Explore

FUTURE QUESTIONS

how does this work with mobile?

problems: 1 browser extension won’t even work, 2 cursor interactions all fall apart.

technically both solvable.. safari extension and i guess something like what neal did - joystick & click button.. but it takes up so much real estate. do you lean into hardware for your phone (a little mouse that you plug in)? it would need to be a lot more streamlined… i don’t browse on my phone so it would need some kind of feed or randomness jump into a spot. need a destination… maybe this is where those intermediate places could be cool? idk

what about AT PROTO?

how does this integrate with where the world is moving with technology? increasing action taking on the internet is done by “agents” but then i think we’ll want somewhere to spend that is more human? there is the more functional internet (markdown websites. apis) for machines and then the things w no function / optimization for humans?

i think html in canvas is also going to be huge for the kinds of interactions enabled… but i really need to experiment still.

prove

⁠

Mobile — direction

Dedicated mobile browser app over Safari extension. Frame mobile as the receiving / reflection surface, desktop as the casting / wandering surface. Skip the extension fight entirely.

What the app could be:

A feed of pages other extension users are spending time on right now — ambient view of where attention is going. Tap to enter via in-app browser. Solves the no-destination problem (the destination is where other people are).

A bottles-only surface — open the app, you have a bottle waiting on a page. Tap, it takes you to that page in the in-app browser with the bottle visible. Mobile = the receiving ritual; desktop = the casting one.

A live map of your own browsing — self-portrait (internet) rendered for you on your phone. Mobile as journal/reflection.

Secondary track: PWA / playhtml snippet that site owners include — bottles + cursors work in any mobile browser without the app. Site-owner opt-in, consistent with playhtml-first philosophy.

Shelve the hardware mouse idea — niche-within-a-niche audience overlaps with people who’d just open their laptop.

Test before committing: would I use this app daily? If it doesn’t make me want to browse on my phone, I’m building for a hypothetical user.

html-in-canvas — what it actually unlocks

Not needed for rich-media bottles. Doodle / voice / animation bottles can ship today with regular DOM + canvas + MediaRecorder.

What html-in-canvas is actually useful for: the atmospheric / environmental layer.

WebGL weather over any page that doesn’t break when it touches normal HTML — particles, distortion, parallax, with the page still interactive inside the canvas’s coordinate space.

Co-presence with texture — another user’s cursor as a glowing thing in 3D space, leaving real visual wakes that warp the environment. CSS fakes this; html-in-canvas makes it native.

Shared 3D layers across sites — a generative environment that persists as users navigate, weather/state shared across pages.

Status: still early (Chrome experimental flag, no Safari support). Don’t build on it for v1. Track it for the 2027 environmental-layer version.

Want to print your doc?
This is not the way.

Try clicking the ··· in the right corner or using a keyboard shortcut (

CtrlP

) instead.