turns out the secret to cracking mobile graphics wasn't metal or vulkan - it was three.js webgl all along lol
untextured low-poly games now feel like corporate memphis back then :)
if opus 5.5 can get a shader-heavy webgl three.js game running at 60fps on an old budget android, what's the point of making games bland? and what's the excuse for shipping bloated, buggy products at all now?
my prediction: if the new tesla roadster ever actually goes on sale, it'll have an even worse production-to-sales ratio than the cybertruck. it may be a "future and progress" only for ppl living in a bubble.
a 140+ db thruster version is ridiculously loud even for a track car, and the rocket version is basically a stunt car. so the base version, which is probably the only one that will actually make it to customers, just isn't anything extraordinary by today's standards.
there are already a few evs that do 0-60 in under 2 seconds, and even the new bmw i3 now has more range than this roadster. bringing back the fan is fun, but.. aside from handling the problem is also how it looks - which is a defining factor for a sports car. never expect too much from tesla, but somehow it's just a reventon from temu 🥲
tesla roadster only proves that you can't build a real future on tech and ideas from the previous century, and scale won't help here.
the big difference between a real future and a prop future is that the real future feels natural and comfortable.
2
8
741
feels like "compression" is word of the year or more :)
With everyone busily using LLMs + Blender/Three.js to make 3D models, I thought I'd try a different geometric representation.
This dragon isn't a mesh, NeRF, 3DGS, or video model: it's 27kb of GLSL code (generated via LLM), defining a closed-form implicit surface. 🧵 [1/n]
9
808
for the record: if the current progress curve keeps accelerating and goes in the right direction, there won't be macro robotics - or it'll be a short-lived dead-end like vr.
if progress slows down and goes in the wrong direction, we could get stuck in the "mechanical humans" paradigm for another decade or two :(
when cars were still new, people were seriously patenting mechanical horses with walking legs, so not every old sci-fi idea needs to be built when we already have new options :)
ALT https://patents.google.com/patent/US200266A/en
ALT https://patents.google.com/patent/US941301A/en
3
421
"dumb" racing games have always been one of the engines of human progress. f1 drivers are actually jedi, making split-second decisions under g-forces, heat, and actual death risk, while the teams play one of the deepest economic strategy games on earth under a cost cap :)
and the rest of motorsport isn't far behind either. racing is just concentrated puzzle-solving!
Car racing is cognitive enhancement:
> 12 hrs on a driving video game, and 60-85 yr olds were cognitively overtaking 20 yr olds
> race car drivers do in one brain region what a novice does in 15
> six months later, the seniors kept the gains with zero practice
4
498
love to see more and more ppl lock in on three.js vibe gamedev! that's right, the window of opportunity looks like 12-18 months max :)
7
600
well said. deep learning is unlikely to produce general intelligence with an extinction threat, no matter what affiliated actors keep pushing into the info sphere - often actually driven by things like realizing progress is slowing down, looking for more investment, trying to monopolize the market etc.
so far, the actual problem is the current employment system and the almost complete lack of discussion about how normies are supposed to make a living as ai-driven layoffs keep accelerating. especially when some of the cringiest ai companies make ads like "you don't need to hire [profession] anymore..." :)
while that may in fact be close to the truth (and i'm not praising "jobs"), the social collapse on the horizon isn't the future we should be aiming for imo. so far, top ai players keep scaring ppl instead of taking even minimal economic responsibility
Not only I am tired of these wild AI speculations of impeding doom from self important people:
I resent them.
I actively resent people proposing to crash the economy or proposing authoritarian control over my life and other people's lives with idiotic and dangerous ideas like chip control or bans or tracking researchers.
Every idiot in history who's taken the approach of "the ends justify the means" to solve an imaginary future disaster created the very disaster they wanted to stop.
See Population Bomb leading to the one child policy and Communism leading to Mao's mass famines and Fascism leading to the death of tens of millions of people and war.
These solutions are evil and worse than any disease that they propose to solve.
It you believe you can actually predict the end of the world then you are as insane as the Heaven's Gate cult that killed themselves in the 90s thinking UFOs were coming to transcend them.
Not only should we not take your policies and fear mongering seriously, we should actively throw out any and all of your proposed solutions because they come from a place of delusion.
I don't give a shit that you work in the industry or think you saw something or that you want to virtue signal on X. You are actively contributing to a horrible future based on baseless speculation that has no grounding in reality.
Don't confuse expertise in a domain with ability to predict impact of that domain or frankly to make predictions a decade out better than a dart throwing monkey.
They are orthagonal skills.
How's Hinton's "we don't need to train Radiologists anymore" working out? How did the Population Bomb work out? The global cooling second ice age? Peak oil?
Just because someone is a bridge engineer does not mean they can predict the impact of bridges on society or that they have any actually useful insight at all on the complex, ever changing system called life.
There are people dying in wars right now. Children starving. Homelessness. Dictatorships.
In short, real problems. And we're supposed to drop everything to stop a made up problem in your head?
Pound sand.
We don't care. And we are not going to remake society based on you scary monsters under the bed delusion.
1
3
763
astra is great, but have you ever tried gemini 3.8 flash (yeah pretty unpop opinion)?
it's really good for quick iterations on details because of its better understanding of video input imo. and since astra is excelling at modeling, it's time to make mobile racing games GREAT for the first time!
6
710
dang turns out it's an EPIC FAIL 😭
all the worst stuff from overtrained 5.5 is still there, whatever was good about 4.5+ is basically dead, and "v6-wild" doesn't make up for it at all. honestly, it's absolute bs - no experimental or deconstruction, just weird floating pitch and tempo at smth like 11khz and 48kbps quality lol
with everything that happened, no complaints about suno, but they could've at least told more clearly beforehand that the party is over 🤷♂️
tbh nice to see oai finally closed the gap so you no longer need to open that ugly buggy orange ui with dario's ghost arguing in the background, thinking forever and everything just to still miss lol
GPT-6 Astra is state-of-the-art on FrontierMath Tier 4, ARC-AGI 3, and TerminalBench-4.0.
GPT‑6 Astra is also a major advance for scientific discovery, with state-of-the-art performance on Terminal-Bench Science 0.1 and HealthBench Pro.
3
374
since i've been pretty good at predicting where things are going, these are 3 absolutely generational articles i decided (should!) to write, forcing myself:
1. From MP4 to Computational Worlds — or Why We Should Hack Physics RN
2. Why Are We Stuck in the 1970s, and Is There a Way Out?
3. The Culture of Wit is Over
VERY BIG, DEEP AND HUGE! absolutely mind-changing regardless of how smart you think you are!
there are just some things that should be spoken clearly, so with this tweet i just wanna mark these ideas in the air. another 2 important things require video as the proper form, but i hope to do those by the end of this year too.
believe me, these are very important announcements of announcements!
5
490
‼️ DEAR VIBE CODE PLATFORMS! PLS STOP THIS BS WORDING LIKE "agentic orchestration layer with context engineering, multi-agent workflows, MCP-powered tool use and autonomous execution" ‼️
you literally LIMIT YOUR REACH with these useless, short-lived terms! even "vibe code" may be too much - there's nothing to code already. don't raise the threshold for ppl artificially when it has already fallen so much in practice! tech these days shouldn't be "techy"!
just "MAKE APP / MAKE GAME / MAKE SITE WITH AI" max!!
2
6
431
