Interview, Fireside Chat, Conference Presentation
Martin Shkreli: Big Pharma's Blind Spot, Betting Against Quantum & What Prison Taught Him
- Plans to increase drug prices and expectation of prosecution for failing to maximize profits.
- Predictions that modern quantum computers operate at hundreds or thousands of hertz rather than gigahertz, lack quantum RAM, and will never outperform NVIDIA machines.
- Belief that quantum computing fans are mistaken about sorting capabilities and that OpenAI is uninterested in quantum technologies.
- Forecast that photonic computing will render current GPUs like H100s, B100s, and B200s obsolete within 20 years and could make NVIDIA's dominance unsustainable if transformers become 1,000 times more efficient.
- Projections of infinite demand for intelligence and work in the AI market, with AI starting businesses and major corporations like McDonald's potentially utilizing 10 million GPUs for strategic decisions.
- Assertion that only consumers and businesses will succeed in the AI marketplace, with current chatbots viewed as early-stage technology that will eventually seem primitive.
- Expectation that AI will produce zero AI-relevant drugs within the next 10 years and doubts regarding AI's ability to assist in drug discovery.
- Belief that Dario Amodei's timeline for human unemployment may be delayed by three years and that delivering models holds more value than the models themselves.
- Observation that some industry figures like Dario Amodei, Dustin Moskovitz, and Sam Bankman-Fried are highly intelligent in one area but incompetent in others.
- Claim that the company "meta" has a competitive advantage due to distribution with Muse and that Sam Altman's hardware hedge is effective.
- Predictions regarding the obsolescence of early computing technologies like punch cards and vacuum tubes due to investor bankruptcy.
- Forecast that the AI industry will eventually reach a point where increasing token count does not improve answers if intelligence is not asymptotic, though no limit to intelligence is expected.
- Expectation that free intelligence in pockets represents the primary future benefit and that figures like the President and Jensen Huang have the resolve to dismiss fear-mongering.
- Observation that Nick Bostrom's "Superintelligence" frightened Elon Musk and that "Jevin always wins," leading to a future AI marketplace where "all flowers bloom."
- Anticipation that EA personnel exhibit narcissistic self-importance driving fear mongering, while Martin Shkreli attributes his legal issues to a failure to admit he was not the best hedge fund manager.