dimo q1. how to choose between three toolboxs (exploratory, move testing) and four hats (artist, scientist, engineer, designer); i if have computational tool to do this called simulation based calibration, would you be interested? teppo and thomas q2. when teppo said multiplicative causal logic, do you mean the observation is convolution of random action variables of different hierarchy? can we ellicit experiment design rules from this e.g. makeing each hierarchy independet? thomas and arnaldo q3. there are computational statitical algorithim of reorganizing of causal representing called collapsing - might this be what arnaldo meeant by blackwellization of experiments? how to make productize this thery in a tool that entrepreneurs can use? scholar, entrep, when the list doesn't form, error in observational judgement world of theory fit (theory is a record of the fit) theory (observational judgement) performance () impose a shape to world form should have a conception (sense of concept; not testing of theory, but making of artifact) emotional intelligence pragmatism would deny essense if desire changes (if value changes), does the agency changes? reflective practionial (validity is propositional consequence of actiong upon it) works for who? process #todo q to scott 1. idea for market (send a mail to scott, attaching his mail, summarizing [longer feedback cycles in deep tech](https://claude.ai/chat/8f04cb43-08f0-4728-8f88-0af6f0bb9bb8), ) q to [hart](https://www.tuck.dartmouth.edu/faculty/faculty-directory/hart-posen) 1. and statement -> lower bar 2. 3. how did you design sequence of papers (your trilogy with hart) weaker prior, my image of entrepreneurial is starting from strong prior and letting it washed out gradually by data this is different from using static weak prior from the start, but more evolving into weaker prior. andrew gelman's argument "Bayesain from defense to offense" might be relevant which is Instead of worrying about prior sensitivity, respect sensitivity and, when prior makes a difference, we want to use good prior information. time inconsistency, rober pollick (consumer theory) donut (mean, ends evolve (adapt evolutionary epistology) neuro andi interplay change the perception to match action, change action to change perception) 1. "what works" time frame (by tomorrow or next month or year) and "who" (entrepreneur, investor, ) when (idea's truth is its cash value) - agency 2. if desire change, (pivot the goal or mission) does agency change? - inductive bias (approaximation, stateistical, optimaization error) - simulation modeling computational I don't ontology in short my q are three: 1. what are your thoughts on this strong to weak evolving prior? 2. what are your image of scientists that are not entrepreneurial? 3. what are your image of scientists that are entrepreneurial? adapt too much from tech measuring, gathering evidence, experimenting | Experiment | Key Variables | Strategic Decisions | Learning Algorithm | |------------|---------------|---------------------|---------------------| | Roadster Supply Chain | Battery supplier (Xcellent), PEM supplier (Chroma), Shipping routes | Outsourcing vs Vertical integration | Bayesian | | Model 3 Production | Automation levels, Worker skills, Factory layout | Automation investment, Manufacturing process design | Evolutionary | | Roadster Demand | Pricing ($109,000), Performance specs (0-60 mph in <4 sec) | High-end market entry strategy | Bayesian | | Cash Flow Management | VC funding rounds, R&D costs, Production expenses | Funding strategy, Cost management | Behavioral | | Battery Pack Quality | Xcellent manufacturing process, Defect rates | Supplier training, Quality control processes | Evolutionary | This table now reflects specific details from the Tesla cases: 1. The Roadster supply chain experiment focuses on the actual suppliers mentioned (Xcellent and Chroma) and the shipping challenges faced. 2. The Model 3 production experiment addresses the automation issues Tesla faced in ramping up production. 3. The Roadster demand experiment uses the actual price and performance specs mentioned in the case. 4. The cash flow management experiment considers Tesla's VC funding rounds and R&D costs. 5. The battery pack quality experiment focuses on the challenges with Xcellent's manufacturing process. Each experiment is mapped to a learning algorithm based on how Tesla approached these challenges, as described in the cases.