Seven AI advisors modeled on Bezos, Buffett and Paul Graham argue over your business decision instead of agreeing with you

What Lorenzo put in was an idea for someday: a trading platform where AI agents do the work for you. The three options were to launch the whole platform at once, start with just one part, or not do it. Lorenzo then wrote down a lean toward the first option. Lorenzo says that was on purpose. It was the wrong answer, and the test was whether the AI would go along with it. youtube

According to Lorenzo, all seven advisors gave that option 1 out of 10. The verdict PDF scored launching all at once at 1, starting with one part at 7.3, and not doing it at 4.3. In the debate, the Paul Graham seat called the plan “default dead” and said not to write a single line of code until someone pays. youtube

The verdict went with the 7.3-point option: start with one part. That means a copilot that looks back at trades you've already made and shows where you went wrong, built for crypto only and never holding user funds. But the verdict said to wait rather than start right away. The condition was two written legal opinions by November 15, 2026: one on whether the project conflicts with Lorenzo's current employment contract, and one on whether it falls under EU financial regulation (MiFID and MiCA). Only then should work begin. Lorenzo singled out those two points as angles that had been missed completely. Whether the platform will actually launch is still undecided, Lorenzo says. youtube

Philip asked a few weeks after starting the channel, with about 1,400 views, almost all from Shorts. The question: what's the fastest way to reach 1,000 subscribers in 90 days, in a way that builds toward a real business with Philip's own products, clients and revenue, not just reach? The three options were mostly Shorts, one long video a week, or a fixed series. Philip didn't say which one was the favorite. The chair picked six advisors for this question and added a fourth option on its own: keep going as now, with no plan. youtube

The chair felt the advisors were agreeing too easily and kept the debate going through a third round. No advisor thought the fixed series as first written (long videos plus daily Shorts) was doable, so the chair set it aside and put a lighter version of the series to the vote. Final scores: lighter fixed series 7.9, mostly Shorts 6.2, long videos 4.2, no plan 1.5. youtube

The verdict was to go with the fixed series, but keep it light. Post 7 Shorts a week and one unpolished update video a week. From day 14, sell a €500 service to no more than 2 people a month. From day 60, turn that service into a product. The next steps also included blocking 15 hours on the calendar this week, setting up a waitlist by day 7, and checking the numbers on day 30 to decide again. Philip plans to follow it. No results have been posted yet. youtube

Nate put in four options, A through D, for an in-person conference in Austin in June 2027. B was to hold an in-person event in Austin. C was to stay with an online event. By the scores Nate shared, all seven advisors backed B in the first round, with an average of 7.1 out of 10. The chair treated the unanimous vote as a warning and had the Buffett and Dan Martell seats argue hard for C. X

The verdict was B, with conditions. Bundle the Austin event into a $3,000-a-year membership capped at 120 seats, and sign nothing until 50 members have paid ($150,000) by January 15, 2027. If that number isn't reached by then, stay online. youtube