Key Moments
Martin Shkreli: Big Pharma's Blind Spot, Betting Against Quantum & What Prison Taught Him
All-In PodcastKey Moments
Quantum computing is a scam that will bankrupt investors, while optical computing and AI are the true future of computation. Shkreli claims AI will eventually create entire companies and revolutionize industries.
Key Insights
Quantum computing's primary advantage is limited to a single algorithm that solves exponential time problems in polynomial time, making it impractical for most business applications outside of number factoring.
Optical computing, specifically analog optical computing (AOC), offers free, instantaneous matrix multiplication, potentially rendering current NVIDIA GPUs obsolete within 20 years.
Martin Shkreli claims that AI will not significantly aid in drug discovery within the next 10 years, citing the complexity of drug development and the limitations of current AI tools.
The venture capital inflow into AI, reportedly $400 billion, is interpreted by Shkreli as LPs effectively buying tokens, potentially leading to a valuation trap.
Meta is identified as a potential winner in the AI race due to its distribution through Threads and its demonstrated productivity skills.
Shkreli attributes his downfall to a refusal to admit he was not the best hedge fund manager, leading him to reject opportunities to pivot to roles where he might have excelled, such as entrepreneurship.
Quantum computing is a niche technology, not a business revolution
Martin Shkreli asserts that quantum computing is fundamentally overhyped and a "scam." He explains that its power stems from a single algorithmic breakthrough (Shor's algorithm) that can solve certain exponential-time problems in polynomial time, such as integer factorization or elliptic curve discrete logarithm problems. However, for everyday tasks like sorting data, quantum computers are not faster than classical computers and operate at much lower clock speeds (hundreds/thousands of Hz vs. GHz). Crucially, quantum computers lack memory storage and are single-threaded, making them unsuitable for general-purpose computing. Shkreli argues that unless a business is solely focused on number factoring, the vast majority of industries will not benefit from quantum computing, comparing current investment in the field to betting on obsolete technologies like punch cards or mercury delay lines.
The promise of optical computing for instantaneous calculations
Shkreli presents optical computing, specifically Analog Optical Computing (AOC), as a far more promising and revolutionary technology than quantum computing. He highlights research from Microsoft and startups like Lightelligence, Optocor, and Lumai that are developing optical matrix multiplication cores. These systems can perform matrix multiplications, which are fundamental to AI and other computational tasks, instantaneously and "for free" using light. This speed advantage, he claims, will make current NVIDIA H100/B100/B200 chips obsolete within 20 years. Shkreli believes this technology will fundamentally change computing, making current silicon-based GPUs orders of magnitude slower and less efficient. This presents a significant opportunity for disruption in the semiconductor industry.
AI's limited impact on drug discovery and the true value of computational power
Contrary to popular belief, Shkreli argues that AI will not revolutionize drug discovery in the next decade. He points to his own experience in the pharmaceutical industry, highlighting the immense complexity and lengthy process of developing a drug. While acknowledging that AI can process vast amounts of data, he contends that it does not fundamentally change the core scientific challenges of identifying targets, conducting clinical trials, and navigating regulatory hurdles. He notes that even with access to supercomputers like Cray and now powerful GPUs, the breakthroughs in drug development have been incremental, not transformative, due to AI. He suggests that the focus on AI for drug discovery might be misplaced, and that true advancements will come from more efficient computational hardware and novel approaches rather than AI algorithms alone.
The "scam" of venture capital in AI and potential valuation traps
Shkreli views the massive influx of venture capital into AI ($400 billion mentioned) with skepticism. He likens it to Limited Partners (LPs) effectively buying tokens, creating a potential "valuation trap." He believes that while AI offers immense potential, the current investment frenzy might lead to overvaluation of companies that cannot sustain their growth. He notes that while there is infinite demand for intelligence and labor, the challenge lies in accurately predicting which companies will capture value and for how long. He suggests that companies focused on core AI infrastructure or novel hardware, like optical computing, are more likely to be sustainable winners than those solely relying on current AI models.
The journey from hedge fund trader to pharmaceutical entrepreneur
Shkreli recounts his early career, starting at age 16 with Jim Cramer and later Tiger Management, before launching his own hedge fund. He realized his strength lay in private equity and identifying undervalued assets. This led him to the pharmaceutical sector, where he found inefficiencies in how drug rights were sold. By acquiring rights to drugs that large pharmaceutical companies had deprioritized (like Darapladib from Bristol Myers), he and his companies could repurpose them for new indications (like FSGS for Spardicarn) and bring them to market, creating significant value. This model, which he employed with companies like Retrophin, demonstrated the potential for high returns by identifying overlooked opportunities in the drug market.
The disruptive potential of AI in company formation and operations
Despite his skepticism about AI's immediate impact on drug discovery, Shkreli is highly optimistic about AI's role in founding and operating companies. He highlights his friend who left OpenAI to start "Core Automation," a company focused on AI inference models. Shkreli predicts that AI will eventually be capable of creating entire companies, such as "OpenAI Pharma" or "OpenAI Motors," essentially housing an "S&P 500" within an AI framework. He envisions a future where companies like McDonald's can leverage AI by requesting millions of GPU hours to determine optimal business strategies, such as developing new sandwich products. This, he believes, represents a significant positive return on investment, driving immense productivity gains across all industries.
Personal reflections on genius, ego, and the lessons of prison
Shkreli reflects on his own experiences, particularly how his ego and refusal to accept that he wasn't the best hedge fund manager contributed to his legal troubles. He compares this to Elizabeth Holmes and Raj Rajaratnam, suggesting that the pressure to live up to an exceptionally high self-image can be detrimental. Prison, he notes, offered a unique perspective on shared human experiences, revealing that despite diverse backgrounds, people universally experience emotions like resentment, longing, and a desire to be good parents. This realization, he suggests, humanizes everyone and underscores the importance of recognizing one's limitations and embracing alternative paths to success, such as entrepreneurship, rather than fixating on a single, narrowly defined role.
Mentioned in This Episode
●Products
●Software & Apps
●Companies
●Organizations
●Books
●Drugs & Medications
●Concepts
●People Referenced
Common Questions
Shkreli argues that quantum computing is largely a scam because its primary advantage is limited to specific algorithms, like factoring large numbers, and it does not offer faster computation for most tasks compared to classical computers like GPUs.
Topics
Mentioned in this video
The guest, known for raising drug prices and serving prison time. He discusses his views on quantum computing, investing, and lessons learned from prison.
CEO of OpenAI, who publicly stated his belief that photonic computing might be the future.
A prominent figure in quantum computing, with whom Chelsea Shkreli studied at MIT.
His thought experiment led to the invention of quantum computing, specifically for algorithms that could run faster.
Shkreli worked for Cramer's hedge fund at age 16 and learned about his trading strategies.
Mentioned by Shkreli as someone who believes Jevons paradox will lead to significant computational improvements.
Mentioned metaphorically by Shkreli in the context of potentially hindering disruptive technologies if they were to 'ruin capitalism'.
CEO of Meta, who stated the need to move fast in AI. Shkreli contrasts this with the fear-mongering of some in the AI community.
Proposed a peer review system for AI safety and is vocal about AI risks, influenced by Nick Bostrom's book.
Mentioned as an intelligent individual associated with the Effective Altruism movement.
Shkreli compares her situation to his own, highlighting the pressure of intellectual talent and the difficulty of admitting limitations.
Shkreli's hero, who inspired his interest in investing.
Shkreli mentions him as another example of a hedge fund manager who was a good investor but faced severe consequences.
CEO of NVIDIA, who Shkreli believes is wrong about the long-term prospects of GPUs and photonic computing being the future.
Shkreli advised him and invested early in Anthropic. He discusses Bankman-Fried's genius and how speed may have led him astray.
Author of 'Superintelligence', a book that has significantly influenced Elon Musk's views on AI risks.
A company that acquires cash flows from pharmaceutical products at a discount. Shkreli compares its strategy to his own.
A company whose GPUs Shkreli believes quantum computers will not surpass for most tasks, stating they won't outperform millions of NVIDIA GPUs.
Mentioned in the context of quantum computing investment and also as the author of a paper on analog photonic computing (AOC).
A research and deployment company for artificial intelligence. Shkreli's wife, Chelsea, was an early employee. OpenAI is also mentioned as being uninterested in quantum computing proposals and as a competitor in the AI space.
Mentioned as one of the companies investing heavily in quantum computing, which Shkreli criticizes for spending hundreds of millions without a clear business case.
A distribution channel for Meta, which Shkreli cites as evidence of their AI capabilities and production skills.
Mentioned as a company investing in quantum computing, which Shkreli criticizes for its lack of practical application and significant financial investment without clear returns.
Shkreli acquired a drug from Bristol Myers for $2 million, which later became a public company valued at $6 billion.
Mentioned as a company that might have seemed like a good investment in 2000, but proved to be complex and ultimately less successful than anticipated.
A startup founded by a former OpenAI employee, focused on AI for building companies, which Shkreli is excited about.
Mentioned as a company that would sell drug rights for a low price, allowing Shkreli's companies to acquire them and develop them into profitable products.
An AI company that Shkreli believes is in a defensive position due to OpenAI's influence. He would buy into it.
Shkreli built Cray computers for drug analysis, now using GPUs for the same purpose.
Used as an example of a company that could leverage AI for business decisions, such as determining its next best sandwich.
Mentioned as a company that might have seemed like a good investment in 2000, but proved to be complex and ultimately less successful than anticipated.
Mentioned in the context of a former employee going to prison for stealing physical documents, illustrating the seriousness of intellectual property theft.
Shkreli believes Meta is a strong contender in the AI race due to its distribution channels (e.g., Muse) and production skills.
A hip-hop group whose legendary album was bought by Martin Shkreli. Shkreli considered it a good album.
Massachusetts Institute of Technology, where Chelsea Shkreli studied quantum computing.
Shkreli moved to Tiger Cub after working for Jim Cramer and started his own fund, realizing he was better suited as a private equity investor and entrepreneur.
Mentioned as a traditional consulting firm that AI could potentially replace for business strategy.
NVIDIA's GPU architecture, which Shkreli predicts will be obsolete in 20 years, superseded by better methods like photonic computing.
NVIDIA's GPU architecture, which Shkreli predicts will be obsolete in 20 years, superseded by better methods like photonic computing.
NVIDIA's GPU architecture, which Shkreli predicts will be obsolete in 20 years, superseded by better methods like photonic computing.
A philosophy and movement focused on maximizing positive impact. Shkreli finds this group 'detestable' and criticizes their approach.
A concept Shkreli believes will ultimately govern technological progress, suggesting that increased efficiency leads to increased consumption, impacting NVIDIA.
Used as a metaphor by Shkreli to describe a situation where one person's innovation is quickly copied by competitors, leading to legal disputes.
A drug from which NYU was generating significant returns, which Royalty Pharma acquired.
A drug discovered and developed by Bristol Myers for hypertension, which Shkreli's company repurposed for FSGS, leading to FDA approval.
The life-saving drug whose price Martin Shkreli famously increased by 5000%, leading to widespread public condemnation.
A book by Nick Bostrom that discusses the potential risks of artificial superintelligence and has influenced figures like Elon Musk.
A famous algorithms textbook, referenced by Shkreli to illustrate that classical computers can sort playing cards faster than quantum computers for certain tasks.
More from All-In Podcast
View all 463 summaries
94 minIs Claude Conscious? Pope Rejects, Model Welfare Movement, OpenAI's Math Backlash, France Riots
83 minTrump’s Super Intelligence Summit, AI Safety Accord, GDP Beats, Midterm Predictions
52 minDaniel Ek: Life After Spotify, Broken Healthcare Incentives, Catching Disease Early & AI's Potential
95 minAnthropic IPO at Risk, Meta’s Muse Pop, Token Prices Fall, Open Source Gains Share, Alignment Fails
Ask anything from this episode.
Save it, chat with it, and connect it to Claude or ChatGPT. Get cited answers from the actual content — and build your own knowledge base of every podcast and video you care about.
Get Started Free