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

The AI Battle

aifuture.substack.com

Post reply on HN
Print view View original

71–80 of 137 posts

Re: The AI Battle

#71

Post by htrp »

LeanderK wrote: post #56
stuckinhell wrote: post #13

AI that can do creative work is scary for its implications on humanity. Are we at best biological AI's ? I've played around with several of them. These artwork AI's are way more scary than GPT-3 et al because they seem to do something so creative with such good visible results. I don't think the genie goes back in the bottle. I think humanity has an existential crisis on its hands. Whats the point in doing things if…

> Are we at best biological AI's ? imho yes. But don't have an existential crisis just yet. It's a very narrow task where the AI systems do well in a way that we don't expected from computers. But it's well chosen, there are still only few tasks where significant progress was made. They really can't make "logical deductions" at all and we have no idea how. Everything you see is learned via a massive amount of labeled…

> We have no idea how to build a system that you could explain some basics on how to repair a simple machine and the system understanding what's going on.

90% of biological humans probably couldn't assemble Ikea furniture from the instructions either .

>I think it's a bit like the early mechanical age where people thought everything in the future will be mechanical (those fun retro-future posts).

I like this, and I'm definitely quoting it.

Re: The AI Battle

#72

Post by slowmovintarget »

lostmsu wrote: post #69
slowmovintarget wrote: post #67

Earlier quoted context omitted.

On "not the path to AGI": Gary Marcus on the Mindscape podcast is worth a listen. https://www.youtube.com/watch?v=ANRnuT9nLEE&list=PLrxfgDEc2N... Transcript: https://www.preposterousuniverse.com/podcast/2022/02/14/184-... "...And then there’s natural language understanding and reasoning, and I would say we have not really made progress at all. GPT-3, which we may wanna talk about, gives the illusion of having natural…

> I don’t really think that it does Either the quote is poorly chosen or reading this article is not worth the time.

That's a false dichotomy. Select the name Gary Marcus, right-click, and search (on my browser it defaults to Duck Duck Go, but that returns the right result).

The Mindscape Podcast is hosted by Sean Carroll. You have a very sharp quantum physicist interviewing an expert in the field of AI research.

The podcast is worth the time, and the quote is representative of an expert's take on the matter. He elaborates, but I don't need to write an essay just to argue on the internet.

Re: The AI Battle

#73

Post by YeGoblynQueenne »

carapace wrote: post #53
ajafari1 wrote: post #31

Earlier quoted context omitted.

Right on. When do you expect a productized AI that can replace your average skilled programmer?

Here's me almost exactly six years ago, predicting "most coding and design tasks will be automated within three to five years." https://news.ycombinator.com/item?id=11718300 If we count GitHub Copilot as the tip of the spear then I was too optimistic. > GitHub Copilot ... was first announced by GitHub on 29 June 2021 https://en.wikipedia.org/wiki/GitHub_Copilot

You were too pessimistic! Automatic programming, a.k.a. program synthesis, has been a thing since the early days of AI and computer science, for example the idea of deductive program synthesis, where a program is generated wholesale from a complete specification in a formal language that is not a computer language goes at least as back as Alonzo Church himself, in 1957:

https://en.wikipedia.org/wiki/Program_synthesis#Origin

Much work has been done since then and inductive program synthesis from incomplete specifications consisting of input/output examples (a form of machine learning) has been possible for quite some time. This is the second time this week I point to the report by Gulwani et al for a modern overview (of both kinds):

Program Synthesis

https://www.semanticscholar.org/paper/Program-Synthesis-Gulw...

Rather than Copilot being the "tip of the spear", it is really a step back, a system that can only generate code but not check its correctness, unlike pretty much every program synthesis system since forever. Although this may actually be an advantage for its commercial application (since it makes it easier to match expectations of the system's performance) it is not really any indication of progress, in any way, shape or form. In truth, Copilot is famous today because it is supported by a large company like Microsoft and because earlier work is not well-known by most people who get their AI knews from blogs and podcasts, who are also not very well aware of the history of the field. But, "tip of the spear"? Oh, no. Unless it's a very blunt, toy spear, more like a tech demo of a spear.

And this is by no means restricted to program synthesis, and Copilot. For example, the first self-driving car to drive on a stretch of real road, with real traffic, without human intervention, was Ernst Dickmann's 1995 robot car:

https://people.idsia.ch/~juergen/robotcars.html

Again: 1995. And yet, fully autonomous self-driving cars are still not here. The revolution hasn't happened and progress has only crawled marginally forward.

Re: The AI Battle

#74

Post by gdubs »

slowmovintarget wrote: post #10

This very much requires acceptance of the worse-is-better model. That "beautiful" drawing of an an astronaut riding a horse is aesthetically... crap. It may take a person 5 hours to paint or draw that, but, arguably, a person wouldn't. GPT-3 might generate a novel, but, generally speaking the prose is jarring and awful to read, because there is no intent (danger: philosophically loaded word) behind it. It's all autom…

You might be disappointed to find out the aesthetic tastes of _most people_ :)

Re: The AI Battle

#75

Post by ilaksh »

qualudeheart wrote: post #20
slowmovintarget wrote: post #10

This very much requires acceptance of the worse-is-better model. That "beautiful" drawing of an an astronaut riding a horse is aesthetically... crap. It may take a person 5 hours to paint or draw that, but, arguably, a person wouldn't. GPT-3 might generate a novel, but, generally speaking the prose is jarring and awful to read, because there is no intent (danger: philosophically loaded word) behind it. It's all autom…

I agree that the horse picture sucks. Other people might say that they like Dall-E’s output on Twitter, for signalling purposes, but I still think they don’t really believe that. Still, it’s only a matter of time till nicer graphics are possible. Most graphic designers and artists are still worse than this thing, especially the type that bought a $200 drawing tablet on Amazon and sell their services on Fiverr. Those…

How about now? Check out the Midjourney Discord bot. You can keep on doing variations and tweaks until you get one you like.

Re: The AI Battle

#76

Post by ilaksh »

CuriouslyC wrote: post #32
slowmovintarget wrote: post #10

This very much requires acceptance of the worse-is-better model. That "beautiful" drawing of an an astronaut riding a horse is aesthetically... crap. It may take a person 5 hours to paint or draw that, but, arguably, a person wouldn't. GPT-3 might generate a novel, but, generally speaking the prose is jarring and awful to read, because there is no intent (danger: philosophically loaded word) behind it. It's all autom…

I don't think these sorts of tools will ever be used to generate a novel or something on that scale. We're building really sexy autocomplete tools that creators will use to fill in the blanks much like the great masters of the renaissance used apprentices to do much of the work in their masterpieces. People will outline what they want, then ask the AI to fill in the blanks, and iteratively refine the result. As long…

Never is a long time.

Re: The AI Battle

#77

Post by ilaksh »

awillen wrote: post #44
CuriouslyC wrote: post #32

Earlier quoted context omitted.

I don't think these sorts of tools will ever be used to generate a novel or something on that scale. We're building really sexy autocomplete tools that creators will use to fill in the blanks much like the great masters of the renaissance used apprentices to do much of the work in their masterpieces. People will outline what they want, then ask the AI to fill in the blanks, and iteratively refine the result. As long…

Ever is an exceedingly long time. I would be shocked if 100 years from now, we didn't have AI-authored bestselling novels. Even then, they might not be literary masterpieces, but certainly AI will be able to write formulaic stuff that sells really well. If you don't believe that, just consider where technology was 100 years ago and what the response would have been if you'd described DALL-E in its current incarnation…

DALL-E is already been outdone my Imagen.

Re: The AI Battle

#78

Post by ilaksh »

stuckinhell wrote: post #13

AI that can do creative work is scary for its implications on humanity. Are we at best biological AI's ? I've played around with several of them. These artwork AI's are way more scary than GPT-3 et al because they seem to do something so creative with such good visible results. I don't think the genie goes back in the bottle. I think humanity has an existential crisis on its hands. Whats the point in doing things if…

We still have decades, maybe even a century or two, before humans become irrelevant. We should be happy we got our lives in right before.

https://marshallbrain.com/second-intelligent-species#:~:text....

Re: The AI Battle

#79

Post by ramoz »

tldr; I think this is just all an early "metaverse" enlightenment & look forward to a cool digital future.

1. Im not sure about all the doom & gloom. This more or less feels like a digital revolution... working mostly in style-transfer space myself. While AI-generated assets are awesome/compelling/shocking, the utility hasn't simply extended into complete automation of meaningful outcomes; so much as accelerated/enhanced digital workflows. & I think "meaningful" outcomes/digital creations will simply advance to require new skillsets.

2. A lot of engineering is still required in the applied domains. e.g. apply real style of some modern artist in new, generated, ways is rather complex and super manual (like re-training models built on wikiart, creating artist-specific datasets, etc). Even for pure-ai creation, the best generated artwork still requires a ton of nontrivial configuration, trial/error, etc... basically a whole new knowledge domain and skillset to acquire. CLIP/alike with diffusions setups have made things seem super simple, but a lot of work still goes into anything meaningful.

3. AI is a utility right now, and probably always will be. Adobe's integrations into tools like photoshop have been suite value-adds for designers (harmonization, color transfers, the big one imo- Super Resolution, and Im in-painting seems to be catching big steam)... Software eats software, photoshop will continue to photoshop and new capabilities will arise as old become automated.

Re: The AI Battle

#80

Post by TaupeRanger »

jaggirs wrote: post #46
TaupeRanger wrote: post #37

I'm not sure how the author could be so blind to the simple fact that computer generated articles and images are not useful. I mean if you want to replicate low quality blog spam and poorly written local news blurbs, maybe you could con a few companies into using your service. Or if you want to destroy the stock image market or the album art market (maybe that kind of "disruption" is somehow beneficial to you?) I gue…

I am not sure how you can be so derogatory yet apparently so ignorant. Have you even seen what DALLE-2 (or imagen) can generate? DALLE-2 does not have to 'prove itself' it is clearly on-par with human designers in terms of quality and price. There even is a product on the market right now called Midjourney that is worse in quality than DALLE-2 yet is growing incredibly fast and already breaks even economically. If NL…

DALLE-2 is not in any way "on par" with human designers, unless they are designers of the things I already mentioned like album art or stock photos. Even with the latter it's not particularly good unless the subject involves something that doesn't require fine detail to look decent. Midjourney is not breaking even because it creates useful, high quality art. It is breaking even because of the novelty of DALL-E and I don't think it will last particularly long in the spotlight.
Post reply on HN