality. Nope. It's still not close to reality. It's as close to reality as it has been for the past 10 years while it was being hyped up to be close to reality. > And that's not even taking into account all the advances we'll have with AI within the next decade that we haven't even thought…

Cruise and Waymo are in production in really difficult cities. I think it's definitely in the last 10 years of not being here.

> Cruise and Waymo are in production in really difficult cities.

Cruise and Waymo are in production in very tightly fine-tuned and carefully monitored situations in two cities. We've yet to see if that can be easily (or at all) adapted to driving anywhere else.

Re: First Impressions with GPT-4V(ision)

#320

Post by chx »

SoftTalker wrote: post #294
chx wrote: post #283

Earlier quoted context omitted.

read the problem again please. it's not that it misidentified what's in the squares, it "miscounted" the number of squares -- because it doesn't count, it's just a plausible sentence generator.

Proving that CAPTCHAs are a poor test for "are you human" as we both failed.

my point still has nothing to do with captchas...
Post reply on HN