> It notes that interactions with enemies need to be improved, since they will often appear fuzzy, and that because its current context length is 0.9 seconds of gameplay (9 frames at 10fps), it will forget about objects that go out of view for longer than this.
Checking out the actual video in the tweet was more impressive than this description setup for me. Definitely more “tech demo” than “game”, but pretty impressive.
Side note —- what an irritating way to put an article together:
- Don’t actually embed the tweet in question that contains the demo video, or even mention there’s a video
- Focus on a few negative replies to the tweet from random people
- The biggest piece of media on the page is a screenshot of a tweet from Tim Sweeney without any context of who he is, or that it’s a reply to the tech demo…
But I guess I clicked on the link, read the article, and gave a bunch of ad impressions, so I’m part of the problem!
Since you had already figured out the gist, I was hoping you'd have shared the demo link, so I don't have them ad impressions! But I notice a YouTube link below so I'm going there instead. :)
> but the fact that it can reliably do this at all is extremely impressive
The better question is could you achieve the same thing using much less computing power and simpler methods? Would a Markov chain be able to do the same thing? If so, I'd have to with the articles sentiment, this burns a bunch of electricity to do something with extreme inefficiency and waste.
This is absolutely incredible: surprising amount of consistency and the whole thing live-generated. It’s incredible how quickly we can get inured to technology that would have been sheer magic when I was younger.
I recall being blindingly excited about Windows 98 and a gradient title bar but these days everything new I see is incredible and everyone is bored with it. I don’t get it. The game is being generated frame to frame!
Hard disagree. I found it to have a surprising amount of inconsistency. All I needed was to look completely up (or completely down) then restore the view and I was in a completely different setting.
The world is. I would struggle to call this a "game" in the sense that it would be highly rewarding to play. It's also impossible to compete with your friends for high scores or best times. It completely ignores everything about what human "games" are and instead focuses on what the "billion dollar copyright theft machine" can crank out this week.
> and everyone is bored with it.
It's almost like the people pushing this technology are not human. Or don't live human experiences. I can't rightfully apprehend how they thought this would be received well.
Perfection is lots of little things done well. The current AI slop is lots of things done poorly. And if you squint your eyes and tilt your head you can _almost pretend_ it's something novel.
I know some may will decry it for what it stands for, but I think this is great for edge case mitigation, imagine civilization games where you can use flexible strategies mimicking real life, versus following programmed decision trees. Gameplay that is responsive to environmental destruction, agent cores in the console that are constantly scheming against you. If done right, we can continue to do new and innovative things. One can get bogged down by the bad 3d animation out there, or marvel at what Pixar and before it, films like "Flight of the Navigator" and "Terminator" did.
This is beyond that, a lot of people are uncomfortable with computer algorithms mimicking human ingenuity and creativity, even at hackernews where we're all trying to figure out how to get rich from it.
Or the camera - think about how it puts all those portrait painters out of a job, and it makes anyone who can click a button think they’re some kind of genius artist, and it’s ruining the environment with all the toxic waste produced from developing the film
I wonder whether it'll be possible to compress enough of the game to make (almost) every possible scenario that you could encounter in the game be playable. Same issue that the previous AI experiment for Minecraft and others had is that objects and enemies seem to pop in and out of nowhere. Could the "learned" probability be high enough for this never to be an issue? You ever think you're seeing something in real life but it's just an optical illusion, it kinda feels like that to me. Obviously this still requires an entire game to be made before you can train on it, but could maybe open up other development and testing of games.
> Obviously this still requires an entire game to be made before you can train on it, but could maybe open up other development and testing of games.
The idea of developing a game where the "code" is totally opaque and non-deterministic honestly sounds like an absolute nightmare. How would you even begin to QA something like that?
> The idea of developing a game where the "code" is totally opaque and non-deterministic honestly sounds like an absolute nightmare. How would you even begin to QA something like that?
I have a fear that we are going to experience a significant regression in our ability to develop software as new "programmers" normalize the idea of "generating" "code" this way. Some kind of dystopian future where people who think an "is-negative" module is a good idea, but coupled with that module having been "generated" by "AI". Bone chilling.
Re: QA
Clearly we just need another generative "AI" to act as QA in an adversarial capacity to the "AI" generating the "code“. Turtles all the way down.
This proposed direction is even worse than generating code, it's eliminating code altogether. The project "source" would just be a big blob of weights that you indirectly prod and poke until it hopefully does what you want, and nobody could understand exactly what's going on under the hood even if they wanted to.
Computer "programs" being big hairy balls of "intent" derived from some corpus of inputs and a prompt is horrifying.
I kind of wish hardware hadn't gotten fast enough to enable this future. Humans being lazy, as they are, and the output of this kind of horror show eventually being "good enough", this is going to get normalized.
Anybody working to enable this future is, to me, acting unethically. This is the "AI" apocalypse I'm worried about, not the AGI singularity fever dreams that garner headlines.
Worse yet: that big blob of weights only works at all because it's been trained on a huge corpus of data from existing games. Doing anything actually novel in a game - like implementing new game mechanics, or even using a distinctive art style - would be next to impossible.
Well, nobody seems to have figured the "generating" part out yet. What we've actually seen so far is crappy approximations of games which already exist (Doom, Minecraft, Quake 2).
Funny you should say that, because the AI Doom demo was 8 months ago and in that time we seemingly haven't got any closer to making novel games. This Quake 2 demo is yet again just a wonky version of a pre-existing game, attempting to be a 1:1 copy of the original, rather than extrapolating something like Quake 2 that isn't literally Quake 2 again.
At this point, HN is the only place I see positive reception of AI products. Everywhere subreddit or gaming forum or social media I follow is overwhelmingly negative. Usually characterized as slop, stealing, low quality, or stealing jobs from craftsmen.
I’m not saying AI is bad, but people are very strongly opposed to these kinds of products. In that sense, HN is the only place that seems to resonate with those.
> It notes that interactions with enemies need to be improved, since they will often appear fuzzy, and that because its current context length is 0.9 seconds of gameplay (9 frames at 10fps), it will forget about objects that go out of view for longer than this.
Checking out the actual video in the tweet was more impressive than this description setup for me. Definitely more “tech demo” than “game”, but pretty impressive.
Side note —- what an irritating way to put an article together: - Don’t actually embed the tweet in question that contains the demo video, or even mention there’s a video - Focus on a few negative replies to the tweet from random people - The biggest piece of media on the page is a screenshot of a tweet from Tim Sweeney without any context of who he is, or that it’s a reply to the tech demo…
But I guess I clicked on the link, read the article, and gave a bunch of ad impressions, so I’m part of the problem!
Link to video of demo: https://x.com/geoffkeighley/status/1908593030141202635?s=46&...
Link to the demo to try yourself: https://copilot.microsoft.com/wham?features=labs-wham-enable...
Since you had already figured out the gist, I was hoping you'd have shared the demo link, so I don't have them ad impressions! But I notice a YouTube link below so I'm going there instead. :)
Realized the hypocritical nature of my post and updated with the demo link! Copying it here too:
https://copilot.microsoft.com/wham?features=labs-wham-enable...
Currently this is a fever dream of non-Euclidean space - but the fact that it can reliably do this at all is extremely impressive.
> but the fact that it can reliably do this at all is extremely impressive
The better question is could you achieve the same thing using much less computing power and simpler methods? Would a Markov chain be able to do the same thing? If so, I'd have to with the articles sentiment, this burns a bunch of electricity to do something with extreme inefficiency and waste.
> The better question is could you achieve the same thing using much less computing power and simpler methods?
If only there were a way to play Quake 2 on a 90mhz Pentium with 16MB of RAM...
This is absolutely incredible: surprising amount of consistency and the whole thing live-generated. It’s incredible how quickly we can get inured to technology that would have been sheer magic when I was younger.
I recall being blindingly excited about Windows 98 and a gradient title bar but these days everything new I see is incredible and everyone is bored with it. I don’t get it. The game is being generated frame to frame!
> surprising amount of consistency
Hard disagree. I found it to have a surprising amount of inconsistency. All I needed was to look completely up (or completely down) then restore the view and I was in a completely different setting.
> The game is being generated frame to frame!
The world is. I would struggle to call this a "game" in the sense that it would be highly rewarding to play. It's also impossible to compete with your friends for high scores or best times. It completely ignores everything about what human "games" are and instead focuses on what the "billion dollar copyright theft machine" can crank out this week.
> and everyone is bored with it.
It's almost like the people pushing this technology are not human. Or don't live human experiences. I can't rightfully apprehend how they thought this would be received well.
Perfection is lots of little things done well. The current AI slop is lots of things done poorly. And if you squint your eyes and tilt your head you can _almost pretend_ it's something novel.
Of course it earns yawns.
I know some may will decry it for what it stands for, but I think this is great for edge case mitigation, imagine civilization games where you can use flexible strategies mimicking real life, versus following programmed decision trees. Gameplay that is responsive to environmental destruction, agent cores in the console that are constantly scheming against you. If done right, we can continue to do new and innovative things. One can get bogged down by the bad 3d animation out there, or marvel at what Pixar and before it, films like "Flight of the Navigator" and "Terminator" did.
> these days everything new I see is incredible and everyone is bored with it
Information saturation
This is beyond that, a lot of people are uncomfortable with computer algorithms mimicking human ingenuity and creativity, even at hackernews where we're all trying to figure out how to get rich from it.
Perhaps if the transistor was invented today, some people would be disgusted at how it mimicks the ingenuity of humans calculating
Or the camera - think about how it puts all those portrait painters out of a job, and it makes anyone who can click a button think they’re some kind of genius artist, and it’s ruining the environment with all the toxic waste produced from developing the film
I would be dubious of the utility if it calculated imperfectly and in a way that couldn't be understood and characterized by people.
If the advice to the public given by experts in the field was that this lack of determinism and understandability was acceptable I would be disgusted.
It won't load for me, but here's a video of someone playing from earlier today:
https://youtu.be/m8Qjg7xJ2SA?si=bdcBYpz3QLK24AVR&t=151
I wonder whether it'll be possible to compress enough of the game to make (almost) every possible scenario that you could encounter in the game be playable. Same issue that the previous AI experiment for Minecraft and others had is that objects and enemies seem to pop in and out of nowhere. Could the "learned" probability be high enough for this never to be an issue? You ever think you're seeing something in real life but it's just an optical illusion, it kinda feels like that to me. Obviously this still requires an entire game to be made before you can train on it, but could maybe open up other development and testing of games.
> Obviously this still requires an entire game to be made before you can train on it, but could maybe open up other development and testing of games.
The idea of developing a game where the "code" is totally opaque and non-deterministic honestly sounds like an absolute nightmare. How would you even begin to QA something like that?
> The idea of developing a game where the "code" is totally opaque and non-deterministic honestly sounds like an absolute nightmare. How would you even begin to QA something like that?
I have a fear that we are going to experience a significant regression in our ability to develop software as new "programmers" normalize the idea of "generating" "code" this way. Some kind of dystopian future where people who think an "is-negative" module is a good idea, but coupled with that module having been "generated" by "AI". Bone chilling.
Re: QA
Clearly we just need another generative "AI" to act as QA in an adversarial capacity to the "AI" generating the "code“. Turtles all the way down.
"The Machine Stops".
This proposed direction is even worse than generating code, it's eliminating code altogether. The project "source" would just be a big blob of weights that you indirectly prod and poke until it hopefully does what you want, and nobody could understand exactly what's going on under the hood even if they wanted to.
Computer "programs" being big hairy balls of "intent" derived from some corpus of inputs and a prompt is horrifying.
I kind of wish hardware hadn't gotten fast enough to enable this future. Humans being lazy, as they are, and the output of this kind of horror show eventually being "good enough", this is going to get normalized.
Anybody working to enable this future is, to me, acting unethically. This is the "AI" apocalypse I'm worried about, not the AGI singularity fever dreams that garner headlines.
Worse yet: that big blob of weights only works at all because it's been trained on a huge corpus of data from existing games. Doing anything actually novel in a game - like implementing new game mechanics, or even using a distinctive art style - would be next to impossible.
Wasn’t is-negative the dystopian past of ten years ago?
I think it's pretty cool you can start generating games this way
Well, nobody seems to have figured the "generating" part out yet. What we've actually seen so far is crappy approximations of games which already exist (Doom, Minecraft, Quake 2).
This seems such an out of place comment when looking back at the last year's progress in generative AI
Funny you should say that, because the AI Doom demo was 8 months ago and in that time we seemingly haven't got any closer to making novel games. This Quake 2 demo is yet again just a wonky version of a pre-existing game, attempting to be a 1:1 copy of the original, rather than extrapolating something like Quake 2 that isn't literally Quake 2 again.
Genie was generating platform games based on a sketch in Feb 24..
https://www.technologyreview.com/2024/02/29/1089317/google-d...
From the way Satya explains it, you can’t, but Microsoft plans to.
At this point, HN is the only place I see positive reception of AI products. Everywhere subreddit or gaming forum or social media I follow is overwhelmingly negative. Usually characterized as slop, stealing, low quality, or stealing jobs from craftsmen.
I’m not saying AI is bad, but people are very strongly opposed to these kinds of products. In that sense, HN is the only place that seems to resonate with those.
You should look at more suitable metrics than reddit or gaming forums.
Don't see reasons not to trust this source: https://explodingtopics.com/blog/most-visited-websites#top-1...
ChatGPT.com is #9, above whatsapp.com, amazon.com, and netflix.com
You're conflating public sentiment and internet traffic. You're using the wrong metric to make your point.
No, I'm just valuing actions over words.