ron

T2 · manager / operator

Former DeepMind researcher and Windsurf (acquired by Google DeepMind) founder; left $2B acquisition proceeds to start Trajectory; building continual learning platform for AI agents with $15M from Conviction.

2 calls·2 names·100% bull·last heard 8 days ago·TBPN
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
$TRAJECTORYTrajectoryposition

Trajectory raises $15M (Conviction) for continual learning; 10x smaller models beat frontier, improve 1% daily

Trajectory builds agents that learn continuously from production usage (edits, retries) rather than static training; 10x smaller models beat frontier models today and compound at 1% daily improvement, serving Harvey, Decagon, Clay, Rogo as 'research lab in their back pocket'.

TBPN2026-05
2ndmedium conviction
$INTOLOGYIntologyposition

Intology automates scientific discovery starting with AI post-training optimization

Domain-agnostic discovery automation: propose experiment, get feedback, learn, iterate; focuses on verifiable domains but building systems that wean off full evaluator requirements; aims for fully autonomous R&D infrastructure deployed in computational environments within two years.

TBPN2026-08

most discussed · click a bar to filter

2 total
$INTOLOGY
Intology
MEDron·TBPN·8 days ago· position
Intology automates scientific discovery starting with AI post-training optimization
Domain-agnostic discovery automation: propose experiment, get feedback, learn, iterate; focuses on verifiable domains but building systems that wean off full evaluator requirements; aims for fully autonomous R&D infrastructure deployed in computational environments within two years.
"we really believe in not building co-pilots... what we are really trying to do here is build fully autonomous systems that are deployed in R&D environments and just run the entire…"
158:00
$TRAJECTORY
Trajectory
HIGHron·TBPN·2 months ago· position
Trajectory raises $15M (Conviction) for continual learning; 10x smaller models beat frontier, improve 1% daily
Trajectory builds agents that learn continuously from production usage (edits, retries) rather than static training; 10x smaller models beat frontier models today and compound at 1% daily improvement, serving Harvey, Decagon, Clay, Rogo as 'research lab in their back pocket'.
"What we're interested in is building like building an agent that is 10 years of experience on the job... we have models that are 10 times smaller than the frontier models that are…"
78:40