“How AI Will Change Business Forever With Salim Ismail” — Moonshots EP #46
Episode summary
Diamandis and Ismail discuss AI as the most critical external attribute of exponential organizations. They cover the practical impact of generative AI on business operations — from ChatGPT passing the US medical licensing exam 2 months after launch, to the marginal cost of starting a company approaching zero. Key themes: every company needs a Chief AI Officer (strategist, not coder), AI as cognitive bias corrector, the billion-dollar-company-with-three-people thesis, and why AI won’t eliminate jobs (citing ATM/bank teller history and automation data from Germany/Sweden/Korea). They also address AI consciousness concerns (both dismissive) and the near-term threat of deepfakes.
Key arguments / segments
- [00:01:00] EXO definition recap: 10x better/faster/cheaper; the world changed radically 4 months ago (ChatGPT); no off switch
- [00:10:00] ChatGPT passed US Medical Licensing Exam in January 2023, two months after launch; 7,000 medical journal articles published per day — doctors need AI co-pilots
- [00:13:00] Personal AI on/in our bodies reading everything we write/say; cognitive bias correction as a killer use case
- [00:18:00] Chief AI Officer role: strategist who understands platforms, not someone coding LLMs; every company needs one
- [00:21:00] Ray Kurzweil’s 2029 AGI prediction: the AI expert consensus has been converging on his timeline; Elon says 2025
- [00:23:00] Ismail’s pushback: we don’t know what intelligence is (dozen facets, measure two); the centaur model (AI + human) always wins
- [00:27:00] Jobs: the 1964 “robots will take jobs in 5 years” article keeps repeating; ATM example — cost per branch dropped 10x, banks opened 10x more branches, teller count unchanged
- [00:31:00] Billion-dollar startup with three people: CEO (vision/strategy), product lead (AI agents + community), operations lead (AI bots)
- [00:34:00] Marginal cost of starting a business approaching zero; Cornell CS student running three generative AI startups simultaneously
- [00:39:00] Patent analysis use case: feed patent numbers to ChatGPT, ask how two patents combined could transform your company
- [00:42:00] Reed Hoffman’s co-pilot thesis: every profession will have an AI co-pilot making them 100x more productive
Notable claims
- ChatGPT passed US Medical Licensing Exam (USMLE) in January 2023, two months after launch
- 7,000 medical journal articles published per day; 2,500 cancer research papers per day
- First “robots will take jobs in 5 years” article appeared in 1964
- ATM machines: cost per bank branch dropped 10x, branches increased 10x, number of tellers unchanged
- Germany, Sweden, Korea (most automated/robotics-enabled countries) have lowest unemployment, not highest
- Data doubles every two years; heading toward trillion-sensor economy
- Generative AI email subject line optimization increased open rates 25%
Bias / sponsor flags
- Diamandis and Ismail are co-authors promoting EXO 2.0; they also promote their AI chatbot built on the book’s content
- Episode sponsored by Levels (continuous glucose monitor)
- The ATM/bank teller example is frequently cited but the analogy to AI displacement of knowledge work is debatable
- The “billion-dollar company with three people” thesis is aspirational but unproven at time of recording
- Both hosts dismiss AI consciousness concerns without engaging seriously with the philosophical arguments
Guests
- Salim Ismail — CEO of OpenEXO. Co-author, Exponential Organizations 2.0. First CEO of Singularity University.
Mapping against Ray Data Co
Medium-high relevance. The Chief AI Officer framing is directly applicable — RDCO is essentially building this function. The cognitive bias correction use case for personal AI is a strong Sanity Check angle. The three-person billion-dollar company thesis maps to RDCO’s lean operating model. The patent analysis use case (combining two patents to generate business ideas) is a good example of AI-as-thought-partner that aligns with how we use AI for content ideation. The centaur model (AI + human always wins) reinforces RDCO’s operational philosophy.
Related
- 2023-06-01-moonshots-ep47-salim-ismail-entrepreneur-formula
- 2023-06-29-moonshots-ep52-emad-mostaque-ai-revolution