thingtime design concept ยท algorithm growth
The doomscroll trainer, made visible

Raise your feed brain.

Every scroll trains your feed algorithm โ€” but today the only sign of life is a pulsing dot. What if the algorithm visibly grew up as you fed it: from an egg, to a hatchling, to a mind of its own?

0 signals trained
tooltip Still an egg ๐Ÿฅš โ€” scroll to start teaching it what you love.
01005002000+
๐Ÿฅš

Egg

0โ€“99 signals

"Still an egg ๐Ÿฅš โ€” scroll to start teaching it what you love."

๐Ÿฃ

Hatching

100โ€“499 signals

"It's hatching ๐Ÿฃ โ€” starting to notice your taste in types and tags."

๐Ÿฅ

Fledgling

500โ€“1999 signals

"Finding its wings ๐Ÿฅ โ€” ranking with real confidence now."

๐Ÿง 

Thinking

2000+ signals

"Thinking for itself ๐Ÿง โœจ โ€” this feed is unmistakably yours."

And every milestone gets a Lopu fanfare
๐Ÿฆ„LopuThingtime AI
Your algorithm just hatched! ๐Ÿฃ
It's learned enough to start finding its own taste. Keep scrolling.
Hooks into what already exists: the pulsing Learning as you scroll ๐Ÿง  dot (AlgorithmMenu.tsx), the "untrained" label in AlgorithmManager.tsx, and the signal counter in useFeedEngagement.ts. Milestone toasts route through useLopu(). Thresholds are deterministic โ€” no RNG.
Motion respects prefers-reduced-motion / the theme's general.motion gate.