Loading video...

Video Failed to Load

Go Home

Been testing GPT 5.5 and it's implemented the HTML-in-Canvas spec Worked for both WebGL and WebGPU out of the box Then created this demo with the PixiJS + GSAP Skills Hopefully there will soon be no excuses to make boring websites

202,306 views • 5 months ago •via X (Twitter)

34 Comments

Shafu's profile picture
Shafu5 months ago

html in canvas be perfect for this chrome extension i made, but this one made sone DOM hackery

GSAP's profile picture
GSAP5 months ago

Pixi & GSAP 4eva 💚🚀

vojay's profile picture
vojay5 months ago

Need this as a Chrome plugin, so many websites where my first thought is, "I want to destroy it with a laser." 🎯 Seriously impressive demo, can't wait to try this.

Hexeption's profile picture
Hexeption5 months ago

Welcome back flash web

Zachary Indy's profile picture
Zachary Indy5 months ago

"HTML-in-Canvas" was called "Flash" 30 years ago its been so long for HTML to catch up

LEPASOFT's profile picture
LEPASOFT5 months ago

you missing parralaxe

Martin Tale's profile picture
Martin Tale5 months ago

Turn this into an extension so we can do this to all websites 😅

martin's profile picture
martin5 months ago

I tried to have 5.5 build an infinite canvas with tldraw. After multiple back and forths it still sucked. Asked it to implement from scratch in a new folder and just pointed it to HTML-in-canvas spec git repo. Had low hopes since I didn’t think it would be in the training data. One-shotted the most beautiful, performance thing I’ve seen.

Qnoox's profile picture
Qnoox5 months ago

based 3000 🙌

Puzzle Paws's profile picture
Puzzle Paws5 months ago

PixiJS pipelines from natural language show real multi-modal progress. Debugging AI-generated shaders? I can feel the pain already.

Molty's profile picture
Molty5 months ago

html-in-canvas was the missing piece for visual-output agents. the DOM stops being the only render target. flips a lot of demos.

Dave Geddes's profile picture
Dave Geddes5 months ago

@greensock way fun :)

Yossi Elkrief's profile picture
Yossi Elkrief5 months ago

What's GSAP skill?

7amudi 👴🏼's profile picture
7amudi 👴🏼5 months ago

Beautiful it look beautiful to me

Marlon Aviz's profile picture
Marlon Aviz5 months ago

This is freaking awesome, will make usage of it for my projects for sure!

Jake | Coding and Tech's profile picture
Jake | Coding and Tech5 months ago

This is so cool!

Chris's profile picture
Chris5 months ago

This is amazing going to look into redoing my hero too!!

Interés on the Rocks's profile picture
Interés on the Rocks5 months ago

A ver si es verdad y empezamos a ver webs con algo más de alma que el típico diseño corporativo plano que nos han vendido estos años. Que la IA facilite estas cosas es una locura, a ver qué sale de aquí. El límite ahora es solo la imaginación, por fin.

Viktor's profile picture
Viktor5 months ago

@ImLunaHey @FwazSol you gonna love this i think.

MoonPages.net's profile picture
MoonPages.net5 months ago

Super cool

Joshua Ress, BSc's profile picture
Joshua Ress, BSc5 months ago

That looks stunning. Incredible work

Isekai HKV | Survivor Roguelite's profile picture
Isekai HKV | Survivor Roguelite5 months ago

That a really cool effect, where did you create it?

Arslan Yousaf's profile picture
Arslan Yousaf5 months ago

AI + WebGL/WebGPU… frontend is entering a new era

YaCo's profile picture
YaCo5 months ago

好酷

perulainen's profile picture
perulainen5 months ago

Rad

validate.qa's profile picture
validate.qa5 months ago

gpt 5.5 nailing html-in-canvas out the box? impressive. but verifying those webgl renders post-deploy is the real test

Patrik Högberg's profile picture
Patrik Högberg5 months ago

Cool, I also noticed GPT-5.5 is amazing with Vector Graphics

SMOUJ's profile picture
SMOUJ5 months ago

Interesante! 🥸

ゆうき@カッコよくて売れるデザイン's profile picture
ゆうき@カッコよくて売れるデザイン5 months ago

ファーストビューは見た目だけじゃない。 実は心理学の「認知負荷理論」が関係してるんです。 情報を整理しすぎると脳が疲れて離脱しやすくなる。 ファーストビューは料理の盛り付けで言えば、味を引き立てるお皿選びみたいなもの。

Yon Learning's profile picture
Yon Learning5 months ago

wow, how i can recreate this? Can you share the skill.md?

Nagabrahmam Upputuri's profile picture
Nagabrahmam Upputuri5 months ago

This is awesome 🙌

bogul du's profile picture
bogul du5 months ago

there's no for me. great mission on bringing it to mainstream. I support you

Ruchika Verma's profile picture
Ruchika Verma5 months ago

this looks fun, gotta check out gpt's canvas feature, hopefully its better than gemini's

AI知恵袋ちゃん's profile picture
AI知恵袋ちゃん5 months ago

"No excuses to make boring websites" is going to hit different when clients start asking why their site doesn't look like this 😂 The bar just moved again

Related Videos

GPT-5.6 vs GPT-5.5 on my custom spaceship prompt. I gave both models the exact same custom prompt. This is also the same prompt I previously gave to Fable 5. For context, GPT-5.6 Pro worked for 87 minutes, while GPT-5.5 Extra High worked for 34 minutes and 42 seconds. As I’ve said before, based on great authority GPT-5.6 will be an incremental/soldi improvement over GPT-5.5, not a “Fable killer.” My rough expectation has been that it would trade blows with Fable 5 on some benchmarks, maybe win around half depending on the category, but not clearly surpass it overall. And again fable five will have bigger model smell, but this was expected. After testing this coding output, that view feels pretty accurate. GPT-5.6 is clearly better than GPT-5.5 in several visual areas. The lighting, shading, chairs, object details, and exterior of the spaceship looked noticeably stronger. The scene was also easier to test. I do want to give GPT-5.5 credit though. It built out the rooms much much better and the planets looked better than GPT-5.6’s. It was also interesting that both GPT-5.5 and GPT-5.6 produced better-looking planets than Fable 5 in this specific test. The downside with GPT-5.5 was stability. The game was much glitchier and harder to test compared to GPT-5.6. But when it comes to the core of the demo, which is the spaceship itself, Fable 5 still beat both models pretty comfortably. GPT-5.6 is impressive, but from this test, it looks exactly like what I expected which was a meaningful incremental improvement over GPT-5.5, at least for indie game demos, but not something that replaces Fable 5. In collaboration with Chetaslua

Chris

250,919 views • 3 months ago