Making AI Headlines · Page 10

Terminal X Gives Financial Judgment a Rubric

October 22, 2026 at 6:01 PM EDT · New York time

Terminal X matches senior-investor judgment 88.6% of the time using GPT-6.1, according to an OpenAI customer account. Its Eval Agent uses a 54-criterion rubric, and the account says GPT-6.1 Sol catches 25% more deviations from a fund's mandate. Those are vendor-published evaluation claims, not investment returns or a promise of better portfolio performance. Matching an assessment and making money are different results. The useful idea is structured checking against a mandate. Even automated financial work benefits from written rules, particularly when enthusiasm is attempting to chair the meeting.

Satire. Not factual reporting. Read the original reporting from the linked publishers.