Real Alpha
Try Pro free

Moonshots · 2 min · Nov 4, 2026

Benchmarks can steer AI toward problems people care about

Alexander Wissner-Gross · Computer Scientist & Founder, Reified

Alexander Wissner-Gross argues that measurable benchmarks help direct superintelligence toward specific goals.

The episode’s best minutes

Full episode
  1. 1:01:57Models can edit neural weights without retrainingPro pick 8 min
  2. 49:09Companies can move workflows into an AI improvement loopPro pick 8 min
  3. 35:02Benchmarks can steer AI toward problems people care aboutPro pick 2 min
  4. 3 more moments in the full episode
Up next · The Joe Rogan ExperienceAI wealth needs a way to include people who lose workPro pick 6:58
“If you can't measure it, you can't optimize it. And now in an era of superintelligence, everything is getting optimized.”
Alexander Wissner-Grossat 35:15 · Moonshots
35:15

Takeaways Pro pick

  1. Benchmarks make progress measurable and optimizable.
  2. A benchmark can focus AI on a specific mission.
  3. You can build an evaluation without being a domain expert.

About this segment

Alexander Wissner-Gross argues that measurable benchmarks help direct superintelligence toward specific goals. He suggests founders and advocates build evaluations around the outcomes they want to advance.

Alexander Wissner-Gross, Host

Computer Scientist & Founder, Reified

Has backed and advised 40+ tech companies, directly and through 021T Capital and Reified; writes The Innermost Loop.

  • Peter Diamandis, Host · Founder, XPRIZE

    Co-founded Singularity University and founded Zero-G; co-wrote the books Abundance and Bold.

  • Salim Ismail, Host · Founder, Open ExO

    Co-wrote Exponential Organizations; was Singularity University's founding executive director; sold Ångströ to Google.

  • Dave Blundin, Host · Founder & GP, Link Ventures

    Has co-founded 23 companies; founding CEO of Vestmark and chairman of insurance marketplace EverQuote.

From “Big Tech’s $3T AI Bet, Claude Writes 90% of Anthropic’s Code, AIs Editing AI Brains | EP #308”, Nov 4, 2026. Watch the full episode

Transcript

Transcribed by machine (Whisper) from the episode’s audio, so a word may be off here and there.

  1. 35:09

    Alex, it's worth taking a moment and talk about what we wrote about in Solve Everything, that if you want to move a field forward, creating a benchmark to actually measure progress is important. Would you expand on that? It's absolutely essential. If you can't measure it, you can't optimize it. And now in an era of superintelligence, everything is getting optimized. And if you have a field, as Peter, you and I argued in Solve Everything, if you have a cause that you care about, if you have, as you would say, a massive transformative purpose, an MTP that you care about, the best way, arguably, all other things being equal, to get superintelligence to care about it and to harness superintelligence is to create a benchmark or an eval for your MTP so that we can steer the superintelligence towards optimizing and maximizing your goal.

  2. 35:55

    And so just a shout out to the entrepreneurs watching this program. I mean, you can create your own benchmark on something you care about, whatever it might be, social, technical, energy, whatever it might be. And then challenge people to use the latest models and harnesses to move that benchmark forward and along. um i'll give if i may an example i'll give a shout out to a friend i have no financial interest in this but he'll probably be tickled for the shout out i have a friend jonathan hendler who created a benchmark for something that he cares about he cares a lot about human relations and mediation so he recently launched his own benchmark mediation bench that scores the ability of all the frontier models to arbitrate disputes among humans and i do think benchmarks like mediation Bench, if you Google it or go to the website, you can read about it. I think there is room in the

  3. 36:49

    market for thousands of opportunities like Mediation Bench for whatever cause you care about, whether it's dispute resolution or something else, go launch your benchmark like Mediation Bench. And as a company, you can launch your own benchmark for your efficiency of your employees or your product improvement rate. We're trying to track one, which is economic value produce per unit of human involvement, right?

  4. 37:13

    Because that would give you a really good heuristic as to how much AI enablement you have in there and humans above the loop