🛝BE25_Ent_decision_making_m2

# Economies of Lifting your Carpet 🚨device of imaginary outcome - rejection sampling🚨 ## Computational Thinking for Entrepreneurial Decision-Making ## 🗄️ Table of Contents |🔐 Question|🧱Solution Module/Literature Brick/Data|🔑Key Message|📊Figure/Table| |---|---|---|---| |**Part 1: Foundations - What lies beneath our decision carpets? (0-20 min)**|||| |How does complexity challenge entrepreneurial decision-making and why do we need to "lift the carpet"?|• Philosophical foundations (Sontag, Bourdieu, Deleuze) • "Subjectivity swept under the carpet" (I.J. Good) • 46,656 varieties of Bayesian thinking|Style emerges when process (model-allocate-search-program) and product (decisions) achieve consistency; entrepreneurs need to make their implicit beliefs explicit|📊 Process diagram showing five updates from data-to-action to probabilistic program 📈 Visualization of style emergence| |What computational thinking principles form the foundation for entrepreneurial style?|• Three key entrepreneurial principles from BE25 framework • Definitions of model-allocate-search-program • Relationship between expressive means and expressed content|Computational thinking provides systematic tools to structure decisions, imagine possibilities, and continuously refine strategies regardless of resource constraints|🗄️ Entrepreneurial Principles and Computational Thinking table showing alignment between principles and computational concepts| --- ## 🗄️ Table of Contents (continued) |🔐 Question|🧱Solution Module/Literature Brick/Data|🔑Key Message|📊Figure/Table| |---|---|---|---| |**Part 2: Imagination - Generating Possibilities (20-40 min)**|||| |How can entrepreneurs better utilize imagination in decision-making?|• Vul (2014): Few samples optimality • "How We Know What Not To Think" • I.J. Good's "device of imaginary results" • Tenenbaum: Generative modeling|💭 The device of imaginary results allows entrepreneurs to work backward from desired outcomes; making decisions based on few samples can be globally optimal when considering time costs|📊 Vul's error rates and expected utility analysis figures (Fig 3-7) 📈 Dimension map of "Yes Simulation" vs "No Simulation" approaches| |What makes imagination different from traditional forecasting?|• Bayesian theorem in reverse (I.J. Good) • Child as hacker model (Tenenbaum) • Theory-driven experimentation (Camuffo)|💭 Using Bayes' theorem in reverse enables entrepreneurs to work backward from imagined outcomes to required conditions, unlike traditional forecasting that projects forward from current conditions|📊 Conceptual diagram of device of imaginary results with lime example 📈 Simulation dimension chart showing scholarly approaches| 🚨device of imaginary outcome - rejection sampling🚨 --- ## 🗄️ Table of Contents (continued) |🔐 Question|🧱Solution Module/Literature Brick/Data|🔑Key Message|📊Figure/Table| |---|---|---|---| |**Part 3: Probability - Developing Precision (40-60 min)**|||| |How should entrepreneurs develop probability precision over time?|• Gelman: Hierarchical Bayesian modeling • Tenenbaum: Learning from minimal data • Nested logit regression models (Ben-Akiva)|⚖️ Entrepreneurs should start with coarse probability estimates and iteratively refine them through strategic experimentation; precision develops through structured feedback loops|📊 Digital vs analog scale visualization of probability precision 📈 Classification table of developing vs fixed precision approaches| |What is the right balance between probability precision and resource constraints?|• "One and done" sampling (Vul) • "Holes in Bayesian Statistics" (Gelman) • Resource rational decision making (Bhui)|⚖️ Developing probability precision requires balancing the costs of thinking vs acting; explicit modeling of theorizing costs enables better resource allocation|📊 Robot localization example showing noise trade-offs 📈 Inside-out noise ratio diagram showing model/inference vs world/measure balance| --- ## 🗄️ Table of Contents (continued) |🔐 Question|🧱Solution Module/Literature Brick/Data|🔑Key Message|📊Figure/Table| |---|---|---|---| |**Part 4: Utility - Optimization Under Constraints (60-80 min)**|||| |How can entrepreneurs optimize decision-making under evolving utility structures?|• Ben-Akiva: Random utility & nested logit models • Bhui: Resource rationality framework • Stevenson: Pursuit of opportunity without regard to resources|💸 Quasi-utility approaches recognize theorizing costs; explicit modeling of resource constraints enables separation of learning and earning activities|📊 Combined Constraint Scores Table (0-10) showing scholarly approaches 📈 Classification table of fixed vs quasi-utility approaches| |What makes entrepreneurial decision-making different from standard explore vs exploit frameworks?|• Bayesian Decision Making in Groups (Jadbabaie) • Entrepreneurship as opportunity pursuit • Resource-Model-Action feedback loops|💸 Entrepreneurs face unique utility challenges: they must discover both the opportunity space and their own utility structure simultaneously while operating under resource constraints|📊 Two-stage model of resource allocation across model-allocate-search-program 📈 Contrast diagram of entrepreneurial vs standard decision processes| --- ## 🗄️ Table of Contents (continued) |🔐 Question|🧱Solution Module/Literature Brick/Data|🔑Key Message|📊Figure/Table| |---|---|---|---| |**Part 5: Integration & Implementation (80-100 min)**|||| |How can entrepreneurs construct their own decision style using the model-allocate-search-program framework?|• 8 Bayesian tribes from 46,656 varieties • Process of lifting your carpet (five stages) • Application to equity valuation example|The model-allocate-search-program framework enables entrepreneurs to develop a consistent style across different decision contexts; explicit computational thinking constructs more robust decision processes|📊 3D diagram showing imagination-probability-utility dimensions 📈 Applied example of decision framework for equity valuation startup| |What concrete next steps can entrepreneurs take to implement this approach?|• Examples of domain-specific applications • Specific tools for model-allocate-search-program • Implementation roadmap|Start by lifting your own carpet: explicitly model theorizing costs, develop precision iteratively, and create computational programs that connect your decision process to your products|📊 Implementation timeline with key milestones 📈 To-do list for practical application| --- ## 🗄️ Entrepreneurial Principles and Computational Thinking |Principle|Core Focus|Why Computational Thinking Helps|Key Computational Concepts| |---|---|---|---| |**1. Structure Your Thinking**|Clearly approach opportunities regardless of current resources|Enables systematic breakdown of complex environments with shifting opportunities and constraints|• Decomposition • Abstraction of process • Algorithmic thinking| |**2. Imagine Possibilities**|Creatively balance vision against resource constraints|Provides tools for representing complex situations and supporting effective decision-making|• Modeling physical and social worlds • Managing trade-offs and complexity| |**3. Continuously Refine**|Openly adapt strategies as resources and opportunities evolve|Emphasizes systematic improvement and clear communication of adjustments|• Iteration and recursion • Design and debugging • Communication of structured thought| --- ## Process of Lifting Your Carpet 0. **Baseline Data-to-Action** - 🧪 Utility & Precision of judgement & Device of imaginary results from 46656 varieties of Bayesians provides axes to classify literature 1. **Action-Based Model Inference** - 🚗 Random utility & Nested logit regression model from Discrete choice analysis (budget constraint) - 💻 Integrative modeling from Integrating explanation and prediction in computational social science 2. **Iterative Action-Model Refinement** - 🔄 Bayesian analysis cycle and workflow from Bayesian statistics and modeling --- ## Process of Lifting Your Carpet (continued) 3. **Strategic Resource Allocation for Model Selection** - 📝 Model selection through experiments from Theory-driven strategic management decisions - ⚡ Balancing expected rewards and informational complexity from Resource rational decision making 4. **Model-Based Resource Discovery** - 🔴 Entrepreneurship as Pursuit of opportunity without regard to resources currently controlled from Perspective on entrepreneurship - ⛳ Bayesian updating, Model evaluation, Bayes factor, Model updating from Holes in Bayesian statistics 5. **Probabilistic Programming for Resource Allocation** - 👶 Child as hacker - cognitive development as hierarchical Bayesian program induction from Bayesian models of conceptual development --- ## 🗄️ Combined Constraint Scores Table (0–10) |Constraint|Ben-Akiva|Gelman|Camuffo|Tenenbaum| |---|---|---|---|---| |💭 **Use of Imagination**|6|8|10|9| |⚖️ **Spatial Perception**|7|7|8|8| |⚖️⚖️ **Spatio-Temporal Perception**|3|8|9|9| |💸 **Utility Heterogeneity**|9|7|6|2| |💸💸 **Resource Heterogeneity**|5|5|5|4| --- ## Simulation Dimension |Yes Simulation|No Simulation| |---|---| |Uses theoretical simulations or mental models|Relies on empirical data without simulation| |- Andrew (2021) Bayesian Statistics|- Moshe (2024) Nested Logit| |- Sendhil (2021) Integrating Explanation|- Arnaldo (2024) Theory-driven Decisions| |- Andrew (2021) Holes in Bayesian Stats|- Rahul (2021) Resource Rational| |- Josh (2020) Conceptual Development|- Stevenson (1983) Entrepreneurship| --- ## Noise from Inside Out **model/inference** and **world/measure** - **action** depends on **inside** noise to **outside** noise ratio - if **inside** noise becomes higher, **measure** more - if **outside** noise becomes higher, **move** more ![Example of robot localization with noise trade-offs](https://i.imgur.com/example.jpg) --- ## Apply to Your Decision Making **Example of "probabilistic AI for equity valuation and allocation"** **Strategy:** - customer (founder VS lawyer) - technology (ADEV VS DSL) - organization (hire technology expert VS customer expert) - competition (collaborate VS compete with lawyers) **Operations:** - segmentation / evaluation - processification / automation - acculturation / evaluation - capitalization --- ## Next Steps 1. **Set language** - subjective → contextual - prior distribution = action-oriented encoding of belief 2. **Introduce missing pieces** - mental model - convergence in bayesian inference behavior - stochasticity ~ heterogeneity/varieties of Bayes - modeling workflow 3. **Apply to your decisions** - control vs flexibility - learning vs earning activities - modeling -> allocating -> searching -> programming --- ## Key Message Style emerges when **process** and **product** achieve **consistency** Start lifting your carpet today: 1. Make implicit beliefs explicit 2. Develop probability precision iteratively 3. Model theorizing costs explicitly 4. Create computational programs to connect your process to products [[segment]]