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·
←

ian fischer

T3 · host / generalist

Founded Poetic building recursively self-improving AI reasoning harnesses. Previously spent a decade as a researcher at Google DeepMind after Google acquired his first YC startup Portable (cross-platform mobile devtools). Background in computer security and systems building.

1 call·1 name·100% bull·last heard 7 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
$POETICPoeticposition

Poetic's recursive self-improving harnesses beat fine-tuning at half the cost

Poetic's meta-system automatically generates reasoning harnesses that outperform fine-tuning, are model-agnostic, cost a fraction of retraining, and improve automatically when new foundation models are released.

Y Combinator2026-02episode →

most discussed · click a bar to filter

  • $POETIC

recurring themes

  • AI Agents1
  • AI Infrastructure1
1 total
$POETIC
Poetic
HIGHian fischer·Y Combinator·7 months ago·The Powerful Alternative To Fine-Tuning· position
Poetic's recursive self-improving harnesses beat fine-tuning at half the cost
Poetic's meta-system automatically generates reasoning harnesses that outperform fine-tuning, are model-agnostic, cost a fraction of retraining, and improve automatically when new foundation models are released.
"what we end up giving you is a harness that sits on top of one or more language models and it just performs better than them. And when the new model comes out that same harness is…"
2:39
8
AI Agentstailwind
Model-agnostic harness layer makes any LLM better, cheaper than fine-tuning
A recursive self-improving system layer ('harness') can automatically optimize prompts, reasoning strategies, and context for any foundation model, delivering superior performance at lower cost and adapting instantly to new model releases without retraining.
7
AI Infrastructuretailwind
Recursive self-improvement at system level beats model retraining for capability gains
Instead of spending hundreds of millions retraining models, startups can deploy a meta-optimization system that continuously improves the reasoning harness atop existing LLMs, achieving SOTA benchmarks (55% Humanity's Last Exam, 54% ARC-AGI v2) with <$100k compute and 7-person team.