Does anyone else feel like most of this AI buzz is destined to go the way of voice recognition? When that tech was new it was paraded around like gods gift to man, but when the high wore off we realized it was just a very shitty button that sometimes did the thing when you pressed it. ChatGPT wordage is okay at informing you I guess. It's extremely verbose, and kind of answers some questions, but using a proper search engine and applying my own critical thinking to get results feels so much better-- the same way it feels so much better to just use your phone/computer vs a voice assistant. Tech is cool though so it's fun to watch
No, we've stumbled across the next industrial revolution and perhaps the next step function evolution of our species (or whatever comes after, "alignment" folks be damned).
I've been bearish about so much. Pointless crypto, tech being cellphone incrementalism, slow speed of basic science, etc.
I'm using this tech to make so much stuff that I could never produce before. It's incredible.
Take my judgment with a grain of salt. I'm in the thick of it and building in the space. That said, never have I ever been so excited.
The statement that tech eats the world? We'll, AI eats tech. And us.
No this is not yet the next industrial evolution. ChatGPT is still at the beginning of its hype curve. It is amazing, however, it is just a language model. It can't reason, just regurgitate.
I’ve been around this stuff for 7 years now, former co-founder was super into AI hype and actually ended up leaving to go work for OpenAI. Nothing is snowballing fast now… it’s been the exact same since at least 7 years ago.
At the time I have to admit I was equally bullishly myopic, the new models seemed insanely cool and the process seemed super-exponential. I predicted AI may replace programming, art and many things within 5 years.
Then I checked in 5 years later - not much really had changed and I revised my predictions way down. Since then I’ve revised them down again actually, as I see them running into scaling issues again, amongst many other issues.
The attitude you take is infectious, I saw it firsthand. But if I look at what concrete is changing, it’s not much.
Humans love watching Chess. It’s exploded in popularity in the last 4 years. No one watches AI chess, though the top players certainly gain in their practice from it. But where’s the revolution?
I know it’s different but I predict the same happens across the “new” areas - image and text generation. They seem amazing, likely beating humans in many ways. But in the end we don’t care much about computers, we care about novelty.
Novelty and humanity drive fashion, music, film, literature… basically all art. It needs to innovate, needs to be extremely personal. It’s not going to disrupt any of them, not even close. But it will be used in the loop.
What is revolutionary is that ChatGPT is indistinguishable from a real person's writing and intelligence on a subject - and not just in text, but in coding. Before ChatGPT any similar program's responses were so obvious bots that you would look at and say "oh that's cool" but it was still garbage you would never use. Now, I've heard people using ChatGPT for self (ai) therapy, asking it what to do, and how to do something. It's a planner, a worker, a designer, an idea machine. It's like Google but actually gets you to the answers you want without the incredible about of garbage that came when SEO websites took over Google. Then you've got the symbiosis with humans, it's not "look at this AI art." It's artists using AI tools to 'produce' 1000 versions of art a day and pick the best one they find most personal. AI art was already entered into a human contest by someone cheating in which it won. Before iPhone there was plenty of similar products but none that put all the pieces together the way an iPhone did. There is a tipping point to AI and ChatGPT is the iPhone of AI.
I don’t see ChatGPT as any different from gpt3 really, it still makes all the same category of mistakes.
It’s very distinguishable from a human in that you can trivially make it contradict, say nonsense, and it still makes all sorts of errors especially past a few paragraphs.
Very few people are using it. AI therapy is really a great counterexample - it’s not just not up to the task, also highly dangerous. Last thing anyone needs in therapy is to talk to a computer that has no long term memory and will say whatever you want it to - pure demoralization.
As far as code - also not really compelling. I’ve tried it and don’t use it, because:
1. It’s worse than Google. Google + stack overflow finds 100% of the same answers, but also gives me many alternatives, verifies they are correct, and adds lots of context.
2. It’s incorrect in subtle ways, often.
3. It’s actually actively bad at programming beyond the simplest smallest isolated problems, which are about 1% of the daily tasks of a software dev.
I actually think the confusion of thinking it’s good at programming and art to be fundamental - it’s good at rough things not precise things. So art is more in its wheelhouse and why people are so interested in it. It’ll do some interesting stuff there.
But it’s very much not precise, it sucks at math, it’s bad at programming for the same reason. In a very very local scope for a simple type of problem constrained only to that local scope it’s ok, but that’s just not really programming.
I did say in the loop it has some value, but it’s truly far from an iPhone in terms of value. About 0% rounded of the population is even aware of it, let alone getting value from it. I have no idea what I’d use it for, its proneness to blunders and bullshit is a real dealbreaker for every application besides illustrations and maybe summarization - and the latters been solved forever, and is a marginal industry.
Again there’s a weird need to hype it up it seems, we’re at some peak of the hype curve and people are buying it. I think one reason that happens is because it’s mediocre at a lot of things, so people think it’ll do everything. But it’s got major reasons why it’s mediocre that haven’t changed from gpt3 or earlier.
It’s a cool thing certainly, but Google is like 1000x more valuable and idk if even Google was a revolution or just a somewhat big improvement.
Four fundamental limiting factors: it’s imprecise (can’t rely on it for unsupervised or technical things), it’s average (not very valuable to art or creative innovation), it’s a liar (not great for many things), and it’s a panderer (therapy, etc).
1. You could literally just input stack overflows data and it would be able to answer the same questions. ChatGPT is running on years old data and isn't even connected to the internet right now, the future will be way better.
2. Again Google is also wrong in subtle and even not so subtle ways. I've seen highlighted answers that are just dead wrong thanks to the SEO takeover. In Google clicks = right answer very often. It's essentially a popularity contest and try finding anything obscure on Google.
3. It's not going to write entire programs but it's amazing for simple one module tasks, but sometimes that is all you need to build into something bigger. You can even ask it tell you how to program and continually ask 'and how do I do that' and find the answer (similar to google which you'll be also be lucky to find anything useful on your first search or first link). It's more like an instructor than Google because with Google you don't know what you don't know. Most programmers are actually just good at searching for answers and this is a short cut. You can get 10 ways to program the same thing if you ask the question slightly differently. Just like I was talking about art, you get 100 solutions if you want and just choose the best. It's not replacing programmers but it sure as hell makes the job easier.
I do get that, but to me it all sums up to “much worse than Google in the best case”.
StackOverflow is almost never wrong at the top answer and again has many coherent responses for many use cases with contextual help, where GPT is often subtly incorrect with no real context (precision struggles). The whole iterative thing is maybe good for learning tools. Just not seeing a revolution, yet.
The struggles for SEO are less than those for GPT, both have to deal with spam and weird data, but GPT you’re also out of date, don’t have much explainability, and basically trust one giant black box of bias that requires higher input-foo whereas Google is also a giant black box of bias but at least you land at SO or Reddit or something where things are concrete and easy to understand.
Again, if this was some iPhone level revolution I’d have no comparable. It’d be way better at something clearly. But it’s not even clearly equal to any of the examples, and imo worse than basically all except “draw me a dragon in a tutu”, caricature artists should be worried!
On second thought, caricature artists are pretty safe since much of that experience is the whole human drawing you thing and also the generative models don’t really do caricature well as of now.
All of those examples have the exact same limitation, they can't replace processes that create novel ideas since they're statistical recombinations of existing ones.
Because it has no mechanism for creating novel ideas...
It's like saying a calculator could spontaneously derive its own proof. No it can't, because it has no ability to do that. It physically does not have a way to create new ideas. It cannot reshape its own circuitry like a brain and it does not have spontaneous thoughts.
ChatGPT is not like other chat bots which operate like a calculator and that is exactly why it is popular. You can put random inputs like "write a short story with 900 words with a bear and wolf about driving to the casino where they win big but can't spend the money" and it will output something total unique based on those inputs.
I hear what you are saying, but I'm not convinced. Can you dig a bit deeper?
What makes an idea "novel"? What types of physical processes can create new ideas or have spontaneous thoughts, and which can't? What is "special" about brains, in your view?
But none of those products solve the problem of the "AI" being stupid as fuck. Because it's not AI, it's just brute forcing.
The current solution of 500b parameters or whatever is literally just a lookup table of existing data. They could replace it with an SQLite database with the same amount of data and get the same solution.
There is no thought, there is no reasoning. It can't come up with new ways of doing anything, it just regurgitates based on its pre-programmed ruleset. That's why self driving doesn't work - it doesn't know what a fucking car is. Ask it what a road is or a car and it has no idea, it just has a million miles of road in its database. Ask it how to drive on a country lane and it crashes because that's not in the database. There's nothing there.
I don't know about that - I've spend 100+ hours asking it to write Plays, TV Scripts, Poems, and lots of code (with Unit tests) - and it's pretty remarkable. Makes far fewer errors in simple stuff than I would - and way faster - I just need to proof it and copy-paste a lot of the time.
What I find amazing is when I ask it to do something like "Write a TV Script about Malcom Saving Kaylee from Aliens" - and then it fills in all the details of the Serenity, places Kaylee in the engine bay , has Mal firing off his pistols, and then retreating back to the Bridge".
Then, I simply type, "Add Elves" - and it rewrites the scenario, with the Elves properly using bows to fight off the aliens, and (this blows my mind), when they retreat to the bridge and Kaylee is worried about the engine, the Elves coherently say that it's okay, that they will use Magic to fix the engine and make it faster - And everything is written far more coherently than I ever would and makes absolute sense.
I get that it's a LLM - but damn, it's been blowing my mind for over a month, and I'm still astonished by it every day. I'm starting to wonder if my own intellect is anything more than an LLM connected to emotions/sense-organs.
> The current solution of 500b parameters or whatever is literally just a lookup table of existing data. They could replace it with an SQLite database with the same amount of data and get the same solution.
I don’t have a deep understanding of how these language models work, but what do I know makes think that saying “It is basically a lookup table” is sort of like saying “A car is basically a bike”.
Sure, there are some very vague similarities, but they are also incredibly different.
A bike and a car both take you places. But the kinds of places they can take you and how effectively they do so differs vastly.
Likewise, both lookup tables and LLMs take an input and then spit something out. But the similarities basically stop there.
That does not mean it's not terribly useful. The computer that helped man land on the moon couldn't reason either but it had a good enough model of the physics and a brand new way to fuse a bunch of sensors into a model of where the craft was. I think of these products as being tools guided by the smart human.
Pretty much this. LLMs are still limited, perhaps by design but it is so. I recently summarised an article by Maggie Appleton on some of the shortcomings of the model and the implication here: https://twitter.com/kirso_/status/1610849402557722625
disclaimer: I am not the author of the original article.
Several examples:
- Language models cannot reason, consult external sources, or sense the world as we can. They also lack access to a shared reality, embodied experiences, and the ability to tell stories grounded in culture, such as referencing obscure concepts or interests.
- Current LLMs fall short due to recycled ideas, unoriginal explanations and not having an opinion. Excellence in writing (and thus thinking) requires a greater sense of critical reasoning, scepticism and going beyond the status quo and comfort.
I'm pretty bullish on it too. I'm not sure it's the next industrial revolution, or at least not yet, but the utility of it while working on my current project has been huge.
It's not anywhere near the point where it can write the app for me, and it has a bad habit of being confidently wrong. But I've been able to use it extremely effectively to cut down on the amount of time it's taking me to learn new things as I build.
I'm a fan, and hope that the next iteration feels like another leap the same way ChatGPT does from GPT3.
ChatGPT/Stable Diffusion are like the Newcomen engine of our time. It took 100 years for the world to see what the Newcomen engine promised, but I think we'll see it in 30. AI is going to totally reshape content creation - what it means to be a creator is going to be completely different for the next generation.
FakeYou.com is my "pandemic project" that lept to millions of users. People make music and memes with it. It's at the low end of creation, but still provides value.
We've built voice conversion and singer vocalist conversion. We solved the real time case first, ironically, and are waiting to finish the web version (bigger reach/audience/impact) to launch them together. It'll be about a week longer. You can turn Taylor Swift songs into Freddy Mercury songs. It's wild. (You can also do this with your own voice!)
We're going to lean heavy into music production for non experts. Stems, instrumental generation, full synthesis, virtual bands.
But the great thing is that I'm a filmmaker. I have another subteam working on mocap, NeRFs (previously live photogrammetry in unreal), and LLMs. We have something really special launching here soon too. Find my Twitch channel. :)
100% of this feels new to the world. And it's building like a tsunami.
The reason why I was asking is I've seen claims thrown around a lot about automation revolutionizing manufacturing, but that revolution has already been underway for eight decades with notable companies moving away from from it because humans are several orders of magnitude more versatile than any machine. Don't get me wrong, there's a lot of room for improvement to make machinery smarter, but I'm just not seeing anything like an industrial revolution coming because of it.
Yes the hype cycle has had gasoline poured on it (to mix metaphors). AI hype already came and went and it seemed like we were settling in to the "plateau of productivity" where we'd get some post hype applications. Now all that's been blown up and we're right back to the overblown "it's the next electricity" stuff. I think expectations are going to get deflated a lot faster this time, given how quick the rise was.
I mean, doesn't being a Bing feature pretty much speak for itself
No. Voice recognition is useless for me. Why talk to a computer when it makes mistakes getting correctly everything I say. Imagine an office were everyone is talking to their computer. It's noisy and silly.
With ChatGPT where it finds answers where a normal Google search is not very helpful, I find ChatGPT very helpful. I pasted some code I didn't understand to it and it explained to me what it does. That was very cool. Google would not have been able to help me.
