Tag
The author developed a method to reduce agent drift in AI systems, enabling Codex to pass a Terminal-Bench task that had zero prior successes. The technique involves having the agent build a service before tackling the task directly.
An LLM developer added research and specification gates to prevent the model from implementing every method it encounters, improving code generation quality.
PACE introduces a proxy method to estimate LLM agentic capabilities using cheaper single-turn tasks, reducing the cost and time of full agentic evaluations.
Recommends using AI for reading papers and references the three-pass method from the classic 'How to Read a Paper' technique.
PerturbSpace is a spatial transcriptomics method that presses a tissue section onto a chip of barcoded microwells, using antibodies to tag cells with location codes before single-cell sequencing, achieving >90% confident spatial assignment.
ECHO is a new, simple, and free method that addresses CLI agents, continual learning, self-improvement, and world models.