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

321–330 of 345 posts

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

#321

Post by sarchertech »

mycall wrote: post #305
sarchertech wrote: post #250

Earlier quoted context omitted.

> When taking a shower, I would like fine control over the water temperature, preferably with a feedback loop regulating the temperature. (Preferably also the regulation changes over the duration of the showering.) That exists, but it’s expensive because of the electronics and mechanics involved. There are so many interfaces with this exact problem. You also almost certainly don’t want non-deterministic hallucination…

Prompt engineering and using multiple AI models in parallel might find ways to cancel out most hallucinations similar to how consensus-based replication works.

It might. If hallucinations are truly random and not correlated to anything shared between models. For example, something inherent to the data they are trained on. Given how locked down I think potential training data is going to become, and the amount of data required, I think that sharing data between models is almost guaranteed.

Also that sounds like an awful lot of computing power for everyday UIs. It also doesn’t solve the non determinism problem.

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

#322

Post by Workaccount2 »

troupo wrote: post #317
Workaccount2 wrote: post #197

Earlier quoted context omitted.

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…

Your coping is painfully obvious.

I'd be sacred too, but at least would be taking a rational approach to it. It's an adapt or die situation. Putting your head in the sand is just gonna get you mowed down.

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

#323

Post by qingcharles »

moffkalast wrote: post #234
qingcharles wrote: post #120

Earlier quoted context omitted.

LOL. I already use it to write half my CSS, and that was before it could see what I was fucking up. I'm looking forward to this.

Real talk, GPT4 is an underrated CSS god, shaking description accurate animations from its sleeve at ridiculous speed. Reminds me of this (apparently now eight year old) meme: https://i.imgur.com/GcZFBaT.png This used to be funny, now it's just Tuesday.

Here's Thursday:

https://chat.openai.com/share/db204acc-7be5-42c1-a0ba-5c5aa7...

It says I have to put my favorite fonts in myself, though.

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

#324

Post by troupo »

Workaccount2 wrote: post #322
troupo wrote: post #317

Earlier quoted context omitted.

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…

Your coping is painfully obvious. I'd be sacred too, but at least would be taking a rational approach to it. It's an adapt or die situation. Putting your head in the sand is just gonna get you mowed down.

Ad hominem is not as good an argument as you think it is.

The only ones scared in this conversation are you and others who literally say you're scared for your jobs and your careers because of a magical boogey man.

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

#325

Post by zamadatix »

fdeage wrote: post #315
zamadatix wrote: post #204

Earlier quoted context omitted.

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 s…

I think it comes down to what you specifically put in the prompt. If you frame your question around the people specifically then ChatGPT will frame its response around as much as it knows or can invent about the people as well. It doesn't really have a good concept itself of what to focus on talking about outside of that, hence the lame generalizations around the very generic prompt I started with.

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

#326

Post by munksbeer »

zarzavat wrote: post #316
nimithryn wrote: post #223

Earlier quoted context omitted.

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.

> Now the computers are coming for that too.

Hopefully.

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

#327

Post by JHonaker »

wilg wrote: post #264
crazygringo wrote: post #109

Earlier quoted context omitted.

> they will become a superior user interface to almost every thing you want to do No they won't. They're actually a pretty terrible user interface from a design perspective. Primarily because they provide zero affordances, but also because of speed. UX is about providing an intuitive understanding of available capabilities at a glance, and allowing you to do things with a single tap that then reflect the new state ba…

There’s no reason chatbots have to be the interface to an LLM. Imagine dynamically generated interfaces redesigning themselves to your needs as you work through a task.

I would hate that. An interface is useful when you can anticipate and remember what things are available and where they are. It doesn't matter if it's a visual UI or keyboard shortcuts. If you start moving stuff around on us, we revolt. (See every UI update on an MS Office product ever.)

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

#328

Post by JHonaker »

jiggawatts wrote: post #216
IKantRead wrote: post #102

Earlier quoted context omitted.

> I hate the idea of having to hold a natural-language conversation with a computer in order to make use of its functionality. I hate the idea of having to use a mouse to click on a visual GUI to navigate a file system in order to make use of its functionality. It's less the case today, even among developers, but it wasn't that long ago that I remember that any serious technical user of a computer took it as a point…

> I hate the idea of having to use a mouse to click on a visual GUI to navigate a file system in order to make use of its functionality. Ironically, studies have shown that mouse-based interfaces are more efficient for practically all filesystem use-cases compared to CLI interfaces. Despite objectively faster-time-to-solution, people self-report that they "feel" that the mouse GUIs are slower. That's because there's…

I believe you, and I'm coming to this conclusion myself, after being a keyboard warrior for years.

Do you have any sources I can read though? I always find it fascinating when our perceptions are so off like this.

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

#329

Post by nuancebydefault »

sarchertech wrote: post #250
nuancebydefault wrote: post #200

Earlier quoted context omitted.

Most user interfaces already have a much finer granularity and number of options than your examples. When taking a shower, I would like fine control over the water temperature, preferably with a feedback loop regulating the temperature. (Preferably also the regulation changes over the duration of the showering.) Choosing to read the NY times indeed is only a few taps away, but navigating through and within its list o…

> When taking a shower, I would like fine control over the water temperature, preferably with a feedback loop regulating the temperature. (Preferably also the regulation changes over the duration of the showering.) That exists, but it’s expensive because of the electronics and mechanics involved. There are so many interfaces with this exact problem. You also almost certainly don’t want non-deterministic hallucination…

It would become less expensive, using less sensors and actuators, when using the predictive and learning abilities of an ai. You can, for safety reasons, keep a mechanical temperature limiter in the loop.

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

#330

Post by nuancebydefault »

bodhi_mind wrote: post #276
sarchertech wrote: post #250

Earlier quoted context omitted.

> When taking a shower, I would like fine control over the water temperature, preferably with a feedback loop regulating the temperature. (Preferably also the regulation changes over the duration of the showering.) That exists, but it’s expensive because of the electronics and mechanics involved. There are so many interfaces with this exact problem. You also almost certainly don’t want non-deterministic hallucination…

Maybe I misunderstood the op, but isn’t the mechanical faucet and our body’s temperature receptors this “fine control with feedback loop”?

Indeed, and to take the UI a step further, humans often prefer automation, if it works reliable. A complicated UI would become simple, just step into the shower.
Post reply on HN