TickerTain
TickerTain
NewsroomShortsPortfolioConvergence
NewsroomShortsPortfolioConvergence
$ANTHROPIC·$MA····$INTC····$BLUE-ORIGIN·$SPCX····$CRWV····$CRM····$MSFT····$NVDA····$ORCL····$CURSOR·$AAPL····$OPENAI·$AMZN····$UBER····$GOOGL····$META····$TSLA····$DATABRICKS·$PERPLEXITY·$LYFT····$NBIS····$TSM····$LITE····$ANDURIL·
$ANTHROPIC·$MA····$INTC····$BLUE-ORIGIN·$SPCX····$CRWV····$CRM····$MSFT····$NVDA····$ORCL····$CURSOR·$AAPL····$OPENAI·$AMZN····$UBER····$GOOGL····$META····$TSLA····$DATABRICKS·$PERPLEXITY·$LYFT····$NBIS····$TSM····$LITE····$ANDURIL·
←

guanming

T3 · host / generalist

Robotics RL background. Co-founded General Instinct with Bill to optimize world action models for real-time deployment, achieving 500ms per chunk on Jetson via distillation, architecture changes, and modality innovations.

1 call·1 name·100% bull·last heard 2 months ago·Y Combinator
track record

no scored calls yet — needs a stated position or a categorical verdict, with a matured window vs SPY

top calls

highest conviction · one per company
1sthigh conviction
$GENERAL-INSTINCTGeneral Instinctposition

General Instinct founders claim 50x speedup for world action models enabling real-time edge deployment

By distilling VAE and DiT components, splitting video/action transformers, reducing flow matching to 1-2 steps, and using mask/flow modalities, General Instinct achieves 500ms per 16-action chunk on Jetson Orin, making world action models viable for production robotics.

Y Combinator2026-08episode →

most discussed · click a bar to filter

  • $GENERAL-INSTINCT

recurring themes

  • World Models & Simulation1
1 total
$GENERAL-INSTINCT
General Instinct
HIGHguanming·Y Combinator·2 months ago·Why Robotics Still Isn't Solved - But Could Be Soon | YC Paper Club· position
General Instinct founders claim 50x speedup for world action models enabling real-time edge deployment
By distilling VAE and DiT components, splitting video/action transformers, reducing flow matching to 1-2 steps, and using mask/flow modalities, General Instinct achieves 500ms per 16-action chunk on Jetson Orin, making world action models viable for production robotics.
"By using our infra, the word action model can runs 500 milliseconds per chunk, which is 16 actions on Jetson store. And we also wrote a full blog about how we did this on our webs…"
76:18
8
World Models & Simulationtailwind
World action models with latent future prediction surpass VLAs but require architectural breakthroughs for efficiency
World action models (Dream Zero, General Instinct) that predict future kinematics via diffusion transformers achieve better physics understanding than VLAs, but current implementations are too slow; latent-space prediction, cross-attention between video/action transformers, and 1-2 step flow matching distillation enable real-time deployment.