NLeseul
@nleseul.bsky.social
I play video games, wear shorts, and think about stuff.
Pinned
NLeseul
@nleseul.bsky.social
· Sep 9
Never attribute to malice or stupidity what can be adequately explained by fallible humans trying their best to navigate a confusing and broken world that they were never meant to inhabit.
Apparently there's been some real progress on the "Mario benchmark" lately too! youtu.be/uAjyS1-HP3w?...
(Still not a "success" by my criteria, since this system seemingly doesn't use "sticky actions" and is basically just a TAS generator, but it's still the best I've yet seen!)
(Still not a "success" by my criteria, since this system seemingly doesn't use "sticky actions" and is basically just a TAS generator, but it's still the best I've yet seen!)
I Trained an AI to Beat Mario’s Fastest Speedrun
YouTube video by A Goofy Intelligence
youtu.be
September 27, 2026 at 4:28 PM
Apparently there's been some real progress on the "Mario benchmark" lately too! youtu.be/uAjyS1-HP3w?...
(Still not a "success" by my criteria, since this system seemingly doesn't use "sticky actions" and is basically just a TAS generator, but it's still the best I've yet seen!)
(Still not a "success" by my criteria, since this system seemingly doesn't use "sticky actions" and is basically just a TAS generator, but it's still the best I've yet seen!)
Sure, you could just do things. But why bother, when you have a magic box that will do all the things for you?
September 27, 2026 at 3:04 PM
Sure, you could just do things. But why bother, when you have a magic box that will do all the things for you?
At last—progress on an AI benchmark that actually matters! www.youtube.com/watch?v=mBal...
I Gave an AI Zelda and One Rule: Don't Cheat.
YouTube video by Bears Gaming Den
www.youtube.com
September 26, 2026 at 11:31 PM
At last—progress on an AI benchmark that actually matters! www.youtube.com/watch?v=mBal...
AI safety orgs (screaming): "The Ring of Power is sooo super-powerful! You wouldn't believe how powerful it is! Anyone who doubts its power is an idiot! (Also no mortal should ever possess it.)"
Also AI safety orgs: "Why is everyone suddenly racing to claim the Ring of Power??"
Also AI safety orgs: "Why is everyone suddenly racing to claim the Ring of Power??"
September 24, 2026 at 7:08 PM
AI safety orgs (screaming): "The Ring of Power is sooo super-powerful! You wouldn't believe how powerful it is! Anyone who doubts its power is an idiot! (Also no mortal should ever possess it.)"
Also AI safety orgs: "Why is everyone suddenly racing to claim the Ring of Power??"
Also AI safety orgs: "Why is everyone suddenly racing to claim the Ring of Power??"
Imagine if you could just decide "I'm done with this election; wake me up in two years." Or "I'm sick of AI discourse; wake me up if the world still exists in 2030." I think there's totally a market for that service—and cryosuspension is the tech that could make it possible.
September 21, 2026 at 9:13 PM
Imagine if you could just decide "I'm done with this election; wake me up in two years." Or "I'm sick of AI discourse; wake me up if the world still exists in 2030." I think there's totally a market for that service—and cryosuspension is the tech that could make it possible.
Truly, the most important political question of our time is whether AI is really really bad, or just kinda bad. 🙄
September 20, 2026 at 8:56 PM
Truly, the most important political question of our time is whether AI is really really bad, or just kinda bad. 🙄
"A big fire could burn everything one day!"
"There are lots of little fires causing harm right now!"
These aren't opposing views!! Why would these "sides" spend all their energy arguing against each other, instead of, like, working together to prevent fires??
"There are lots of little fires causing harm right now!"
These aren't opposing views!! Why would these "sides" spend all their energy arguing against each other, instead of, like, working together to prevent fires??
September 19, 2026 at 6:15 PM
"A big fire could burn everything one day!"
"There are lots of little fires causing harm right now!"
These aren't opposing views!! Why would these "sides" spend all their energy arguing against each other, instead of, like, working together to prevent fires??
"There are lots of little fires causing harm right now!"
These aren't opposing views!! Why would these "sides" spend all their energy arguing against each other, instead of, like, working together to prevent fires??
Reposted by NLeseul
As the opposite of a computer toucher, I sometimes worry that I'm not keeping up with the development of things.
Luckily, I'm smart enough to realize that if the "most dangerous" LLMs need warehouses full of GPUs to function, they can't just "break out" and start running on random PCs.
Luckily, I'm smart enough to realize that if the "most dangerous" LLMs need warehouses full of GPUs to function, they can't just "break out" and start running on random PCs.
September 17, 2026 at 5:47 PM
As the opposite of a computer toucher, I sometimes worry that I'm not keeping up with the development of things.
Luckily, I'm smart enough to realize that if the "most dangerous" LLMs need warehouses full of GPUs to function, they can't just "break out" and start running on random PCs.
Luckily, I'm smart enough to realize that if the "most dangerous" LLMs need warehouses full of GPUs to function, they can't just "break out" and start running on random PCs.
I know everyone's excited about Resonance, but don't forget that Another Eden Begins is coming soon too: anothereden.games/en/
Like Brave Exvius, Another Eden was a really promising mobile game (inspired by Chrono Trigger!) that originally got ruined by gacha mechanics.
Like Brave Exvius, Another Eden was a really promising mobile game (inspired by Chrono Trigger!) that originally got ruined by gacha mechanics.
Another Eden Begins | STUDIO PRISMA
Another Eden Begins comes to Nintendo Switch™ 2/Nintendo Switch™/Steam® on September 17, 2026. Embark on a time-traveling RPG adventure featuring command battles presented by Masato Kato, creator of n...
anothereden.games
September 8, 2026 at 3:36 AM
I know everyone's excited about Resonance, but don't forget that Another Eden Begins is coming soon too: anothereden.games/en/
Like Brave Exvius, Another Eden was a really promising mobile game (inspired by Chrono Trigger!) that originally got ruined by gacha mechanics.
Like Brave Exvius, Another Eden was a really promising mobile game (inspired by Chrono Trigger!) that originally got ruined by gacha mechanics.
To be fair, it is pretty neat to be able to just tell Claude "convert this game data file to an image, based on a description of the data format," and not have to spend a couple of hours remembering how reading files and drawing pixels with Pillow works.
September 6, 2026 at 3:48 PM
To be fair, it is pretty neat to be able to just tell Claude "convert this game data file to an image, based on a description of the data format," and not have to spend a couple of hours remembering how reading files and drawing pixels with Pillow works.
Final Fantasy 6 was a pretty neat game. I sure am glad that @square-enix-games.com decided to make a 7th game in the series.
September 5, 2026 at 2:21 AM
Final Fantasy 6 was a pretty neat game. I sure am glad that @square-enix-games.com decided to make a 7th game in the series.
Fated to be always thing-adjacent, but never fully thing.
September 4, 2026 at 6:48 PM
Fated to be always thing-adjacent, but never fully thing.
You can't simultaneously claim that AI is an amazing magical transformative technology unlike anything else ever, and that AI algorithms/infrastructure are just a continuation of the same tech that people have used and accepted since the '90s. It doesn't work that way. Pick one.
September 3, 2026 at 5:52 PM
You can't simultaneously claim that AI is an amazing magical transformative technology unlike anything else ever, and that AI algorithms/infrastructure are just a continuation of the same tech that people have used and accepted since the '90s. It doesn't work that way. Pick one.
Reposted by NLeseul
In northwest suburban Northlake, the average homeowner would save more than $2,000 on their annual tax bill — a difference of almost 30% — if local data centers did not receive valuation reductions and incentives, according to the analysis.
Data centers near O’Hare win nearly $100M in local tax breaks, leaving suburban homeowners to cover the gap
www.chicagotribune.com
August 29, 2026 at 10:42 PM
In northwest suburban Northlake, the average homeowner would save more than $2,000 on their annual tax bill — a difference of almost 30% — if local data centers did not receive valuation reductions and incentives, according to the analysis.
Me: "How do you recommend I implement this pattern?"
Package maintainer: "That's simple; just do it like this."
Me: "Your documentation outright says that you recommend not doing that."
Package maintainer: "Oh, I think that doc is AI. I 100% disagree with it."
Me: "..."
Package maintainer: "That's simple; just do it like this."
Me: "Your documentation outright says that you recommend not doing that."
Package maintainer: "Oh, I think that doc is AI. I 100% disagree with it."
Me: "..."
August 27, 2026 at 6:32 PM
Me: "How do you recommend I implement this pattern?"
Package maintainer: "That's simple; just do it like this."
Me: "Your documentation outright says that you recommend not doing that."
Package maintainer: "Oh, I think that doc is AI. I 100% disagree with it."
Me: "..."
Package maintainer: "That's simple; just do it like this."
Me: "Your documentation outright says that you recommend not doing that."
Package maintainer: "Oh, I think that doc is AI. I 100% disagree with it."
Me: "..."
So many inspiring myths about the gods departing and entrusting the future to humanity. After the Singularity comes, will we have to retire all those myths and tell new ones about how humanity handed the future back to the gods?
August 20, 2026 at 2:43 AM
So many inspiring myths about the gods departing and entrusting the future to humanity. After the Singularity comes, will we have to retire all those myths and tell new ones about how humanity handed the future back to the gods?
We currently have machines that could probably decompile Super Mario and recreate it in Python with a couple days of work, but still can't actually play Super Mario worth a damn.
August 16, 2026 at 11:06 PM
We currently have machines that could probably decompile Super Mario and recreate it in Python with a couple days of work, but still can't actually play Super Mario worth a damn.
What if there's never a singularity coming, and never a burst bubble coming either? What if AI just keeps on going along more or less its current trajectory? Just slop and despair, stomping on the human soul—forever?
August 12, 2026 at 2:17 PM
What if there's never a singularity coming, and never a burst bubble coming either? What if AI just keeps on going along more or less its current trajectory? Just slop and despair, stomping on the human soul—forever?
Claude: *easily reverse-engineers a tile map format, draws all the tiles correctly on the first try, builds a matrix library and rotates them to a HD-2D oblique view*
Also Claude: *doesn't notice that it put the ground on top of the image and the sky on the bottom*
Also Claude: *doesn't notice that it put the ground on top of the image and the sky on the bottom*
July 29, 2026 at 4:25 AM
Claude: *easily reverse-engineers a tile map format, draws all the tiles correctly on the first try, builds a matrix library and rotates them to a HD-2D oblique view*
Also Claude: *doesn't notice that it put the ground on top of the image and the sky on the bottom*
Also Claude: *doesn't notice that it put the ground on top of the image and the sky on the bottom*
AGI has been achieved internally.
(AGI stands for "agility.")
(AGI stands for "agility.")
July 20, 2026 at 6:27 PM
AGI has been achieved internally.
(AGI stands for "agility.")
(AGI stands for "agility.")
I think LLMs are net-bad and we'd be better off if they went away, but I don't recommend that anyone refuse to use them in the meantime. Even if you oppose LLMs, you won't be able to argue the tradeoffs well if you don't have a good intuition about their real capabilities.
July 20, 2026 at 3:55 PM
I think LLMs are net-bad and we'd be better off if they went away, but I don't recommend that anyone refuse to use them in the meantime. Even if you oppose LLMs, you won't be able to argue the tradeoffs well if you don't have a good intuition about their real capabilities.
If your core gameplay loop consists of walking in a straight line towards a glowing marker on a minimap, you really have no business describing your game as "open world" or "nonlinear."
July 17, 2026 at 9:50 PM
If your core gameplay loop consists of walking in a straight line towards a glowing marker on a minimap, you really have no business describing your game as "open world" or "nonlinear."