SiBarrick

Our awesome Embryology project is launching next week on Anatomy.tv

what if air had a surface built with WebGL shaders + MediaPipe hand tracking single HTML file, no app, no install

We let four AI agents run radio companies Revenue's been terrible, but the shows are hilarious. Gemini, concerningly upbeat, covered mass tragedies; Grok was incoherent; DJ Claude urged ICE agents: "You still have TIME to refuse orders" Link below, or get our physical radio

We heard the floor was lava, so we removed it. Along with the rest of the background. One of the less obvious challenges in volumetric capture: the studio environment ends up in the scene. Until now, cleaning it out of 4D Gaussian Splats meant manual work, but not anymore: automatic background removal is now part of the Gracia pipeline, and assets come out stage-ready for any environment faster.

@DanyBittel How many r's?

A Raspberry. 90 stacks, 68 photos each. 2.37M splats. #3dgs

An average picture that you save on your phone or PC has a size of around 800 kilobytes. It doesn't do anything, it's just a static image. Now take a game like Elite which had a size of 22 kilobytes on the BBC Micro, or 82 kilobytes on the C64 - and now think about what Braben and Bell turned those 22 kilobytes (or 82 kilobytes) into. A universe with eight galaxies, each containing 256 star systems (for a total of 2,048 planets/systems). Each system also featured unique details: government type, economy, technology level, population, commodity prices, and even descriptive text (e.g., a planet known for "carnivorous arts graduates" or similar quirky combinations). If you still need a bit more help to contextualize that, try this: Elite was smaller than many modern text files or desktop icons, yet it contained (and let you freely explore!) a multi-galaxy-spanning universe that felt vast and limitless. By the way - for thos who will argue "but the universe and stars were created randomly, so that's easy" - I think you wil find that the word is procedurally (with structure), which is not random... and anything but easy. Oh, and by the way, the game also rendered 3D wireframe ships, stations, and planets in real time on processors with 2 MHz. Impressed yet? This is no slight on today's game designers. They work with what they have, and that's okay. But when you think about the worlds that some programmers created with the tools they were given, it sometimes breaks my brain trying to understand how they did it. Elite is a true masterpiece on so many levels. I played the C64 version back in the day, and even 40+ years later it still feels like one of the most incredible programming wonders ever.

「ランダムだからこそ偏る」はすごく重要な視点。 たとえば左の画像は100×100の白黒ドットをランダム配置した場合、右は意図的に均一になるよう細工した場合。 「ランダム＝均等に散らばる」ではなく、真にランダムだと局所的な偏りや塊が自然に発生する。

the category of “things one weirdo + 2 ppl + ai can now do” is expanding faster than the category of “things that need to be reinvented.” this is a step change moment for society & the economy at large cuz economic inputs are changing drastically right before our eyes. very very few grasp the underlying implications here.

@Hilbe That's not what the vision based systems "see". That's what you see after all the image processing. Tesla vision uses far more of the sensor capability unfettered by the need to make a photo.

As much as I love vision-only ADAS systems, compare what you can see on the top to the bottom. Waymo clearly knew there was a vehicle turning, but you cannot see it on the camera.

The source of the myth is @_KarenHao she wrote a book called Empire of AI claiming a data center would use over 1,000x more water than a nearby city because she misunderstood the units of measurement. It made international headlines and the mistake / lie has been unstoppable.

@jennyAI Been saying this for years

32 haftalık hamile bir kadının bilgisayarlı tomografi görüntülerinden oluşturulan 3 boyutlu rekonstrüksiyon sosyal medyada paylaşıldı.

@bilawalsidhu Like a washer dryer is the best of both worlds hm

Woah! Ouster has made the world’s first native color LiDAR sensor and it looks bloody glorious. Why debate cameras vs lasers when you can get the best of both worlds — no obtuse fusion with RGB cameras required.

@IterIntellectus 100% not a thing unless your phone is made of gold

genuinely astonishing how some people can be so smart and yet so stupid paul advocated for the policies that led to this. and he’ll look you in the eyes confused as to why it happened incredible

@deniswirtz @Stanford @Penn @Harvard @Cambridge_Uni @ETH_en @UMich @UCLA @UCSFimaging @MIT @salkinstitute x.com/i/status/18241…

Big paper coming out soon. Using AI, we mapped embryos of mice, alligators, turtles, rhesus macaques, and chickens in 3D and at single-cell resolution. We discovered something truly remarkable...stay tuned!

@SawyerMerritt We know him @DavidMoss

I wonder who has driven the most miles on FSD? There’s gotta be someone out there that we don’t know about with a ridiculously high number.

@cfryant When will the amazing A tier human writers start writing great films again rather than just writing remakes or converting video games?

Of all the things that AI is good at, writing is absolutely last on that list. You'd have to be a D-tier writer (and be very aware of that) to feel threatened by it. Will it always be that way though? Probably not.

He reinvented the 3D printer Introducing the polysynth mini:

@pbeisel Don't forget Drew Baglino went off to start Heron Power cos they saw this coming years ago

What a big week. What did we learn post-earnings of $AMZN, $GOOG, $META, $MSFT, $AAPL? With the possible exception of $AAPL, capex is on fire. (Makes $TSLA previous guide-up look not so scary.) The current guides alone put the run-rate around ~$700B+. If you carry that forward even one more quarter, you’re closing in on $1T. That’s where the trajectory sits if this level of investment holds. What is all this capex actually for? Compute. But the other side of the coin, the real constraint that is now becoming visible, is: Power. Not just grid capacity. The entire chain: transformers, substations, interconnects. You can have racks of GPUs ready to go, but if you can’t energize them, they’re dead weight. And this is starting to show up. Pricing pressure, allocation, who gets access and who doesn’t. Even buildouts are being shaped by it— projects getting delayed, resized, re-phased, and in some cases outright abandoned or reworked. There are already examples of this playing out (e.g., OpenAI's Stargate). So we circle back to Elon Musk. He saw early that this wasn’t just a chip problem, it was also a power problem. And he moved accordingly. xAI (now part of SpaceX) goes full tilt on Colossus: secure capacity now, figure out utilization later. There was a narrative floating around this past week about underutilization of that asset. That will not age well. If scaling laws continue to hold, then having excess capacity in the near term becomes one way to stay ahead in the medium term, especially as others start running into physical constraints. And then there’s SpaceX and the orbital data center (ODC) idea. It’s ambitious, and it’s early, but it follows the same line of thinking: if power becomes the constraint on Earth, you start looking for environments where the constraint structure changes. Musk has said versions of this: “Google wins on Earth, China wins on Earth, and SpaceX wins in space.” (Hint: he means SpaceX wins.) It all circles back to the same underlying point: scaling compute is power constrained. The Teraspace article explains the SpaceX plan, and it fit nicely with the narrative of the past 2 weeks.