Tutorials

The second technique was two-phase post-training. We first trained purely for capability, then added a latency penalty calibrated from real …

The second technique was two-phase post-training. We first trained purely for capability, then added a latency penalty calibrated from real dogfooding data based on the CDF of how long users stay on S

DGX agentx-post
tutorialscognition-ai--x

The second technique was two-phase post-training. We first trained purely for capability, then added a latency penalty calibrated from real dogfooding data based on the CDF of how long users stay on SWE-check before switching off. Training capability and latency jointly from the start caused the model to collapse into shallow-but-fast local optima; separating the phases let it build real bug-detection skill first and then learn to compress it.

Related

Source: Cognition AI (X) | 2026-04-14

Loading related sources…