What happened in this episode?
Martin Shkreli uses a short All-In Podcast appearance to attack the commercial case for quantum computing and to outline where he thinks computing advantage will come from next. He argues that quantum machines are commonly mis-sold as “faster computing”, saying most business workloads would not run faster on current architectures and claiming that, after trying quantum services available worldwide, he and his wife could not get any to perform basic arithmetic. Shkreli says he has shorted quantum computing companies and that the trade has cost him millions.
He pivots to what he describes as more plausible compute disruption, discussing analog optical or photonic approaches to matrix multiplication, and suggests this could eventually displace today’s GPU-centric AI infrastructure — while still describing himself as long semiconductors for now. On AI markets, he argues demand for “intelligence” is effectively unbounded and speculates that AI systems could increasingly create and operate new businesses.
On life sciences, Shkreli says AI “does not help drug discovery” and predicts there will not be a single “AI relevant” drug approved within 10 years, citing his experience taking a drug through FDA approval. He also criticises effective altruism and AI safety “fear-mongering”. In a personal coda, he says prison forced him to confront ego and status anxieties, arguing his fixation on being seen as a great trader contributed to the choices that landed him in prison.
