**1st Annual Workshop @MIT Sloan** # **Bayesian Entrepreneurship Day 1: What is Bayesian Entrepreneurship** The conversation revolved around the importance of creating a community for Bayesian entrepreneurship, with speakers emphasizing the intersection of Bayesian learning and entrepreneurship. The workshop presented a research program exploring the application of Bayesian entrepreneurship. The speakers discussed challenges in intrapreneurship, including language and standardization, and the role of experiments in creating a common language and conceptual framework. They also explored the intersection of entrepreneurship and innovation, with a focus on the role of theory and practice. The speakers highlighted the importance of updating probabilities based on new information and using theories to guide decision-making. They also discussed the challenges of balancing opportunity and risk in entrepreneurship, with a focus on the significance of experimentation and prior knowledge. ## **Transcript** [https://otter.ai/u/GmuseONdv1IK6vyKH6MO-A2bWm4?utm_source=copy_url](https://otter.ai/u/GmuseONdv1IK6vyKH6MO-A2bWm4?utm_source=copy_url) ## **Suggested Items** * [ ] Share data from randomized control trials on Bayesian entrepreneurship for other researchers to explore * [ ] Look into philosophical literature on abduction to better integrate Bayesian and non-Bayesian perspectives * [ ] Investigate what types of probabilities (e.g. Knightian uncertainty vs frequentist) are being referenced in critiques of Bayesian approaches * [ ] Explore effectuation further to find common ground between Bayesian and non-Bayesian frameworks * [ ] Develop clearer links between the Bayesian entrepreneurship theory and analysis of observational data * [ ] Provide more details on boundary conditions between Bayesian and non-Bayesian approaches to avoid polarization * [ ] Consider how Bayesian entrepreneurship may apply differently across types of science and technology innovations * [ ] Examine the selection process into entrepreneurship to understand differences between optimistic and contrarian entrepreneurs * [ ] Look at how entrepreneurial processes may evolve over the business lifecycle from more flexible to more stable ## **Outline** ### **Bayesian entrepreneurship with academics.** * Group of experts gather to discuss Bayesian entrepreneurship and its applications. ### **Entrepreneurship and language barriers in a conference.** * Alfonso reflects on the success of the conference, citing a grand plan and experiment to create a community and common language for discussing a fragmented field. * Alfonso discusses the challenges of understanding and addressing complex issues in entrepreneurship, highlighting the need for a common language to facilitate communication and decision-making. * Alfonso suggests that entrepreneurs often face difficult choices that require a deep understanding of various factors, and that digging into these decisions can provide valuable insights. ### **Using probabilities in decision-making.** * Alfonso discusses the importance of a standard language in communication and decision-making, highlighting its potential downsides and the need for analytics and excellent descriptions of phenomena. * Alfonso presents a framework for vision entrepreneurship, emphasizing the importance of parsimonious comprehensiveness and provoking discussion. ### **Entrepreneurship and startups.** * Scott discusses Bayesian entrepreneurship, a framework that helps bring together various ideas and perspectives within a common frame. * Scott uses the example of Zappos, a company founded by Nick Swinmurn, to illustrate how this framework can be applied to almost any startup. ### **Entrepreneurship, experimentation, and persuasion.** * Scott hypothesizes that investors thought the shoe industry was stupid, but then found two investors who shared their view and invested their own personal cash. * Tony Hsieh presented his own internal numbers to Sequoia, but they were unpersuasive, leading Hsieh to design an experiment to demonstrate value to real customers. * Scott shares a story about Zappos’ early days, highlighting the importance of entrepreneurs choosing their own ideas and how it impacts experimentation, persuasion, and strategy. * The speaker emphasizes that entrepreneurs get to choose one thing for sure - their idea - and how it affects other aspects of the business, creating a unique dynamic that is worth exploring. ### **Entrepreneurship, opportunity identification, and contrarian ideas.** * Scott emphasizes the importance of choosing the right opportunity, citing the vast majority of papers focus on ideation and learning but neglect the decision-making process. * Scott highlights the need for a balance between experimentation and learning, with the goal of building a successful business that earns more money than costs. * Scott emphasizes the importance of contrarian ideas in entrepreneurship, citing Peter Thiel as an example. * The speaker discusses the value of optimism in entrepreneurship and that entrepreneurs must be optimistic about their ideas. **Entrepreneurship, probability, and decision-making.** * Shannon emphasizes the importance of understanding the phenomenon that current theories cannot fully account for to understand the different perspectives in entrepreneurial decision-making. * Scott discusses the limited literature on entrepreneurship, focusing on the impact of the microeconomic and institutional environment on decisions. * Scott highlights the challenge of abstracting away from the subjective level of entrepreneur beliefs, while also acknowledging the productive approach of entrepreneurial finance in taking an abstracted view. **Entrepreneurship, experimentation, and opportunity perception.** * Scott discusses the idea of “agreeing to disagree” and the importance of recognizing that people have different priors and perspectives. * Scott highlights the work of Eric Van den Steen, who has been pushing the idea that people are heterogeneous in their priors and that this should be taken into account when studying problems. * Scott discusses the importance of purposeful experimentation in entrepreneurship, highlighting the work of Mary Tripsas and Fiona Murray (2004). ### **Using Bayesian inference for entrepreneurial decision-making.** * Scott discusses the concept of priors in decision-making, explaining that entrepreneurs come with relatively optimistic priors and that this can have a direct impact on their choices and outcomes. * The speaker references Ramana, who would be happy with the experiment setup, as the actual state of the world can be either success or failure, with a signal s (bad news or good news) and probability lambda. * Scott discusses the importance of prior beliefs in shaping the demand for experiments, highlighting how one's prior beliefs can influence the type of experiment they conduct and how they update their beliefs based on the results. * Scott notes that the experiment entrepreneurs conduct can be influenced by their prior beliefs, with different priors leading to different types of experiments and updates, and highlights the importance of considering these factors in decision-making. ### **Entrepreneurship, Bayesian learning, and causal logic.** * Speaker reflects on Bayesian thinking and its role in entrepreneurship, questioning whether "Bayesian" is the right term to describe the core of what entrepreneurs do. * Speaker 3 discusses the challenge of describing what entrepreneurs do or should do, and the importance of using the right language to describe their actions. * Speaker 3 emphasizes the normative nature of the agenda, which is to help entrepreneurs do their work better, rather than simply explaining what they do. * Psychologists: Causal logic helps learn from sparse data in belief formation, learning, and theory development. ### **Causal logic and entrepreneurship.** * Judea Pearl argues that causal logic is critical for entrepreneurial success, enabling learning in small amounts and solving subproblems to unlock magical outcomes. * Speaker 3 discusses the importance of testing assumptions in causal logic, highlighting the difference between confidence in assumptions and confidence in the underlying theory. * Speaker 3 emphasizes the need to mix belief revision literature with their own task of revising within a theory in a methodical way. ### **Entrepreneurship, learning, and theory.** * Elon Musk's team experiments with different theories to achieve space travel at a low cost. * Speaker 3 discusses the challenges of building causal models of future states, citing Heisenberg effects and the difficulty of learning from data. * Speaker 3 argues that theories are important for recognizing surprises and updating decisions in entrepreneurship. ### **Bayesian entrepreneurship research program.** * Intrapreneurs use Bayesian approach to make decisions based on prior knowledge and evidence. * Researchers compared Bayesian and non-Bayesian decision-making approaches among intrapreneurs, examining their termination decisions, PTO teams, and revenues. * Scientific entrepreneurs are more efficient in search process but also experience heightened methodic doubts. * Speaker 4 discusses their research on efficient search and intrapreneurship, sharing insights from a forthcoming paper and inviting others to explore the available data. ### **Entrepreneurship, innovation, and causal links.** * The speaker represents a user perspective and finds the white paper helpful, suggesting experiments to inform and persuade investors. * Speaker 2 discusses the importance of testing theories in entrepreneurship, particularly in the context of technical advances. * Speaker 2 discusses the potential of effectuation theory to help entrepreneurs think about differences across science and technology, and how it can be applied in a framework that economists can use. * Speaker 2 expresses interest in exploring the idea of effectuation further and encourages others to share their thoughts and comments. ### **Entrepreneurship, Bayesian logic, and experimentation.** * Speaker 5 criticizes the white paper for being overly optimistic and inconsistent in its use of terms like "creation" and "discovery." * Speakers discuss Bayesian and Knightian uncertainty in decision-making, with a focus on finding common ground between the two perspectives. * Entrepreneurs' optimism and experimentation are linked, with more optimistic individuals more likely to engage in experimentation and persuasion. * Speakers discuss the importance of probabilistic thinking in decision-making, with one speaker advocating for a Bayesian approach and another expressing frustration with the limitations of probabilistic thinking. * Speaker 8 suggests selection process for entrepreneurs may be unpacked, while Speaker 1 questions whether optimist entrepreneurs will notice potential resources. ### **Entrepreneurship, effectuation, and boundary conditions.** * Speaker 2 emphasizes the importance of considering non-Bayesian and Bayesian perspectives in data analysis, highlighting the distinction between updating probabilities for events with a zero probability and not caring about disability. * Speaker 3 discusses the role of cognitive and non-cognitive factors in expanding awareness and creating future states, with a focus on normative agendas. * Discussion on the limitations of current entrepreneurship theories and the need for new approaches to account for unforeseen phenomena. ### **Entrepreneurship education and research methods.** * Speakers discuss similarities and differences in their approaches to entrepreneurship education. ### **Entrepreneurship, opportunity, and choice.** * Entrepreneurs make choices, balancing clear opportunities with broad impact. * Speaker 1 argues that people make choices grounded in their assessment of what might be fabricated for them, and provides examples to illustrate this point. # **1.5: Priors in Entrepreneurial Decision-Making** Entrepreneurs must consider prior information, fair process distribution, and economic insights to make informed decisions. Speaker 1 emphasized the importance of recognizing and accounting for priors' heterogeneity, while Speaker 2 highlighted the challenges of decision-making in uncertain environments. Speaker 3 discussed the role of mentors and external advice in entrepreneurial decision-making, and Unknown Speaker emphasized the need for a test-and-learn approach. The conversation also touched on the importance of developing and testing entrepreneurial strategies, and the value of using a common language and shared understanding among entrepreneurs. ## **Transcript** <span style="text-decoration:underline;">https://otter.ai/u/lxQhCK56CJJYoaq0vjPp8Iir7zo?utm_source=copy_url</span> ## **Suggested Items** * [ ] Distribute copies of a share file in a fair process * [ ] Find speaker's office and get copies of the share file from there * [ ] Send out the Zappos case that was presented earlier * [ ] Elaborate on optimistic vs thoughtful thinkers in a later session * [ ] Have a session tomorrow on the choice of experiments when needing to persuade * [ ] Provide references for the decision theory concepts discussed * [ ] Break for coffee * [ ] Continue the conversation when we return at 3pm ## **Outline** ### **Prior knowledge and its impact on decision-making.** * Speaker 5 seeks help from doctoral students to distribute copies of an experiment on intrapreneurial priors. * George Bell is hesitant to purchase Google due to price and conditions, while Larry Page considers selling for $150,000 plus 1% of excise. * Speaker 1 discusses Bayesian approach to valuation, highlighting the role of priors and likelihood functions in shaping decision-making. ### **Priors and optimism in intrapreneurship.** * Speaker 1 highlights the importance of acknowledging priors heterogeneity in intrapreneurship research. * Speaker emphasizes individual differences, agency, and optimism in entrepreneurship. * Speaker discusses the concept of "logically pragmatically grounded optimism" as a key factor in decision-making. ### **Bayesian analysis and decision-making.** * Speaker 1 ponders how different eras form and choose relatively standard priors without relying on external shocks or events. * Speaker discusses Bayesian approach to decision-making, focusing on concentration of distribution probabilities. * Entrepreneur imagines new market emerging from technology development, seeks to increase probability of occurrence. ### **Bayesian reasoning and decision-making.** * The speaker discusses the importance of probability theory in understanding the relationship between technology and market development. * The speaker explains how probability theory can help identify the expected value of respect in new markets, weighted by the probability of state of interest. * Decision scientists address selective attention and blind spots in reasoning about uncertain future states. * Speaker 1 discusses Bayesianism, including reverse Bayesianism and prior on priors, in the context of decision-making and entrepreneurship. ### **Bayesian analysis and prior information.** * Speaker 1 explains how Bayesianism can accommodate for unknown states, and the implications for decision-making. * Speaker 1 discusses the importance of priors in Bayesian analysis, including their role in updating beliefs and selecting experiments. * Speaker discusses theories, parsimony, causal reasoning, and disentangling entities in AI research. ### **Entrepreneurial assessment and optimism.** * Speaker 2 presents on optimism, overconfidence, and tolerance for ambiguity in entrepreneurship. * Speaker 2 outlines a framework for evaluating entrepreneurial decisions, emphasizing the importance of probability judgments and risk aversion. * Speaker 2 shifts focus from probabilities to possibilities, highlighting attitudes and beliefs as motivational factors that can deviate from true values. ### **Decision-making and uncertainty in entrepreneurship.** * Expertise can lead to overconfidence, but being self-delusional about skills may not be beneficial for decision-making. * Decision theorists have found that tolerance of ambiguity is a promising construct for entrepreneurship, with two components: ambiguity aversion and a taste for ambiguity. * Speaker 2 discusses the concept of "ambiguity insensitivity" in decision-making, which refers to treating different possibilities as the same despite their differences. ### **Decision-making and mentorship in entrepreneurship.** * Decision maker uses limited recall of past experiences to make inferences about future outcomes. * Mentors share valuable experience and insights, helping intrapreneurs navigate entrepreneurship. * Speaker 2 discusses the importance of understanding why certain groups seek advice and how it relates to their level of confidence and trust in their ideas. ### **The role of advice in entrepreneurship decision-making.** * Speaker 3 shares their process of incorporating Bayesian ideas into their work in entrepreneurship, including defining terms and framing a white paper around Bayes theorem. * Speaker 3 discusses the role of advice in entrepreneurship, highlighting how advice can come from various sources and influence priors, decision-making, and levels. * Speaker 3 also notes that mentors and peers provide different types of knowledge, with mentors offering strategic advice and peers providing practical help with implementation. ### **Overcoming premature satisficing in entrepreneurship.** * Advice from mentors in intensive bursts helps entrepreneurs overcome premature satisficing. * Entrepreneurs need negative feedback, transparency, and structured learning to update their business ideas and succeed. ### **How entrepreneurs use advice to broaden their theories.** * Entrepreneurs must consider multiple theories to succeed, even if it means broadening their initial ideas. * Entrepreneurs often push back against advice, but can be convinced to use it by engaging in a dialogue and testing it. ### **Entrepreneurial strategy testing and mentor feedback.** * Speaker 3 discusses strategies for entrepreneurs, including testing ideas and gaining feedback through pitching. * Speaker 3 discusses using test pitches to cheaply test strategy ideas and gather mentor feedback. * Speaker 3 emphasizes the importance of collecting information before testing and using advice to expand options before actually testing anything. * Speaker 3 emphasizes the importance of qualitative research in identifying anomalies and refining theories. * Speaker 3 quotes "assumptions, blind hypotheses" from a book and reflects on the many choices entrepreneurs make without realizing it. ### **Economics and decision-making strategies.** * Speaker 4 emphasizes the importance of framing and supply/demand analysis in understanding Viking economics. * Speaker 4 discusses the importance of using language to analyze and understand issues in a particular field. # **1.75 Entrepreneurial Experimentation and Learning** Entrepreneurs and investors shared insights on navigating complexities in entrepreneurship. Speaker 1 emphasized rejecting a single optimal outcome and updating theories based on feedback. Speaker 2 highlighted the need for optimism and a scientific approach to problem-solving. Speaker 3 discussed the value of Bayesian thinking, while Speaker 4 shared anecdotes about Google's founders. Speaker 6 emphasized distinguishing between worrying about venture viability and the optimal path. Speaker 7 discussed the attention problem in entrepreneurship and the need for informed decision-making. Speakers also discussed trust and expertise in decision-making, experimentation and learning in corporate strategy, and the challenges of managing a business with a theory-driven approach. Unknown Speaker added notes on related touchpoints and the importance of understanding the theory of what one is trying to do. ## **Transcript** <span style="text-decoration:underline;">https://otter.ai/u/zOKiw9Jv_vdnsu368EyIM4TKw54?utm_source=copy_url</span> ## **Suggested Items** * [ ] Send the working paper to participants ahead of the session. * [ ] Share more details on the neuroscience research showing how brains are prediction machines if anyone is interested. * [ ] Think about heterogeneity of designers similar to buyers when evaluating models and measures of success. * [ ] Reflect on how negative feedback may indicate value for entrepreneurs rather than lack of merit. * [ ] Consider how an entrepreneur's network for advice impacts their access to quality guidance. * [ ] Examine the differences between B2B and B2C entrepreneurship, especially around experimentation. * [ ] Develop more common language and conceptual tools from the research while retaining nuanced insights. * [ ] Spend more time developing strong theories as a foundation before taking action as managers. * [ ] Consider the lifecycle of entrepreneurs and how their theory formation may vary across domains and over time. * [ ] Focus next year's discussion on capital structure, governance consistency, and sequencing of theories for entrepreneurs. * [ ] Examine the strategic role of signaling and eliminating noise without fully disclosing theories upfront. * [ ] Analyze how entrepreneurs can assess which theories are feasible for them to understand and implement. ## **Outline** ### **Brain function, learning, and design heterogeneity.** * Speaker discusses how the brain operates as prediction machines, using sensory data to update models, and how this relates to machine learning and AI. * Speakers reject single-minded approach to success, acknowledge socially constructed reality. ### **Entrepreneurship, optimism, and learning.** * Entrepreneurs want positive feedback, but negative advice can be valuable for learning and reducing variance. * Entrepreneurs need a good theory of how to access resources to be successful. ### **Entrepreneurship, technology, and business growth.** * Speaker 3 finds insightful way forward after realizing past victims' payments are important. * Google founders received critical feedback on their website, leading to a trillion-dollar business model. ### **Intrapreneurial experimentation and learning.** * Speaker 4 emphasizes the importance of trust and expertise in decision-making. * Scott discusses upcoming meeting with Andrea and Fiona, hoping for regular follow-ups. * Speaker 1 introduces the topic of intrapreneurial experimentation and learning, and distinguishes between testing hypotheses within a state space and testing theories across state space. * Speaker 1 provides two case examples from their company, including the idea of a motto and bike sharing, to illustrate the distinction between these categories of language. ### **Testing hypotheses for scooter sharing service.** * Speaker 1 discusses the problem of demand for a service, considering the potential market for young people and the importance of parking locations. * Speaker 1 hypothesizes about the probability of high demand given different parking locations, highlighting the ambiguity of the situation. * Speaker 1 proposes an experiment to test the hypothesis that scooter sharing services should be targeted towards college students or young professionals. * Speaker 2 suggests that there are two different ways of thinking about the problem, with different expenses and probabilities of high demand. ### **Experimentation and decision-making in entrepreneurship.** * The speaker discusses the difference between testing a theory and fine-tuning in a safe space, and how to approach comparing alternative theories. * Speaker 1 explains that the probability of success of the Airbnb of things is very small, so they commit to publicly if it is very, very high. * Speaker 1 also discusses the shape of the curve for the expected value of the experiment on the alternative theory, and how it turns out that the probability of success of the MVP of things is very small. * Speaker 1 assumes that the probability of being accused is equal to the prior, and uses this assumption to inform their decision-making in the experiment. ### **Intrapreneurship and decision-making processes.** * Speaker 1 discusses Bayesian vs decision scientist models for unknown events, with implications for optimism and hypothesis testing. * Intrapreneurs behave like events, anticipating problems but lacking actionable solutions. * Experimenting outside one's own space can lead to successful outcomes, but resources and privacy concerns may be an issue. ### **Entrepreneurship, venture viability, and criticality.** * Speaker 6 shares insights on entrepreneurship, AI, and the importance of founder personality. * Speaker 6 emphasizes the importance of testing the whole system, rather than just one piece, to ensure venture viability. * Speaker 6 highlights the distinction between worrying about venture viability versus the optimal path, with examples from MIT spin-offs and DOCSIS. * Speaker 6 aims to create a weather company by launching satellites, believing they can do better than existing weather companies. * Speaker 6 discusses the importance of learning spillovers in technology areas, and how it can inform decision-making about investment and viability. * Speaker 6 also mentions the potential for a "quadrum kind of stuff" in the space of working on third or fourth quarter issues, and how it may not be taken seriously by others in the room. ### **Entrepreneurial experimentation and learning.** * Entrepreneurs use experiments to test and update their priors and demand for experimentation increases with uncertainty. * Speaker 7 discusses the importance of generating alternative theories and hypotheses in the entrepreneurial process. * Speaker 7 proposes a framework for entrepreneurship that includes three steps: idea generation, testing and learning, and making an explosive choice. ### **Using experiments to update theories in data-driven decision-making.** * Entrepreneurs may not collect enough data despite lower costs, leading to updates to prior beliefs. * Speaker 7 highlights the issue of noise and bias in field experiments, which can lead to uninformative signal and hinder learning. * Even when experiments provide informative signal, they may not lead to learning due to three problems identified by Speaker 7. ### **Entrepreneurs' tendency to ignore important data.** * The speaker discusses the "attention problem" in entrepreneurship, where entrepreneurs may not be aware of or paying attention to the data they need to test their theories. * The speaker's experiment on 3000 firms in the personal care industry found that 60% of them were unable to guess competitor prices, highlighting the potential for learning failures in key decisions. * Speaker 7 observes customers and investors, but ignores data on seaweed farming despite variation in pod size. ### **Data-driven decision making and the potential for incorrect conclusions.** * Speaker 7 discusses the "Trump data problem," where data collected may not be informative for predictive purposes, even with more data. * The speaker uses the analogy of an entrepreneurial startup to illustrate how click-through rates may not be the right data for measuring business potential, and how targeting the wrong sample can lead to poor results. * Speaker 7 discusses how entrepreneurs may be misled by their own biases and the persuasive power of wrong models, leading to updates away from the truth. * The likelihood of observing data with a wrong model may be higher than under a truthful model, especially when data is random and the wrong model is more compelling. ### **AI's impact on entrepreneurship and decision-making.** * Speaker 7 discusses how AI can generate theory and machine learning can detect patterns, potentially making it easier to evaluate and select among theories. * The second step of learning from experimentation may become more valuable in the future as both theory generation and choice become easier with AI. * Speaker 7 discusses the importance of linking theory and data in entrepreneurship, highlighting the need for entrepreneurs to understand both to make informed decisions. * Speaker 7 raises questions about the challenges of updating towards the truth in entrepreneurship, including the difficulty of making the learning process more efficient. ### **Experimentation and moderation in a conference session.** * Speaker 3 mentions Rome's studies and their impact on experimentation. ### **Experimentation's impact on startup valuation.** * Speaker 4 highlights the importance of experimentation in evaluating startup value, demonstrating how it can change investment decisions and strategies. ### **Entrepreneurship, experimentation, and learning.** * Speaker 9 discusses the importance of experimentation and data fitting in entrepreneurship, highlighting the value of learning through experimentation and the potential for data fitting to reveal insights (1:19:46-1:22:14) * Speaker 1 emphasizes the crucial role of assumptions in data fitting and testing theories, and how flexible functional assumptions can help in testing a theory (1:22:05-1:22:14) * Entrepreneurs experiment and seek to fine-tune their approaches, differing from the idealized image of entrepreneurs in the work microphone. * Participants discuss the importance of experimentation in business, with some arguing that it can lead to valuable learning experiences, while others suggest that it may overshadow other forms of learning. ### **Using sparse data for entrepreneurship.** * Speaker 5 highlights the challenge of inventing space for entrepreneurs, who may not be using it correctly due to complexity and difficulty. ### **Entrepreneurship and career journey.** * Andrea On is a financial services executive with experience at Salomon Brothers and founding i on, a company that provides critical workflow and pricing data. * Elon Musk describes his journey from a strong career to becoming an entrepreneur, sharing his vision and beliefs that guided his choices and strategies. * Speaker 5 learned to delegate tasks and trust the market, realizing that IT departments were not focused on their needs. ### **Decision-making strategies in finance.** * Speaker 5 discusses their journey from managing money to building a technology company, highlighting their desire to learn entrepreneurship and reduce opportunity costs. * Speaker 5 discusses building a culture of effective decision making with data-driven tools for financial institutions. * Speaker 5 argues that learning without theory is possible, but requires a systematic research approach to design experiments and understand relationships. ### **Entrepreneurship, business strategies, and investment.** * Entrepreneurs in b2b typically have a sense of what the industry needs and want to scale their business, while b2c is a discovery of a market that is a match between idea and client willingness to pay. * Speaker 5 emphasizes the importance of a unique theory of value creation in private equity vs. venture capital, with implications for seller and buyer needs. ### **Organizational design, mergers, and AI investments.** * Speaker 5 highlights the importance of organizational design and human capital matching for entrepreneurs. * Speaker 5 discusses merging with NASDAQ, reputation, and growth. * Speaker 5 developed a theory that banking was where trading used to be, and they bought 3 companies to prove their theory, including a pathology company to improve credit decisions. * Speaker 5 underinvested in the system and kept buying companies to stay efficient, despite having over 10,000 employees in their company. ### **Entrepreneurship, investing, and business strategies.** * Organization structure and information quality are key to decision-making, says Speaker 5. * Investors prioritize entrepreneurs' ability to execute and maintain governance over time, rather than just relying on their vision or story. * Speaker 5 discusses the importance of selecting the right theory for a project, considering both feasibility and consistency with other theories. * Sequencing theory is mentioned as a key concept, with the need to balance micro and macro perspectives, and to strategically choose when to disclose complex theories. **1st Annual Workshop @MIT Sloan** # **Bayesian Entrepreneurship Day 2** Michael and Speaker 2 discussed the impact of Ed Roberts' passing and his contributions to MIT's innovation and entrepreneurship ecosystem. Speaker 3 shared their personal journey in entrepreneurship, highlighting the challenges they faced in the field and their experiences with parallel ownership in the pharmaceutical industry. Speaker 3 emphasized the importance of understanding the innovation landscape and identifying the right areas for investment, as well as creating a reward system that incentivizes entrepreneurs to take risks. Speaker 3 presented an alternative approach to science and innovation, emphasizing the need for discontinuous and detached hypotheses that challenge the current reality. Speaker 3 discussed the challenges of entrepreneurship in the face of uncertainty, emphasizing the importance of imagination, creativity, and learning in overcoming these challenges. ## **Transcript** <span style="text-decoration:underline;">https://otter.ai/u/_2AvNFd1pSQzCYNmQ4qdisC36rM?utm_source=copy_url</span> ## **Suggested Items** * [ ] Send an email to Noubar Afeyan with any disagreements or thoughts on his views expressed in the meeting * [ ] Consider if Flagship Pioneering's parallel company building approach could work in other scientific fields like chemistry * [ ] Explore how artificial intelligence could help understand complex biological systems * [ ] Examine the reward systems and opportunity costs for entrepreneurs to motivate participation in parallel platforms * [ ] Analyze how the parallel approach impacts resilience in down economies when individual companies fail * [ ] Create order and process to navigate uncertainty systematically when innovating ## **Outline** ### **The loss of a valued colleague and MIT alumnus Ed Roberts.** * Ed Roberts, a valued colleague and MIT alumnus, passed away recently, inspiring tributes and remembrances from his colleagues. ### **Entrepreneurship and company creation at MIT.** * Robert Metcalfe, MIT professor and entrepreneurship pioneer, passes away. * Flagship ventures and Fractal's work in company creation and building in Cambridge, MA. ### **Entrepreneurship and innovation in the tech industry.** * John started a company after graduating, then realized a given company was limiting for innovation after going public in 1992. * Engineer turned entrepreneur struggles with serial entrepreneurship and investor expectations. ### **Parallel ownership and innovation in the biotech industry.** * Speaker 3 explores parallel ownership, questioning why it's inherently serial and experimenting with different patterns in the mid-90s. * Speaker 3 starts a firm called Neupogen for new company generation Intel, leveraging institutional learning and professional systematic approach with other companies in parallel. * Dr. Arnold said that the institution has created 48 companies and is involved in the development of over 10,000 drugs. * Dr. Arnold acknowledged that the institution's work is not without challenges, with the feeling that it could all fall apart at any moment. ### **Entrepreneurship, innovation, and the limits of expertise.** * Speaker 3 believes entrepreneurs and innovators are meant to defy experts to become experts themselves, but the curse is being viewed as an expert. * Speaker 3 recognizes the limitations of science, including the uncertainty of data reliability and the conservative nature of institutional knowledge. ### **Innovation and funding in science and technology.** * Speaker 3 argues that the current scientific funding system inhibits innovation by prioritizing incremental progress over bold, disruptive ideas. * Speaker 3 describes a strategy for generating R&D ideas by drawing a circle around known concepts and exploring new ideas outside of it. * Venture capital flows into startups in similar spaces because experts provide similar answers, increasing the likelihood of success. * Speaker 3 discusses the importance of being in a "reasonable zone" for innovation, suggesting that it's crucial to have conviction and confidence in one's ideas to persistently work towards success. * Speaker 3 also shares their personal experience as an immigrant, feeling like they were at a disadvantage despite being at MIT, highlighting the importance of having a mindset that allows for growth and innovation. ### **Leaps of faith in innovation and entrepreneurship.** * Speaker 3 discusses the importance of imagination and leaps of faith in scientific discovery, arguing that these qualities are necessary for making new and innovative discoveries. * Speaker 3 believes that science fiction can serve as a societally permissible outlet for people to express their beliefs and ideas, even if they may not be feasible or possible in the present time. * Speaker 3 describes a process of experimentation and iteration, relying on a team of 25 people who conduct 100 explorations a year to develop new ideas. * Speaker 2 asks questions about the reliance on the team and the people involved in the process, and the potential for success despite the risks and uncertainties involved. ### **Innovation and creativity in scientific research.** * Senior practitioners at a flagship AI organization have grown up with the company and gained confidence in their abilities over 15 years. * Speaker 3: Seeking world-class scientists to work on unconventional projects, pushing them out of their comfort zones. ### **Risk and uncertainty in investment.** * Entrepreneurs take calculated risks to create value, but most people are scared to do so. * Speaker 2 discusses uncertainty in biology and AI, arguing that nature is inherently uncertain and that companies work to mitigate this uncertainty through scientific research. * Speaker 2 references Andrew Lowe's book on the topic of uncertainty and its relationship to unknown provinces, outcomes, and risk. * Speaker 3 argues that uncertainty is not high risk, but rather a lack of knowledge or understanding, and that investors should not assume a high discount rate for uncertain investments. * Speaker 3 criticizes the way investment bankers approach uncertainty, trying to assess and mitigate risk instead of acknowledging and planning for it, and suggests a more flexible and adaptive approach to dealing with uncertainty. ### **Risk and uncertainty in entrepreneurship.** * Speaker 3 argues that uncertainty is not an absolute thing, but rather a distribution of valuables with different sizes and probabilities of success. * Speaker 3 believes that only small valuables exist closer to the present due to competition, while big valuables are often at the extreme end of uncertainty. * Speaker 3 highlights the importance of taking calculated risks in drug development, citing the example of mRNA vaccine development for COVID-19. * The speaker notes that while there have been 48 successful mRNA vaccine developments, the process of scaling up production to make a billion doses is a significant risk. ### **Engineering microorganisms for carbon-neutral fuel production.** * Speaker 3 discusses using bacteria to create renewable fuels, mentioning their company's focus on engineering microorganisms for biofuel production. * Engineers create microorganisms to convert CO2 into diesel fuel. ### **Entrepreneurship, risk-taking, and innovation in the energy industry.** * Incumbent oil industry's infrastructure and government ties hindered transition to renewable energy. * Speaker 3 discusses the importance of parallelism in accelerating learning and deriving value, citing the example of creating multiple parallel companies to explore different ideas and reduce risk. * The speaker highlights the reward system in place, where entrepreneurs are incentivized to take on research projects and receive shares in the companies that arise from their work. ### **Entrepreneurship, risk-taking, and learning cycles.** * Speaker 3 discusses the challenges of scaling a business, including the importance of prioritizing resilience and diversifying investments. * Speaker 3 shares their belief in the value of venture capital, citing the importance of having a continuous continuum of early-stage companies with potential for growth. * Speaker 3 discusses the importance of exploring uncertainty in entrepreneurship, suggesting that creating order in a disordered environment can be a protective measure. * Speaker 2 compares the speaker to Admiral Nelson, acknowledging their insightful comments on entrepreneurs and consumers. # **2.25 Entrepreneurial Persuasion** The conversation revolved around the importance of experimentation in entrepreneurship research. Speakers highlighted the challenges of applying academic theories to real-world scenarios and emphasized the need for rigorous design and data collection. They discussed the role of experimentation in creating a mechanism for founders to adjudicate disputes with investors, streamlining the process of collecting data, and validating experiment design. The speakers also touched on the trade-off between experiment cost and valuation, and proposed a university-investor partnership to generate benchmarks. Overall, the conversation emphasized the crucial role of experimentation in gaining insightful results and making informed decisions in entrepreneurship research. ## **Transcript** <span style="text-decoration:underline;">https://otter.ai/u/6V-stK4Ke2zFfUB-CxBVEzJ3yXM?utm_source=copy_url</span> ## **Suggested Items** * [ ] Create a paper synthesizing ideas on entrepreneurial persuasion * [ ] Develop computational simulations to model entrepreneurial information requirements over time * [ ] Make Bayesian entrepreneurship concepts clear for those unfamiliar with advanced theory * [ ] Consider both shared belief teams and attracting those who disagree for progress * [ ] Empirically test model implications like experiment types and diversity of beliefs * [ ] Advise entrepreneurs to break financing into stages with abandonment options * [ ] Develop benchmarks to improve learning from early experiments * [ ] Have universities validate experiment design through relational contracts with investors * [ ] Be careful with word choice and terminology like Bayesian persuasion vs persuasion * [ ] Make Bayesian entrepreneurship appealing beyond advanced theory * [ ] Consider naive versus sophisticated entrepreneurs and investors * [ ] Examine costs and information quality of high vs low bar experiments * [ ] Incorporate organizational structure into high/low bar experiments * [ ] Develop statistical methods for firms to learn from large scale experiments ## **Outline** ### **The role of theory in business and economics.** * Expert discusses the challenges of applying theoretical models to real-world problems in economics. * Speaker 1 discusses the concept of a "moat" in business, highlighting the difference between competing on execution versus investing in intellectual property protection. * Speaker 1 references a white paper on the topic, which emerged from a process of questioning and discussion with others in the field. ### **Entrepreneurship and decision-making.** * Scott's wife Kathy shared how he had a different perspective on scientific work, valuing empathy and doing the right thing, which led to a friendship despite their differences. * Kathy's conversation with Scott highlighted the importance of considering customer feedback in a more systematic and unbiased way, using randomization to get to the truth. * Speaker 1 discusses the importance of modeling in entrepreneurship, using the example of oil prices to illustrate how probabilities can be assigned to uncertain events. * Speaker 1 highlights the challenge of convincing oneself to continue pursuing an entrepreneurial venture, despite potential failure, and discusses the role of beliefs and probability in this process. ### **Decision-making in economics with optimal information.** * Speaker 1 discusses the "workhorse model" for decision-making under uncertainty, which involves balancing the costs of exploration and the potential payoffs of exploitation. * The speaker highlights the importance of considering the entrepreneur's perspective and the potential biases that can arise when using econometric techniques to analyze data. * Speaker 1 explains the concept of "choosable" in economics, which refers to the ability to make decisions based on available data. * Speaker 1 discusses the trade-offs between minimizing false positives and false negatives in experimentation, and how these choices are influenced by prior beliefs and product. ### **Entrepreneurial experiments and their success measures.** * The speaker emphasizes the importance of conducting low BAR experiments in entrepreneurship, which are easy to pass and provide valuable insights into what works and what doesn't. * The speaker notes that people often understand how to interpret low BAR experiments differently than high BAR experiments, with low BAR experiments providing affirmation and high BAR experiments providing the truth. * Speaker 1 discusses the importance of experimentation in product development, highlighting the value of minimal viable products (MVPs) in reducing false positives and startups. * Speaker 1 also mentions the concept of "high BAR" experiments, which require minimal effort and provide valuable insights into customer preferences. ### **Optimizing experiments for entrepreneurs.** * Entrepreneurs prioritize low-bar experiments over high-bar experiments, despite potential biases. * Speaker discusses the use of noise in decision-making, with a focus on the Bayesian approach. ### **Entrepreneurial optimism and experimentation.** * Entrepreneurs choose high bar experiments to convince investors, as they are more optimistic and want to maximize upside potential. * Minimum Viable experiments are valuable for businesses and investors, as they provide convincing evidence and demonstrate potential for growth. * Entrepreneurs' optimism influences likelihood of starting a venture, with lower probability of success for those who need to persuade others. * Pioneer of experiments on unbiased data to gain insights, but challenges remain in interpreting results and understanding data-generating process. ### **Multistage financing and real options in entrepreneurial finance.** * Speaker 2 discusses multistage financing as a fundamental application of Bayesian entrepreneurship ideas, highlighting the importance of tying milestones to value inflection. * Speaker 2 explains how to calculate the post-money valuation of a company at the B round, using the example of a biotech investment. * Scott, a skeptical investor, is willing to pay $8.9 million to fund an experiment that could improve his expected value, despite only having a 10% chance of success. ### **Entrepreneurship, investment, and experimentation.** * Speaker 2 discusses the importance of experimentation in entrepreneurship, emphasizing the need for informative experiments that provide valuable insights. * The speaker highlights the benefits of high-bar experiments for both entrepreneurs and investors, as they provide more information and increase the value of the venture. * Entrepreneur raises money at higher valuation, potentially increasing chances of success but also increasing risk of failure. * Guidance on how much money to raise based on the amount of time needed to run experiments and generate information for next round of financing. * Airbnb's early high valuation led to significant dilution in later rounds. ### **Low learning efficacy in early-stage tech ventures.** * Speaker 2 argues that the growth of venture capital financing over the past decade has been uneven, with most of the increase going to IT, consumer products, and business services, while sectors like healthcare and hardware have seen little growth. * Speaker 2 suggests that this dearth of capital in certain sectors may be due to a lack of investor interest or a lack of resources available to build these companies, particularly in universities. * Speaker 2 discusses low learning efficacy in the materials sector, citing examples of long-term stability of electric catalysts and drug development. * The rate of failure from preclinical to clinical trials is high, with less than 10% of drugs making it through Phase 123 and launch, due to a lack of predictive power in lab experiments. ### **Experimentation and entrepreneurship in venture capital.** * Speaker 2 discusses the lack of well-developed benchmarks for customer discovery and experiment design in the venture capital industry, which can lead to moral hazard and friction in the investment process. * Speaker 2 references a paper co-authored with Simon and others on the idea of moral hazard in experiment design, which explores the degree to which entrepreneurs might want to run less informative experiments to avoid killing the killer experiment. * Professor suggests co-designing informative experiments with investors to validate entrepreneurial ideas. * Speaker 2 highlights the need for clarity on the definition of "expert" in the context of experimentation. ### **AI research and data collection.** * Speaker 2 questions the market size and feasibility of a medical device company's approach, citing lack of common language and ineffective distribution model. * Entrepreneurs and experts discuss the importance of considering outliers in experimentation and data collection. * Efficient data collection process streamlined, with 7000 data points collected in 8 years. ### **Startup financing and experimentation.** * Speaker 5 highlights the challenge of making progress in a field despite feedback and suggests experiments to enhance legitimacy and social acceptance. * Lean startup is a language for founders to adjudicate disputes with investors, not just a method for creating a business. * Entrepreneurs and investors advised to break up financing and R&D into stages for flexibility and abandonment. * Entrepreneurs advised to get outside investment due to insider incentives and ability to continue investing in rounds. ### **Experimentation costs and investor naivety.** * Experimenters debate whether early-stage investors prioritize homophily or high-fidelity investments. * Experiment cost-benefit analysis is crucial for venture success. * Entrepreneurs and investors discuss optimal experiments, naivety, and sophistication, with questions on cost and informed decision-making. * Speakers discuss econometric theory, modeling, and experimental design, with a focus on trade-offs and constrained set of experiences. ### **Experimentation and statistical methods in tech companies.** * Organizations with hierarchical structures may struggle with decision-making and risk management. * Experimentation in tech companies requires commitment to statistical methods to learn from data. ### **Experiment design and Bayesian learning.** * Speaker 9 highlights importance of experimental design, pedagogical sessions, and Bayesian learning. # **2.5 Entrepreneurial Pedagogy and Practice** ## **Transcript** <span style="text-decoration:underline;">https://otter.ai/u/A5_O-sC7uQGE4Wso73WmbXpaT7Y?utm_source=copy_url</span>