That's interesting you were getting pushback on whether LLMs could rank internal research. We've found o3, gpt-5.x do a great job with earnings post-mortems, summarizing what our analysis got right and what we missed. One challenge building workflow tools for investment teams is the only eval that really matters is the PM and their time is very expensive.
Of course it's hard to know what really happened on the Avid Biosciences situation. I did some research and couldn't find many subsequent incidences of companies being notified they were in technical default for a deadline that was stale (ie should have been caught earlier, but a new LLM was able to catch). Makes me think a lot of these opportunities are going to be in the long tail and highly varied.
That's interesting you were getting pushback on whether LLMs could rank internal research. We've found o3, gpt-5.x do a great job with earnings post-mortems, summarizing what our analysis got right and what we missed. One challenge building workflow tools for investment teams is the only eval that really matters is the PM and their time is very expensive.
Of course it's hard to know what really happened on the Avid Biosciences situation. I did some research and couldn't find many subsequent incidences of companies being notified they were in technical default for a deadline that was stale (ie should have been caught earlier, but a new LLM was able to catch). Makes me think a lot of these opportunities are going to be in the long tail and highly varied.