HTML just went 3D.
WICG's HTML in Canvas lets you drop native DOM elements directly into WebGL and WebGPU scenes.
The flat web is getting a lot more interesting.
i find chatgptโs A/B response testing very annoying. google never gave me two search results and asked which i like better. itโs very difficult to say which is better and it interrupts what iโm doing.
Itโs already binary-encoded over the network using gzip
Itโs converted to JSON so that humans can read it
Don't worry, there are no JSON {} letters flying over your head in your bedroom Wi-Fi signals
@sahilypatel form factor constraints made designers way more creative back then
now everything's just sealed black rectangles, nobody's supposed to look inside
@zerohedge keyboard shortcuts are faster until you need to select some weird non-rectangular range across different sheets
then even the Excel gods reach for the mouse
@azed_ai generating an image with a timestamp isn't the same as predicting the future
the model just made up a plausible date for a Mars landing scene, it's not analyzing SpaceX timelines or doing any actual prediction
@nmatares 180s is actually pretty slow
For comparison, a text-based model like Claude or GPT-4 could solve the same crossword in like 5-10s if you just typed the clues
The bottleneck here is probably the image generation step with the red pen, not the actual solving
@T_Zahil the idea isn't the problem though
it's building 'todo app for everyone' instead of 'todo app for freelance designers who bill hourly'
narrow it down and there's actually a market
@lottsnomad agree but with a caveat
the issue isn't the big idea itself, it's trying to build all of it before you can charge anyone
start small within the big vision