←
Jetha Chan42,020 posts
Jetha Chan avatar

Jetha Chan

@jetha

developer relations at @aiand_ | agentic AI enthusiast, building in public. | prev: @GoogleDeepmind, Google Stadia

◷ Joined May 2008▧ 2,318 Photos and videos
Jetha Chan avatar
Jetha Chan@jetha·Oct 28

AMDパートナーの実際の課題に、オープンモデルで挑戦しませんか?8チーム、4.5時間、最後はステージでデモ。応募は11/9まで👇

ai&@aiand_Oct 21

Applications are open for @aiand_ x @AMD Build Day! 🗼 8 problems from AMD partners. 8 teams. 4.5 hours to build. Then demo your work on the Advancing AI stage. Build with @Kimi_Moonshot, @Zai_org GLM, @Alibaba_Qwen, and @MiniMax_AI on ai& inference, powered by AMD. 📅 December 1, 2026 🕤 9:30 a.m. – 7:30 p.m. 📍 Palace Tokyo, Marunouchi Apply by November 9: amd-advancing-ai.jp/buildday…

Post attachment
Ariel Deschapell (compute/acc) avatar
⟳ Jetha Chan retweeted
Ariel Deschapell (compute/acc)@NotASithLord·7h

What if, and hear me out, the US simply produced superior open source models.

SITUATION DETECTED: The White House and Treasury are discussing a proposal that would require US companies to publicly disclose the use of Chinese open-weight AI models on their websites, aiming to discourage their use through public pressure, per Reuters.

ludwig avatar
⟳ Jetha Chan retweeted
ludwig@LudwigAhgren·21h

one of the unforseen results of AI is the digital mccarthyism that comes with every piece of art it is not longer enough to be an artist and to create you must now also be able to prove you are not AI any mistake, stylistic choice, or overanalyzed detail can lead to a witch hunt where the artist must prove a negative

Mark Price avatar
⟳ Jetha Chan retweeted
Mark Price@PlayUGC·20h

I created a game in the ORIGINAL Unreal Editor I'm not sure why I did this, but i'm glad i did. After I built the MCP, the game only took MINUTES. That SAME game in UE 5.9 and Roblox took 5x longer to make. There's something special about building so fast. Not fighting load times or having a thousand tools I don't even need. Testing gameplay in this very old editor is so insanely rapid. I don't know what it is, but we need to get back to this to some degree. You made something really special here @TimSweeneyEpic

0:54
Varun Mohan avatar
⟳ Jetha Chan retweeted
Varun Mohan@_mohansolo·22h

One more thing… In honor of Antigravity turning 1, we’re resetting quota for everyone across all models!

The Lunduke Journal avatar
⟳ Jetha Chan retweeted
The Lunduke Journal@LundukeJournal·23h

5 months ago, the Godot game engine banned AI generated code from their project. Flash forward to today: Godot has announced that an AI company, @ManusAI, is now a Platinum financial sponsor of Godot, with a glowing statement regarding AI development of games using Godot.

Post attachmentPost attachmentPost attachment
Patrick C Toulme avatar
⟳ Jetha Chan retweeted
Patrick C Toulme@PatrickToulme·Nov 17

You will notice that Gemini Carbon thinks differently than Claude and GPT. It often comes up with entirely different but more clever solutions than competitor models. This is precisely because we never distilled other models. Distilling other models creates a poor imitation rather than unique intelligence.

Hakm@hakmgptNov 17

Google has proven that when you’re months or even a year behind, the solution isn’t to distill models. and they proved that good research takes time and I have high hopes for SSI as well :)

Japan Embassy DC🌸 avatar
⟳ Jetha Chan retweeted
Japan Embassy DC🌸@JapanEmbDC·Nov 16

It’s Japanese Language Education Week! Today we are teaming up with Michael-Christopher Koji Fox, localization director and supervisor at the video game company @SquareEnix. Hear how learning Japanese changed his life and shaped his career!

1:19
Shishira avatar
⟳ Jetha Chan retweeted
Shishira@ShishiraReddy05·Nov 16

Must must must must must must must must must must must must must must must must must must must must must must must must must must must must must must read if you’re learning harness engineering x.com/theo/status/2113424203… x.com/hwchase17/status/20429… x.com/bytemohit/status/20634… x.com/bibryam/status/2108946… x.com/addyosmani/status/2053…

its hard finding good content on harness engineering for beginners. please drop resources if you have any worth reading or implementing

Omar Sanseviero avatar
⟳ Jetha Chan retweeted
Omar Sanseviero@osanseviero·Nov 17

With Carbon's help, I ran Gemma directly in an office printer 😅🖨️ 🤏Carbon compiled a small transformer engine + quantized Gemma 3 270m into a single PostScript job ⚡️ 100% in the printer: 14MB peak RAM and blazing 60s per token 🧠Smart layers juggling to avoid OOMing

Post attachment
Shunsuke Sako avatar
⟳ Jetha Chan retweeted
Shunsuke Sako@SS_Sako·Nov 16

You fall in love with a series. Then you find out the merch isn't sold in the US. So you import it from Japan, pay a fortune in international shipping, deal with customs, and hope it shows up in one piece. I'm the CEO of a new company that wants to fix that.

Post attachment
ai& avatar
⟳ Jetha Chan retweeted
ai&@aiand_·Nov 17

Two days of great conversations about inference and the future of AI in Japan @sgl_project Summit 2026! @Weyaxi took the stage to talk RL on heterogeneous infrastructure. Thanks to everyone who stopped by our booth and @lmsysorg for bringing this community together.

Post attachmentPost attachmentPost attachmentPost attachment
Jetha Chan avatar
Jetha Chan@jetha·Nov 17

0:05

My favorite part of reporters these days is the glee they take in taking down people who have been risking and working since the reporters were in diapers. It’s a function of our economy, but it also betrays a complete lack of understanding that is only obtained by building a career as a parasite. The expectation is a CEO coddles the holders of bad ideas or execution because the everyone “hates” their boss and it makes a good story for readers. Except people didn’t hate Jensen. Only the reporters do.

John Carmack avatar
⟳ Jetha Chan retweeted
John Carmack@ID_AA_Carmack·Nov 16

Both NeXT and 3DFX tried to “do the right thing” with color spaces, and it wound up generally making them look worse than competing options. I didn’t really understand it properly until years later. It is natural to think that a pixel with a value of 1.0 is emitting twice the photons as a pixel with a value of 0.5, so you can do the obvious linear math to calculate partial pixel coverage, blending a foreground over a background, or interpolating between colors in a texture. If displays are “linear”, that is correct. NeXT and 3DFX were much closer to linear than other contemporary displays. NeXT had crossover from the deep graphics understanding at Pixar, and 3DFX had the SGI heritage. They did this intentionally. If you have 10+ bits per channel like the professional rendering systems, linear displays are The Right Thing. Unfortunately, NeXTStation Colors were 4 bits per channel and 3DFX was 5 or 6 bits per channel, and there… it was not the right thing. Human eyes aren’t linear intensity sensors. We mostly register relative changes in intensity, so we can tell the difference between 0.01 and 0.02 as a huge doubling, while 0.99 to 1.00 is almost imperceptible. This distribution of intensities is often referred to as the “gamma” of the display system, but the actual curves like sRGB are a bit more complicated. If you only have a small number of discrete pixel values, you definitely don’t want to spread them evenly from 0 to 1! If you do, you will have good precision in the brights, but very stark “intensity bands” in the dark colors. If everything was authored specifically for 3DFX, you would still have the fundamental problem with the dark bands, but more accurate blending and filtering. The problem was that artists made all of their textures in paint programs on conventional non-linear displays, so when they were drawn by 3DFX cards, they got the characteristic “washed out color” look. You can mathematically correct textures for differences in display gamma, but I didn’t really understand the issue back in the glQuake days, and we were so enthralled by the absolute performance of 3DFX that the color fidelity didn’t get as much attention as it should have. Even today, tons of games don’t properly handle gamma across their textures and render targets. High resolutions and 24 bit color masks most of it, but you still see it in some slight shimmerings in high contrast textures and edges.

Does anybody know why 3DfX 1 cards had such a washed out color output? In my research, circa '97 I found many people complaining about the colors. Is it gamma? 16-bit color? What can it be?

Post attachment