wbk ᕦ(ò_ó )ᕤ

@ket74384019 @SebAaltonen @JeremyNguyenPhD @bensig Natasha is a known talented graphics engineer Unity/Bungie/AMD etc so "She announces the speakers..." is grossly unfair. Sure at that level technically the industry is still majorly male dominated, but there are others equally as talented, even more so outside games.

@SebAaltonen @JeremyNguyenPhD @bensig I scanned through that website. She announces the speakers and gives a generic talk on 20y history of the conference remembering the most notable papers, among which zero were by women. So where are those female Quilez, Aaltonen, Muratori, Hashimoto, Tsoding, at least Primagen?

Milla Jovovich (actress from The Fifth Element) created a world-beating Claude memory system with @bensig?! - 100% on LongMemEval — first perfect score ever recorded. Free and 100% open source. Github link in the quoted post from Ben. I'm keen to hear how it works for you.

@ducha_aiki @SebAaltonen I don't think anyone was arguing that better products won't do well. Only that they are not guaranteed too, even if they stand a better chance of doing so - and discussing to what degree that balance may shift in the near future.

@wizardbrainz @SebAaltonen Right, that's why 7 years of Silksong development didn't pay off, and it is overshadowed by AI slop.

Unpopular opinion: Nobody will make money by single shotting a game or app. There's always somebody else who is willing to put in more effort. You spend 15 minutes, they spend a whole week. Of course their product is better. You beat their 1 week product with 1 month product, etc

@SebAaltonen @longiy I agree in general currently. But do you think AI could learn what makes Megabonk or Vampire Survivors fun? I actually think that is not far off.

@wizardbrainz @longiy Steam top 20 games have been making almost half of the total revenue on the platform. Top games attract most players. I don't think a poor quality single shot game will have any chance. Especially today when most top games are big multiplayer games with social connections.

@longiy @SebAaltonen Yes I think the Pareto Principle will remain undefeated somehow.

@wizardbrainz @SebAaltonen Things that are abundant and easy will always have less perceived value, so the moat is still going to be some sort of skill which is multiplied by the ai but proportionally things will stay the same. You can be Shein, H&M or Hermes

@SebAaltonen @longiy Yeh I didn't mean an exact equivalence. As you say scarcity etc plays it out differently. It's horrifying if gamers have to play bad games as they can't discover the good! Despite nearly 2000 games on Steam I can't finish this lifetime, I still for now buy new ones on sale.

@wizardbrainz @longiy That is possible, but fast fashion (Shein, etc) is significantly cheaper than real clothes. Digital products are almost free to copy. So many free-to-play high quality products and so many $10 AAA Steam sales. You can't sell bad quality products any cheaper than $0.

@longiy @SebAaltonen The majority of games, even the less commercial creative ones are products though - whose utility is measurable. I see @SebAaltonen thinking as product centric too btw. And it's looking like AI will sail past the slop to the Uniqlo/H&M fast fashion stage very soon.

@wizardbrainz @SebAaltonen Well, making slop is inherently unfashionable and not trendy in all creative endeavours. You are applying product thinking to art (yea heayh not all games all art yaddayada) which doesn't work because utility is not very well measurable.

@SebAaltonen I'd caution that the future definition of a game/movie/youtube/entertainment is going to be very different from ours - kids born today will see them as facets of the same thing. I don't think that invalidates the core of what you are saying though.

@wizardbrainz Also if you don't like playing games, then it is unlikely that you like playtesting your own game either. And without playtesting it is very unlikely that your game is going to be fun.

@SebAaltonen I think we are just now entering another 'social casualisation' step with apps/games in general - maybe HypeHype was orientated a little early. We basically agree restricted to the traditional Steam/Console gaming audience of course!

Yes, there might be some exceptions, but in general, you are likely a gamer, game dev or game influencer. If you don't have any gaming connections and don't like playing games, it's very unlikely that your single shot game project will succeed. If you don't play games, how do you know how to build a good game?

@SebAaltonen "A random non-gamer can never make a popular game." - I don't think that is true unfortunately ;-) Unlikely to make a good one sure! Again to my point of apps becoming fashion trend items, it doesn't require them to actually be good to be popular...

Yeah. Motivation is a key. And now that AI makes software/game dev available for a broader audience, motivation matters even more. Some gamers are super into a genre and know what matters and are willing to playtest their game like crazy (they play games anyways). A random non-gamer can never make a popular game. If you don't play games, you don't have the vision and don't have the motivation to push it forward and to do all the necessary live-ops to maintain it.

@SebAaltonen I would say you actually furthered my point. With so much content the problem becomes discovery which is led by key influencers and fashion trends.

@wizardbrainz So far I have only seen sloppy AI games. We made a UGC game platform with 500,000 games on it. But they were slightly bit worse than games on competing platforms, so nobody cared about them. Why would anybody waste time on playing bad content when so much good content exists?

@SebAaltonen I certainly don't disagree with this part btw. And motivation is a key differentiator. 'Things becoming easy with AI' still doesn't replace that, if anything it could have the opposite effect.

Iteration is the key especially in game products. AI can't find you the fun, you have to test the game repeatedly and iterate to improve it. AI can't replace your motivation either. Shipping a high quality game requires a lot of work. 15 minutes of motivation isn't enough.

Down the Apollonian fractal sphincter we go. Cached temporal GI experiment at 4k with no denoising.

@SebAaltonen @int_16h At some point do we want to be plugging our input devices directly into the GPU eventually?

@int_16h Frame generation could theoretically improve the latency same way as async time warp improves latency on VR headsets. Write mouse data at 1000Hz in background thread to persistently mapped buffer. Frame generator reads mouse movement. Minimizes mouse look latency.

6X, 10X, 100X, does it even matter? No one cares about frames per second. What matters is latency. Frame generation does not improve latency.

@timsoret @mrdoob They certainly are not on the path to form happy Amish communities

I follow many anti-AI accounts just for inspiration for my game. It’s fascinating to see their wishful thinking proven wrong in real time, their frustration growing further week after week. I’m waiting to see them wake up, but no, the delusional denial is becoming more & more desperate. I think many are on the verge of radicalization. AI entrepreneurs & engineers are in great danger.

Most anti-AI accounts are just monetizing outrage. They don't care about the tech, they just know anger pays. Their followers don't realize they're being farmed for $500-$1000/month in Twitter payouts. Don't bother arguing with them, you're just funding the grift.

@timsoret @SebTellier Noice intro

Going to see the great @SebTellier in concert at Koko tonight. Difficult to explain how important his music is to me. I wish his work was more known outside of France.

@SebAaltonen Regardless of peoples (currently valid) complaints we have come a long way since pong!

Next level uncanny valley :D

@Grummz Do the drivers still crash your PC?

Believe it or not, they just dropped a new Sound Blaster card.

@matiasgoldberg I researched this a couple of years ago. Came to conclusion it is the future, but it is going to have to work a bit differently...

I've been researching the same thing. That's a lot of user rejection though. But ONLY APPLY IT TO THE CHARACTERS. DLSS5 seems to be applying it to the whole scene. The style controls are off for all games except the EA FC and Resident Evil one, which are the most impressive.

Procedural grass #gamedevlife