PhD Student Reports Astra Outpacing Claude on Research Idea

Francois Chaubard, a Stanford AI PhD student and former Focal Systems founder, posted that he launched a research idea the night before using his autoresearcher. By morning Astra had iterated on the idea extensively. Claude produced only a naive baseline then stated the task was theoretically impossible and recommended publishing a negative result. Chaubard suggested Claude should tool-call to Astra for advice. The post includes a dark-mode line chart titled Autoresearch pro.