Thoughts on code agents, rock climbing, and more.

The growing concern of closed AI

Anthropic's restrictions on LLM research raise a broader question about who controls access to transformative technology.

· Rohit Malhotra
The growing concern of closed AI

Proof in Production: Evaluating Effectiveness of SWE Agents with A/B Tests

SWE-Agents are setting new benchmark records, yet underperform in the real world. Beneath the hype, only A/B testing can prove their effectiveness.

· Rohit Malhotra
Proof in Production: Evaluating Effectiveness of SWE Agents with A/B Tests