Logo
    Search

    Patrick Collison (Stripe CEO) - Craft, Beauty, & The Future of Payments

    enFebruary 21, 2024

    Podcast Summary

    • Not every path to success is the sameExplore different paths, find what resonates, and consider long-term study for technical careers

      While San Francisco and its cultural orientation towards independent innovation and striking out on one's own can be valuable, it may not be the best fit for everyone, especially those pursuing careers that require deep technical expertise and long-term study. Instead, these individuals may benefit from a different path, one that encourages the acquisition of foundational skills and knowledge over a longer period. The world needs diverse contributions, and not everyone should follow the same trajectory. Additionally, San Francisco's status-oriented culture may not be conducive to the deep focus and dedication required for groundbreaking discoveries and innovations. Ultimately, the advice for those in their twenties is to explore different paths and find what resonates with them, as there is no one-size-fits-all approach to success.

    • Balancing the value of established institutions and unconventional pathsFind what you care about and pursue it with a commitment to continuous learning and growth, while acknowledging the value of established institutions and the importance of learning from those who have come before us.

      The pursuit of knowledge and expertise in any field requires a balanced approach. While it's important to challenge conventional wisdom and explore unconventional paths, it's equally important to acknowledge the value of established institutions and the importance of learning from those who have come before us. In the context of the discussion, this means that while San Francisco and its entrepreneurs may be held in high regard, it's not the only path to success or the only source of valuable insights. Similarly, in the world of science, while it's important to question and challenge, it's also crucial to acquire the necessary bench skills and specific knowledge that comes from direct experience and mentorship. The herd mentality can be a powerful force, but it's important to find the gradient of maximal learning and not just follow the prevailing mood for the sake of being contrarian. Ultimately, the key is to find what you care most about and pursue that with a commitment to continuous learning and growth.

    • Balancing Importance, Difficulty, and Progress in ScienceIn science, choosing the right problems to pursue involves balancing importance, difficulty, and solvability, while also learning from high-standard environments. Flexible funding can significantly impact research programs by allowing for changes at the micro level.

      In both business and science, choosing the right path involves balancing importance, difficulty, and progress. In science, this means pursuing new problems that are important, challenging, and solvable, while also learning from mentors and environments that embody high standards. However, when it comes to funding and research, a direct correlation between investment and output is not always the case. The conversion rate between inputs and outcomes can vary greatly, and other constraints like permits, labor shortages, and regulations can also impact progress. It's important to focus on concrete, tactile problems and potential solutions at the micro scale, rather than making assumptions based solely on financial inputs. For instance, a survey of Fast Grants recipients revealed that 79% of them reported that their research programs would change significantly if they had flexible funding. This highlights the importance of considering the actual problems and potential solutions at the micro level, rather than just focusing on the amount of funding available.

    • Constraints on NIH research agendasRemoving financial constraints on NIH grantees might significantly change their research directions, but a more nuanced approach is needed to understand the complexities of bound and constrained research agendas, considering cultural, organizational, and financial factors.

      The constraint on research agendas for NIH grantees may significantly change their research directions if removed. However, focusing solely on financial frameworks to analyze this question may not be effective. Instead, it's essential to consider how bound and constrained NIH grantees should be in choosing their research agendas. The speaker is skeptical of financially oriented frameworks and suggests that there could be substantial improvements by making just one switch. The speaker also questions the constancy of US GDP growth and wonders if the US is uniquely productive or if we're adding unproductive capacity. A micro example given is the comparison of SpaceX and NASA R&D budgets, where NASA's efficacy has not fully followed the trajectory of its inputs. The speaker is not sure if the binding constraint is the number of top scientists or if it's cultural or organizational. They keep going back to the cultural and sociological point, suggesting that the issue may be more complex than just financial inputs and outputs.

    • Effective structures and practices in research groupsNew research organization ARC funds scientists to pursue their interests, offers infrastructure and career paths, fostering curiosity-driven and diverse research.

      The success of certain research groups, like the one at Caltech that produced six Nobel laureates, can't be attributed solely to the brilliance of their researchers or good fortune. Instead, they found organizational structures and cultural practices that worked effectively. However, replicating these structures in the field of biology research is a challenge. The traditional model of university-based research, with principal investigators applying for grants and managing their labs, may not be the most efficient way to produce breakthrough discoveries. ARC, a new research organization, aims to address this by funding scientists to pursue their own research interests, providing in-house infrastructure, and offering career paths for those who don't want to become lab managers. This curiosity-driven, supportive environment could lead to more diverse and groundbreaking research.

    • The impact of funding decisions on groundbreaking technologiesFunding decisions can significantly influence the development of groundbreaking technologies like CRISPR gene editing and mRNA vaccines, potentially altering their trajectory or even preventing their discovery.

      The development of groundbreaking technologies like CRISPR gene editing and mRNA vaccines might have taken a different course if funding decisions had been different. While it's impossible to know for sure what would have happened, it's plausible that these advances may not have occurred or would have been less likely in a different research environment. Looking forward, the potential applications of gene editing go beyond direct therapeutic uses. CRISPR can be used as a tool for discovery and data generation, acting as a "new kind of telescope" to understand the effects of genetic perturbations on cells. This is particularly exciting for complex diseases that are not infectious or monogenic, such as autoimmune diseases, cancers, cardiovascular disease, and neurodegenerative diseases. By figuring out the genetic interactions between genes and diseases, we may be able to shed light on their underlying causes and develop new treatments. However, it's important to consider the dual use possibilities of biotechnology and focus on minimizing risks while advancing scientific knowledge.

    • The risks of advanced technologies should not be underestimated, but the real concern might be making them more accessible.Stay humble about the future, acknowledge unknowns, and prepare for potential challenges and opportunities as technology advances.

      The potential risks of advanced technologies, such as biology and AI, should not be underestimated, as they could be used maliciously to cause harm. However, the relevant margin for concern may not be at the frontier of these technologies, but rather in making them more accessible. The world already faces numerous problems with diseases and other non-pathogenic threats, and enhancing our capabilities to deal with these issues could also help mitigate potential future threats. The unpredictability of technological development and the importance of adaptability were highlighted by the COVID-19 pandemic, emphasizing the need for competent individuals and institutions to respond effectively when crises arise. Overall, it's crucial to approach the future with humility and a recognition of the unknowns, while also being prepared for the potential challenges and opportunities that lie ahead.

    • Small initiatives can make a big impactSmall, agile initiatives can have significant results, while larger organizations may not be as quick or effective in responding to new challenges. Identifying and hiring experts in specific domains can lead to successful outcomes.

      Small, agile initiatives, like Fast Grants, can have a significant impact, even if they don't have the grandeur or impressive backing of larger institutions. The Fast Grants example, led by the speaker and colleagues, including Tyler Cohen and his wife, demonstrates this principle. While larger organizations, such as the NIH or NSF, have their merits, they may not be as quick or effective in responding to new challenges. The speaker suggests that identifying and hiring experts in specific domains could lead to more successful outcomes. Regarding the financial infrastructure for AI agents, the speaker believes that autonomous transactions are already happening in a rudimentary way and will likely continue to evolve, with decisions being made by AI agents to some degree. However, there are intriguing questions surrounding the legality, responsibility, liability, and transaction velocities of these agent transactions. The speaker emphasizes the need for careful consideration and planning as these developments unfold.

    • The economy is evolving with autonomous transactions and crypto's roleIdentifying market inefficiencies, motivation, ideas, and organizing talent can lead to successful businesses in an evolving economy.

      The economy is expected to change and expand in the next decade with the rise of autonomous transactions, and crypto could play a role in this change due to its exemption from AML regulations. The concept of Stripe as a game-changing payments company is debatable, as similar companies existed before it, but the market inefficiency and the lack of motivation, ideas, and determination to organize talent were key factors that prevented its emergence earlier. Despite the common belief that moats are overrated, Stripe and the fintech ecosystem prove that there are still effective organizations that exist, and the binding constraint on their number is largely due to motivation, ideas, and people's willingness to organize talent. Organizational and cultural moats can also exist, and Stripe is an example of a company that was able to build such a moat. Hamilton Helmer's book "7 Powers" provides insights into the various sources of market power, but it's still strange that nobody had done Stripe before Stripe. In summary, the key takeaway is that the economy is evolving, and the ability to identify and capitalize on market inefficiencies, combined with motivation, ideas, and the willingness to organize talent, can lead to the creation of successful businesses.

    • Maintaining alignment between goals and incentivesOrganizational success depends on aligning goals and incentives, even during generational turnover. Factors like a legally binding mission statement and external pressures can impact this alignment, and the debate continues on the role of shareholder capitalism in corporate longevity.

      The longevity and success of an organization often depend on its ability to maintain alignment between its stated goals and the incentives of its leadership and employees, even as generational turnover occurs. This alignment can be influenced by various factors, including the organizational structure and the external pressures it faces. For instance, organizations with a legally binding mission statement, like some Danish corporations held by nonprofit foundations, may have an advantage in staying true to their original goals. However, the role of shareholder capitalism in potentially shortening the lifespan of organizations is also a valid consideration. Ultimately, it's a complex issue with no clear-cut answer, and the balance between human aspirations and corporate longevity is an ongoing debate.

    • Understanding the historical context of trade policies and corporations' growthStripe's role in reducing learning curve for competitors might lessen the need for trade policies, but targeted policies can still benefit certain industries and Stripe Climate applies the idea of aggregating demand to expand carbon removal supply.

      The median age of corporations and the role of trade policies, such as tariffs, in shaping their growth, are complex issues with a historical context that goes beyond simple free trade versus protectionism debates. Stripe, a company facilitating global trade, could potentially reduce the learning curve for less efficient competitors, making trade policies less necessary. However, there are exceptions where targeted trade policies might be beneficial. This discussion also touched upon Stripe Climate, which applies the idea of aggregating demand to expand carbon removal supply. In 2018, Stripe identified the importance of carbon removal and started supporting the sector by contracting with carbon removal companies and forming Frontier, an advanced market commitment. Overall, this conversation emphasized the importance of understanding the historical context and nuances of economic theories and their practical applications.

    • Effective tools for technological innovation with insufficient market incentivesAdvanced Market Commitments (AMCs) raised $1bn for carbon removal tech and inspired new companies. Potential for AMCs in biomedical field for socially beneficial innovations lacking economic incentives, like mannose as oncology treatment.

      Advanced Market Commitments (AMCs) can serve as effective tools for advancing technological innovation, particularly in areas where market incentives are insufficient. The discussion highlighted the success of the Carbon Removal AMC, which raised $1 billion and inspired the creation of numerous new companies. The speaker also suggested potential applications for AMCs in the biomedical field, where socially beneficial innovations may lack sufficient economic incentives. For instance, the use of mannose as a potential oncology treatment is an example of a promising yet unpatentable innovation that could benefit from an AMC. Additionally, the speaker touched upon the importance of balancing craft, scale, and speed in business, with some industries, such as luxury goods, demonstrating that appreciation for craft and beauty can coexist with large-scale growth.

    • The significance of craft and aesthetics in businessStrive for high-quality work and design to attract top talent and create successful businesses, even if it's not easily quantifiable.

      The importance of craft and aesthetics in business, though often unquantifiable, significantly influences customer perception and the quality of work. Stripe, a company that sells primarily to businesses, has observed this trend and believes that taking the practice of work seriously is crucial for attracting and retaining top talent. The distinction between interface and implementation can vary depending on the nature of the business. For instance, a simple interface might be more important for a social media platform like TikTok, whereas Stripe prioritizes a well-designed architecture to provide customers with the ability to do certain things. The idea of a self-contained universe for modeling is a useful analogy for understanding Stripe's approach. While there are differences between the two, the concept of a transaction, which is fundamental to both, provides a connection. Ultimately, the pursuit of craft and aesthetics is crucial for creating a successful business, even if it's not easily quantifiable.

    • The Importance of Good API Design for Long-Term SuccessWell-designed APIs and architectures can provide compounding benefits for decades, while poorly designed ones can lead to stagnation. Companies like Stripe and Mathematica prioritize multidecadal abstractions, considering both immediate needs and long-term implications.

      Good API design and architecture play a significant role in the longevity and success of platforms. When done right, it can provide compounding benefits over multiple decades, even in the face of rapid technological changes. For example, UNIX and Mathematica have shown that well-designed architectures can endure for decades. Conversely, getting it wrong can lead to stagnation. Companies like Stripe and Mathematica prioritize multidecadal abstractions, considering not just the immediate needs but also the long-term implications and potential future evolutions. The card networks, such as Visa and Mastercard, are examples of successful architectures that have evolved to address various use cases over the years. The person responsible for designing the card networks, Dee Hock, was remarkable for his foresight and ability to create a design that could adapt to various use cases. However, it's important to remember that even the best designs don't always get it right, and there are trade-offs to consider.

    • Understanding Interchange Fees' True ValueInterchange fees are about more than just revenue sharing; they also involve distribution incentives, credit issuance, and customer support. Critics should consider the benefits of card networks before dismissing their value.

      While Stripe may appear to be giving away interchange revenue to other companies, it's essential to consider the full picture. Interchange is not just about revenue sharing but also about distribution incentives, credit issuance, and customer support. If we look at countries where card networks didn't arise, like Germany, we can appreciate the benefits of what DeHaan, Visa, Mastercard, and the card networks have accomplished. Critiques of interchange are most valuable when they come from experts who have studied the payment ecosystems of other countries. If we were to design payments from scratch today, it's unclear if the current costs, including credit risk, fraud, dispute handling, and customer support, should amount to 2-3% of each transaction. Central banks in various countries, like Brazil, India, Switzerland, and Japan, are experimenting with designing national payment schemes from scratch, and their rapid adoption rates suggest that this reinvention could change the payment landscape significantly. It's important to remember that a significant portion of interchange fees goes towards consumer rewards, not just covering costs.

    • Understanding the complexities of transaction taxes and evolving tax structuresBusinesses operating internationally face significant challenges from transaction taxes and evolving tax structures, requiring a deep understanding of various jurisdictions and their requirements to navigate these complexities and minimize financial impact.

      The economics of transaction taxes and the evolving landscape of tax laws pose complex challenges for businesses, particularly those operating internationally. While the specific transaction fees may not be massively inefficient, the changing tax structures, including the imposition of sales taxes on businesses without a physical presence in a jurisdiction, add a layer of complexity that goes beyond transactional economics. These taxes can amount to significant percentages, making it essential for businesses to understand and navigate the intricacies of various jurisdictions and their tax requirements. Stripe's role in facilitating business operations, including payment processing and identity fraud detection, has proven to be a valuable complement to these challenges, enabling businesses to focus on their core operations while Stripe handles the financial complexities. Looking ahead, Stripe's mission to help businesses manage and orchestrate everything pertaining to their revenue will likely expand to accommodate an increasing number of relevant business functions.

    • Stripe's focus on global money orchestrationStripe's approach to solving complex challenges in global money transactions, by plugging into existing systems and enhancing their capabilities, creates significant value and impacts the world.

      Stripe, as a company, has focused on solving the complex and often overlooked challenges in the realm of global money orchestration, rather than just offering a pretty interface for financial transactions. This approach, which involves plugging into existing systems and enhancing their capabilities, has the potential to create significant value and make a substantial impact on the world. Additionally, the company values both the writer and the reader in its literate culture, recognizing the benefits of organizing thoughts through text for both parties. The textual nature of Stripe's approach to knowledge sharing, inspired by the printing revolution, allows for easier identification of discrepancies between theories and reality, leading to continuous improvement and innovation.

    • Productivity gains from AI in writing and collaboration culturesAI integration in production services and workflows can be complex, but offers significant productivity gains for organizations with heavy writing and collaboration cultures. Prioritizing reliability and uptime is crucial for businesses dealing with large portions of the global economy.

      Organizations with a heavy writing and collaboration culture may experience significant productivity gains from AI, particularly in the form of Language Models (LLMs). However, the integration of LLMs into production services and workflows can be complex, requiring collaboration tools and a centralized system for managing and experimenting with different models. The high stakes nature of businesses like Stripe, which deal with significant portions of the global economy, necessitates a focus on reliability and uptime, making the efficient integration of AI a priority. Additionally, the flexibility of LLMs can be particularly valuable in dealing with the uncertainty and edge cases inherent in human interactions. Overall, the effective implementation of AI in organizations with a strong writing and collaboration culture can lead to high returns, but requires careful planning and infrastructure development.

    • Stripe's Significant Economic Impact and High Operational ExcellenceStripe contributes $1 trillion annually to the global economy with a 99.995% reliable platform, emphasizing process, reliability, and continuous improvement for software development and customer experience.

      Stripe, a tech company, contributes around 1% of the global GDP, approximately $1 trillion per year, by facilitating mostly final goods transactions. This significant economic impact comes with great responsibility, as Stripe aims for high reliability and minimal downtime, deploying services around 1000 times a day with 99.995% reliability, translating to about 100 minutes of unavailability annually. This level of operational excellence requires substantial investment and cultural emphasis on process and reliability, a value that may not be universally appreciated in the tech industry. Stripe's success lies in its commitment to defining and measuring desired outcomes, implementing automated systems, and continuously improving the system to minimize deviations and prevent potential issues. This dedication to maintaining high standards for software development and customer experience is crucial for organizations aiming to build and deliver the best software at scale.

    • Stripe's growth fueled by global Internet economy expansion and company's increasing share within itStripe's growth is driven by the expanding global Internet economy and the company's growing presence within it. Opportunities for efficiency gains, extending capital, and helping businesses navigate international markets contribute to Stripe's continued growth. Exponentially growing startups and optimizing business growth are key focus areas.

      Stripe's growth is driven by both the expansion of the global Internet economy and Stripe's increasing share within it. Stripe serves customers whose growth outpaces the Internet economy as a whole, and there's still significant room for growth before Stripe's activity approaches the ceiling of global economic growth. Additionally, there are vast opportunities for efficiency gains and optimizations in the business world, particularly in areas like extending capital to businesses and helping them navigate international markets. Exponentially growing startups are also crucial for Stripe's continued growth, as they contribute to the overall expansion of the digital economy. The basic questions surrounding business growth, such as why some companies succeed while others don't, remain largely unanswered, and Stripe aims to tackle these questions by providing software solutions and services. Overall, Stripe's mission is to optimize the meta system of business, making it more efficient and connected, and help businesses of all sizes grow and reach their full potential.

    • Stripe's Growth and Collaboration with Established BusinessesStripe's growth may not match the economy's rate, but partnerships with established businesses lead to innovation, consumer surplus, and economic growth.

      Stripe's growth may eventually match the economy's rate, but the company has a long way to go due to numerous optimization opportunities. Additionally, established businesses are significant sources of innovation, often improving existing technologies and processes. John Warrillow emphasizes the underrated role of big businesses in innovation. Stripe collaborates with established businesses to help them sell existing products in new markets or through more convenient methods, leading to consumer surplus and economic growth. Regarding John's personal life, Charlie Munger acknowledged the similarities between John and Warren Buffett's intense, lengthy partnerships, but there's no definitive record of Munger commenting on their relationship specifically. From observing John's marriage, we can learn about the importance of a coequal, intense partnership in both personal and professional contexts.

    • Working with close friends and partnersStarting a business with enduring relationships can lead to significant success and personal fulfillment, as each individual brings unique strengths to the table.

      Working with close friends and partners can lead to significant and fulfilling ventures. The speaker, who has had experience starting businesses with close associates, including Patrick Collison, Silvana Martinez, Tyler Cowen, and John Collison, emphasizes the importance of enduring relationships in business. Despite the common advice against working with friends or partners, the speaker's experience has been rewarding and complementary. Each individual brings unique strengths to the table, and the companies would be less effective without them. The speaker also highlights the personal and emotional fulfillment that comes from working with people you care about. Overall, the speaker's message is that the potential benefits of working with close associates outweigh the potential risks.

    Recent Episodes from Dwarkesh Podcast

    Tony Blair - Life of a PM, The Deep State, Lee Kuan Yew, & AI's 1914 Moment

    Tony Blair - Life of a PM, The Deep State, Lee Kuan Yew, & AI's 1914 Moment

    I chatted with Tony Blair about:

    - What he learned from Lee Kuan Yew

    - Intelligence agencies track record on Iraq & Ukraine

    - What he tells the dozens of world leaders who come seek advice from him

    - How much of a PM’s time is actually spent governing

    - What will AI’s July 1914 moment look like from inside the Cabinet?

    Enjoy!

    Watch the video on YouTube. Read the full transcript here.

    Follow me on Twitter for updates on future episodes.

    Sponsors

    - Prelude Security is the world’s leading cyber threat management automation platform. Prelude Detect quickly transforms threat intelligence into validated protections so organizations can know with certainty that their defenses will protect them against the latest threats. Prelude is backed by Sequoia Capital, Insight Partners, The MITRE Corporation, CrowdStrike, and other leading investors. Learn more here.

    - This episode is brought to you by Stripe, financial infrastructure for the internet. Millions of companies from Anthropic to Amazon use Stripe to accept payments, automate financial processes and grow their revenue.

    If you’re interested in advertising on the podcast, check out this page.

    Timestamps

    (00:00:00) – A prime minister’s constraints

    (00:04:12) – CEOs vs. politicians

    (00:10:31) – COVID, AI, & how government deals with crisis

    (00:21:24) – Learning from Lee Kuan Yew

    (00:27:37) – Foreign policy & intelligence

    (00:31:12) – How much leadership actually matters

    (00:35:34) – Private vs. public tech

    (00:39:14) – Advising global leaders

    (00:46:45) – The unipolar moment in the 90s



    Get full access to Dwarkesh Podcast at www.dwarkeshpatel.com/subscribe
    Dwarkesh Podcast
    enJune 26, 2024

    Francois Chollet, Mike Knoop - LLMs won’t lead to AGI - $1,000,000 Prize to find true solution

    Francois Chollet, Mike Knoop - LLMs won’t lead to AGI - $1,000,000 Prize to find true solution

    Here is my conversation with Francois Chollet and Mike Knoop on the $1 million ARC-AGI Prize they're launching today.

    I did a bunch of socratic grilling throughout, but Francois’s arguments about why LLMs won’t lead to AGI are very interesting and worth thinking through.

    It was really fun discussing/debating the cruxes. Enjoy!

    Watch on YouTube. Listen on Apple PodcastsSpotify, or any other podcast platform. Read the full transcript here.

    Timestamps

    (00:00:00) – The ARC benchmark

    (00:11:10) – Why LLMs struggle with ARC

    (00:19:00) – Skill vs intelligence

    (00:27:55) - Do we need “AGI” to automate most jobs?

    (00:48:28) – Future of AI progress: deep learning + program synthesis

    (01:00:40) – How Mike Knoop got nerd-sniped by ARC

    (01:08:37) – Million $ ARC Prize

    (01:10:33) – Resisting benchmark saturation

    (01:18:08) – ARC scores on frontier vs open source models

    (01:26:19) – Possible solutions to ARC Prize



    Get full access to Dwarkesh Podcast at www.dwarkeshpatel.com/subscribe
    Dwarkesh Podcast
    enJune 11, 2024

    Leopold Aschenbrenner - China/US Super Intelligence Race, 2027 AGI, & The Return of History

    Leopold Aschenbrenner - China/US Super Intelligence Race, 2027 AGI, & The Return of History

    Chatted with my friend Leopold Aschenbrenner on the trillion dollar nationalized cluster, CCP espionage at AI labs, how unhobblings and scaling can lead to 2027 AGI, dangers of outsourcing clusters to Middle East, leaving OpenAI, and situational awareness.

    Watch on YouTube. Listen on Apple PodcastsSpotify, or any other podcast platform. Read the full transcript here.

    Follow me on Twitter for updates on future episodes. Follow Leopold on Twitter.

    Timestamps

    (00:00:00) – The trillion-dollar cluster and unhobbling

    (00:20:31) – AI 2028: The return of history

    (00:40:26) – Espionage & American AI superiority

    (01:08:20) – Geopolitical implications of AI

    (01:31:23) – State-led vs. private-led AI

    (02:12:23) – Becoming Valedictorian of Columbia at 19

    (02:30:35) – What happened at OpenAI

    (02:45:11) – Accelerating AI research progress

    (03:25:58) – Alignment

    (03:41:26) – On Germany, and understanding foreign perspectives

    (03:57:04) – Dwarkesh’s immigration story and path to the podcast

    (04:07:58) – Launching an AGI hedge fund

    (04:19:14) – Lessons from WWII

    (04:29:08) – Coda: Frederick the Great



    Get full access to Dwarkesh Podcast at www.dwarkeshpatel.com/subscribe
    Dwarkesh Podcast
    enJune 04, 2024

    John Schulman (OpenAI Cofounder) - Reasoning, RLHF, & Plan for 2027 AGI

    John Schulman (OpenAI Cofounder) - Reasoning, RLHF, & Plan for 2027 AGI

    Chatted with John Schulman (cofounded OpenAI and led ChatGPT creation) on how posttraining tames the shoggoth, and the nature of the progress to come...

    Watch on YouTube. Listen on Apple PodcastsSpotify, or any other podcast platform. Read the full transcript here. Follow me on Twitter for updates on future episodes.

    Timestamps

    (00:00:00) - Pre-training, post-training, and future capabilities

    (00:16:57) - Plan for AGI 2025

    (00:29:19) - Teaching models to reason

    (00:40:50) - The Road to ChatGPT

    (00:52:13) - What makes for a good RL researcher?

    (01:00:58) - Keeping humans in the loop

    (01:15:15) - State of research, plateaus, and moats

    Sponsors

    If you’re interested in advertising on the podcast, fill out this form.

    * Your DNA shapes everything about you. Want to know how? Take 10% off our Premium DNA kit with code DWARKESH at mynucleus.com.

    * CommandBar is an AI user assistant that any software product can embed to non-annoyingly assist, support, and unleash their users. Used by forward-thinking CX, product, growth, and marketing teams. Learn more at commandbar.com.



    Get full access to Dwarkesh Podcast at www.dwarkeshpatel.com/subscribe
    Dwarkesh Podcast
    enMay 15, 2024

    Mark Zuckerberg - Llama 3, Open Sourcing $10b Models, & Caesar Augustus

    Mark Zuckerberg - Llama 3, Open Sourcing $10b Models, & Caesar Augustus

    Mark Zuckerberg on:

    - Llama 3

    - open sourcing towards AGI

    - custom silicon, synthetic data, & energy constraints on scaling

    - Caesar Augustus, intelligence explosion, bioweapons, $10b models, & much more

    Enjoy!

    Watch on YouTube. Listen on Apple PodcastsSpotify, or any other podcast platform. Human edited transcript with helpful links here.

    Timestamps

    (00:00:00) - Llama 3

    (00:08:32) - Coding on path to AGI

    (00:25:24) - Energy bottlenecks

    (00:33:20) - Is AI the most important technology ever?

    (00:37:21) - Dangers of open source

    (00:53:57) - Caesar Augustus and metaverse

    (01:04:53) - Open sourcing the $10b model & custom silicon

    (01:15:19) - Zuck as CEO of Google+

    Sponsors

    If you’re interested in advertising on the podcast, fill out this form.

    * This episode is brought to you by Stripe, financial infrastructure for the internet. Millions of companies from Anthropic to Amazon use Stripe to accept payments, automate financial processes and grow their revenue. Learn more at stripe.com.

    * V7 Go is a tool to automate multimodal tasks using GenAI, reliably and at scale. Use code DWARKESH20 for 20% off on the pro plan. Learn more here.

    * CommandBar is an AI user assistant that any software product can embed to non-annoyingly assist, support, and unleash their users. Used by forward-thinking CX, product, growth, and marketing teams. Learn more at commandbar.com.



    Get full access to Dwarkesh Podcast at www.dwarkeshpatel.com/subscribe

    Sholto Douglas & Trenton Bricken - How to Build & Understand GPT-7's Mind

    Sholto Douglas & Trenton Bricken - How to Build & Understand GPT-7's Mind

    Had so much fun chatting with my good friends Trenton Bricken and Sholto Douglas on the podcast.

    No way to summarize it, except: 

    This is the best context dump out there on how LLMs are trained, what capabilities they're likely to soon have, and what exactly is going on inside them.

    You would be shocked how much of what I know about this field, I've learned just from talking with them.

    To the extent that you've enjoyed my other AI interviews, now you know why.

    So excited to put this out. Enjoy! I certainly did :)

    Watch on YouTube. Listen on Apple PodcastsSpotify, or any other podcast platform. 

    There's a transcript with links to all the papers the boys were throwing down - may help you follow along.

    Follow Trenton and Sholto on Twitter.

    Timestamps

    (00:00:00) - Long contexts

    (00:16:12) - Intelligence is just associations

    (00:32:35) - Intelligence explosion & great researchers

    (01:06:52) - Superposition & secret communication

    (01:22:34) - Agents & true reasoning

    (01:34:40) - How Sholto & Trenton got into AI research

    (02:07:16) - Are feature spaces the wrong way to think about intelligence?

    (02:21:12) - Will interp actually work on superhuman models

    (02:45:05) - Sholto’s technical challenge for the audience

    (03:03:57) - Rapid fire



    Get full access to Dwarkesh Podcast at www.dwarkeshpatel.com/subscribe

    Demis Hassabis - Scaling, Superhuman AIs, AlphaZero atop LLMs, Rogue Nations Threat

    Demis Hassabis - Scaling, Superhuman AIs, AlphaZero atop LLMs, Rogue Nations Threat

    Here is my episode with Demis Hassabis, CEO of Google DeepMind

    We discuss:

    * Why scaling is an artform

    * Adding search, planning, & AlphaZero type training atop LLMs

    * Making sure rogue nations can't steal weights

    * The right way to align superhuman AIs and do an intelligence explosion

    Watch on YouTube. Listen on Apple PodcastsSpotify, or any other podcast platform. Read the full transcript here.

    Timestamps

    (0:00:00) - Nature of intelligence

    (0:05:56) - RL atop LLMs

    (0:16:31) - Scaling and alignment

    (0:24:13) - Timelines and intelligence explosion

    (0:28:42) - Gemini training

    (0:35:30) - Governance of superhuman AIs

    (0:40:42) - Safety, open source, and security of weights

    (0:47:00) - Multimodal and further progress

    (0:54:18) - Inside Google DeepMind



    Get full access to Dwarkesh Podcast at www.dwarkeshpatel.com/subscribe

    Patrick Collison (Stripe CEO) - Craft, Beauty, & The Future of Payments

    Patrick Collison (Stripe CEO) - Craft, Beauty, & The Future of Payments

    We discuss:

    * what it takes to process $1 trillion/year

    * how to build multi-decade APIs, companies, and relationships

    * what's next for Stripe (increasing the GDP of the internet is quite an open ended prompt, and the Collison brothers are just getting started).

    Plus the amazing stuff they're doing at Arc Institute, the financial infrastructure for AI agents, playing devil's advocate against progress studies, and much more.

    Watch on YouTube. Listen on Apple PodcastsSpotify, or any other podcast platform. Read the full transcript here. Follow me on Twitter for updates on future episodes.

    Timestamps

    (00:00:00) - Advice for 20-30 year olds

    (00:12:12) - Progress studies

    (00:22:21) - Arc Institute

    (00:34:27) - AI & Fast Grants

    (00:43:46) - Stripe history

    (00:55:44) - Stripe Climate

    (01:01:39) - Beauty & APIs

    (01:11:51) - Financial innards

    (01:28:16) - Stripe culture & future

    (01:41:56) - Virtues of big businesses

    (01:51:41) - John



    Get full access to Dwarkesh Podcast at www.dwarkeshpatel.com/subscribe

    Tyler Cowen - Hayek, Keynes, & Smith on AI, Animal Spirits, Anarchy, & Growth

    Tyler Cowen - Hayek, Keynes, & Smith on AI, Animal Spirits, Anarchy, & Growth

    It was a great pleasure speaking with Tyler Cowen for the 3rd time.

    We discussed GOAT: Who is the Greatest Economist of all Time and Why Does it Matter?, especially in the context of how the insights of Hayek, Keynes, Smith, and other great economists help us make sense of AI, growth, animal spirits, prediction markets, alignment, central planning, and much more.

    The topics covered in this episode are too many to summarize. Hope you enjoy!

    Watch on YouTube. Listen on Apple PodcastsSpotify, or any other podcast platform. Read the full transcript here. Follow me on Twitter for updates on future episodes.

    Timestamps

    (0:00:00) - John Maynard Keynes

    (00:17:16) - Controversy

    (00:25:02) - Fredrick von Hayek

    (00:47:41) - John Stuart Mill

    (00:52:41) - Adam Smith

    (00:58:31) - Coase, Schelling, & George

    (01:08:07) - Anarchy

    (01:13:16) - Cheap WMDs

    (01:23:18) - Technocracy & political philosophy

    (01:34:16) - AI & Scaling



    Get full access to Dwarkesh Podcast at www.dwarkeshpatel.com/subscribe

    Lessons from The Years of Lyndon Johnson by Robert Caro [Narration]

    Lessons from The Years of Lyndon Johnson by Robert Caro [Narration]

    This is a narration of my blog post, Lessons from The Years of Lyndon Johnson by Robert Caro.

    You read the full post here: https://www.dwarkeshpatel.com/p/lyndon-johnson

    Listen on Apple Podcasts, Spotify, or any other podcast platform. Follow me on Twitter for updates on future posts and episodes.



    Get full access to Dwarkesh Podcast at www.dwarkeshpatel.com/subscribe

    Related Episodes

    Business Rap: December 22, 2017

    Business Rap: December 22, 2017
    Find out how to find the right Investor for your business with top experts; Mike Murshak of Loomis Law, Paul Jaques of Spartan Innovations and Thomas Stewart of Conquer Accelerator. Get the scoop on how to handle chronically late employees on "Ask Phil" with Phil Zeller of Dale Carnegie - Michigan, and get Chad Livengood of Crain’s Detroit Business' reaction to Detroit Free Press' decision to fire Pulitzer Prize winning journalist Stephen Henderson in an interview with Michael Patrick Shiels.

    #38 Thibauld Favre (Fairmint.co) - Le "Continuous Financing", une alternative au VC?

    #38 Thibauld Favre (Fairmint.co) - Le "Continuous Financing", une alternative au VC?

    “Le principe du Continuous Financing est i) de permettre à tout un chacun (stakeholder, client, grand public) d’investir dans votre société à n’importe quel moment et ii) de permettre à vos investisseurs la liquidité de leurs investissements, …  et cela est aujourd’hui possible grâce à la technologie de la blockchain”

    Cette semaine 🌟

    Cette semaine nous échangeons avec Thibauld Favre, CEO de Fairmint.co, plateforme qui permet aux entreprises de mettre en place le “Continuous Financing” pour son entreprise et ainsi de recevoir des financements de manière globale, continue et sécurisée, directement depuis leur application ou leur site web.

    Dans cet épisode Thibauld revient sur son parcours d’entrepreneur, d’abord entrepreneur startup avec AllMyApps (le 1er app store pour PC! 😮), puis entrepreneur non-profit avec LaPrimaire.org 🗳️ avant de nous expliquer pourquoi il a lancé il y a quelques mois Fairmint.co afin de rendre accessible et démocratiser le Continuous Financing pour les entreprises en croissance.

    Retrouvez Thibauld sur Twitter ou sur linkedin!

    Pour recevoir toutes les semaines chaque nouvelle épisode dans votre boîte mail vous pouvez vous inscrire avec le lien ci-dessous! 🔥💫

    Subscribe now

    Au cours de cet épisode, nous discutons 💬

    ⚡ Des key learnings de l’aventure AllMyApps, le 1er appstore pour PC!

    ⚡ De l’importance d’initiative comme LaPrimaire.org pour l’avenir de la démocratie

    ⚡ Comment fonctionne le Continuous Financing et pourquoi ce modèle est vertueux pour le financement d’une entreprise

    ⚡ Comment la stack technique de la blockchain rend ce processus d’investissement possible et sécurisé

    ⚡ Quelle est la typologie d’entreprises concernée par ce type de financement?

    ⚡ Comment ce modèle est compatible avec les financements classiques (VC, dettes, etc…)

    ⚡ Pourquoi la finance va être bousculée et ré-inventée par la technologie blockchain

    Si vous avez aimé ce post, pourquoi ne pas le partager dans votre réseau ? 💌🤗

    Share Equity 101

    Liens, outils et ressources 📚


    Encore là? 😉

    Si vous appréciez ce podcast ou souhaitez me soutenir, il suffit d’aller sur Apple Podcast et de laisser une note et un petit commentaire, ça prend 30 secondes et cela m’aide ÉNORMÉMENT pour faire découvrir Equity 101 à d’autres personnes!

    > Donner une note au podcast

    Les épisodes précédents

    🎙️ #36 #37 - Leila Nicolas (GO Capital) - L’investissement dans les biotechs

    🎙️ #35 - Guillaume Moubeche (Lemlist) - L’hypercroissance sans lever de fonds!

    🎙️ #34 - David Salabi (Cambon Partners) - Comment préparer un exit?

    🎙️ #33 - Jean-David Chamboredon (ISAI) - Comment organiser un board efficace?

    🎙️ #32 - Julien Etaix (Upfront) - Les clés de lecture de l'investissement du Corporate Venture (CVC)

    Pour recevoir toutes les semaines chaque nouvelle épisode dans votre boîte mail vous pouvez vous inscrire avec le lien ci-dessous! 🔥💫

     Subscribe now  

    #44 Alban Denoyel (Sketchfab) Romain Vidal (Cap Horn) - La relation Founders - VC

    #44 Alban Denoyel (Sketchfab) Romain Vidal (Cap Horn) - La relation Founders - VC

     

    C’est reparti pour un épisode hors série, je vous rappelle le principe on réunit un investisseur VC et un entrepreneur pour discuter, échanger, confronter les points de vue et aller plus loin sur un sujet donné! 

    Et pour animer le débat j’ai le plaisir d’avoir à mes côtés Julien Petit, pour ceux qui ne le connaissent pas encore il faut vite aller s’inscrire à sa newsletter “The VC insider”!  

    Cette semaine 🎙️🌟

    “C'est en intégrant le programme Techstars que l’on nous a vraiment formé à la construction de la relation avec les VC; c’est maintenant devenu une “best practice” mais en tant que first time entrepreneur je n’avais pas tous ces éléments naturellement en tête;  Techstars nous a formé au follow-up, à construire et maintenir une relation, partager ses progrès, ses KPI de façon régulière avec tous les gens qui peuvent être des potentiels investisseurs. Pour le VC “time is the best due diligence”, plus il a le temps d’observer une équipe et des fondateurs executer une vision, plus il peut dérisquer son investissement.”

    Pour écouter l’épisode : iTunes | Spotify| Google| Deezer| Ausha

    Pour ce second épisode hors série nous avons la chance d’avoir deux invités de luxe, Romain Vidal, Partner chez Cap Horn et Alban Denoyel, CEO de Sketchfab.

    Romain a démarré sa carrière en tant qu’entrepreneur avant de rejoindre le fonds Cap Horn à sa création en 2010, Cap Horn est un fonds de capital amorçage / Series A / Series B fondé en 2010, qui investit dans les startups françaises B2B et notamment dans Ledger, Critizr, Finalcad, Miuros… Pour en savoir plus sur Romain vous pouvez retrouvez l’épisode fait ensemble au mois de mai dans lequel nous parlons collaboration startups - grands groupes notamment. 

    Alban est entrepreneur depuis plus de 10 ans et à lancé Sketchfab en 2012, c’est aujourd’hui la plateforme de référence pour héberger, publier et trouver des plans 3D pour l’impression 3D, la réalité virtuelle et la réalité augmentée. Créée en 2012 la société est aujourd’hui basée à Paris et New York et a réalisé 3 tours de financement, un premier tour de seed début 2013 (400k€) auprès de Business Angels, puis une Serie A (1,5M€) avec Balderton & Partech fin 2013 et enfin une Serie B en 2015 avec FirstMark Capital 👏.

    Dans cette épisode nous avons pris le temps d’échanger et partager les points de vue d’un investisseur et d’un entrepreneur sur le sujet de la relation Entrepreneurs / Fondateurs avec les VCs, pourquoi c’est un sujet crucial pour le développement et la croissance de l'entreprise, comment on construit une relation, comment gérer les relations et les divergences de points de vue.

    Retrouvez Romain sur Linkedin ou sur Twitter et Alban sur Linkedin ou sur Twitter !

    Pour recevoir toutes les semaines chaque nouvelle épisode dans votre boîte mail vous pouvez vous inscrire avec le lien ci-dessous! 🔥💫

    Subscribe now

    Au cours de cet épisode, nous discutons 💬

    ⚡ Des enjeux de la relation VC - founders dans la trajectoire d’une startup

    ⚡ L’importance de la proximité géographique entre le VC et le founder notamment dans le contexte COVID

    ⚡ Comment construire la relation pre-investissement, pourquoi c’est important, quel est l’importance du fit relationnel dans un investissement

    ⚡ Quelles informations partage-t-on avec son VC? Quel niveau de transparence?

    ⚡ Quels sont les sujets qui peuvent générer des désalignements intérêt?  

    ⚡ Que penser du support proposé par les VCs?

    ⚡ Comment gérer les relations VC-founders dans le cadre d’un board?

    Si vous avez aimé cet épisode, pourquoi ne pas le partager dans votre réseau ? 💌 et si vous voulez supporter le projet vous pouvez m’offrir un café☕🤗

    Share Equity 101

    Encore là? 😉

    Si vous appréciez ce podcast ou souhaitez me soutenir, il suffit d’aller sur Apple Podcast et de laisser une note et un petit commentaire, ça prend 30 secondes et cela m’aide ÉNORMÉMENT pour faire découvrir Equity 101 à d’autres personnes!

    > Donner une note au podcast

    Les épisodes précédents

    🎙️ #43 - Pierre-Antoine Dusoulier (iBanFirst) - “Choisir ses batailles pour être imbattable”

    🎙️ #42 - Jean Lebrument (Brigad) - Mettre en place les OKR dans son organisation

    🎙️ #41 - Ilan Abehassera (Willo & Diaspora Ventures) - L'investissement Business Angel

    🎙️ #40 - Réza Malekzadeh (Partech) & Benjamin Zenou (Simplifield) - Quand et comment partir aux US?

    🎙️ #39 - Benjamin Joffe (SOSV) - l'investissement dans le Hardware

    Pour recevoir toutes les semaines chaque nouvelle épisode dans votre boîte mail vous pouvez vous inscrire avec le lien ci-dessous! 🔥💫

    Subscribe now