[[eval(josh, recovering rationality of venture's adaptation)]]
[[🧍♀️2🧍♀️🌏_A2AE_charlie-jb]]
The framework maps academic content through a Target (🎯) defining the section's goal, implemented via Algorithm Bricks (🧱) detailing methodological steps, materialized through Keys (🔑) in language/plots/tables, and connected by Bridging (🌉) elements that spark curiosity about the next target while leveraging current insights.
[[🗄️product2_EDT]]
| Section | | 🎯 Target | 🧱Algorithm Brick | 🔑Key in 🗣️language, 🎞️plots, 🗄️table | 🌉Bridging |
| ---------------------------------------------------------------------------------- | --- | ------------------- | ---------------------------------------------------------------------------------------------------------------------------------------------- | ------------------------------------------------------------------------------------------------------------------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ |
| knowhow of using column relationships to generate paper<br><br>[[🎯🧱💻marr 3lev]] | | goal of the section | Rationalizing agent's movement vector in spatio-temporal environment as discrete choices on signal, belief, uncertainty, utility, timestep<br> | 🎯Computation Target of each row should be solved with<br>- process in 💭Algorithm and Brick 🧱<br>- productized as 🔑Key | conditional on next row's 🎯 Target, and current row's 🔑Key in 🗣️language, 🎞️plots, 🗄️table, generate sentence that makes readers curious about the the next target, given key so far. |
## e.g.1
[[📝🧭Vectorizing Adaptation]]
| Section | | 🎯 Target | 🧱Algorithm Brick | 🔑Key in 🗣️language, 🎞️plots, 🗄️table | 🌉Bridging |
| ---------------------------------------------------------------------------------- | --- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ | --------------------------------------------------------------------------------------------------------------------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ |
| knowhow of using column relationships to generate paper<br><br>[[🎯🧱💻marr 3lev]] | | **Input:** Agent actions in uncertain environment<br>**Output:** Optimal action sequences<br>**Goal:** Environmental UNCERTAINTY for agent's FREEDOM<br>**Logic:** Equating spatial and temporal action paths | Rationalizing agent's movement vector in spatio-temporal environment as discrete choices on signal, belief, uncertainty, utility, timestep<br><br>♻️🧠🏎️📍 Process:<br>🧠B. Believing: $\color{Green}{p_c} \sim Beta(\alpha_c, \beta_c), \color{Purple}{p_r} \sim Beta(\alpha_r, \beta_r)
lt;br>🏎️P. Predicting: $\color{Green}{pred_c} = exp(N(\color{Green}{p_c}, \sigma)), \color{Purple}{pred_r} = exp(N(\color{Purple}{p_r}, \sigma))lt;br>📍U. Utility-based action: $\color{Red}{a^*} = \underset{\color{Red}{a \in \{a_c, a_r\}}}{\arg\max} \: \Delta\color{Red}{U}(\color{Green}{pred_c}, \color{Purple}{pred_r}, \color{Red}{a})lt;br><br>**♻️🌏⏰ Relaxations:**<br>⏰🌏TS: $\underset{\color{Red}{a}}{argmax} U(\color{Green}{\Delta c}, \color{Purple}{\Delta r}, \color{Red}{a})$ -⏰-> $\underset{\color{Red}{a},t}{argmax} U(\color{Green}{\Delta c}, \color{Purple}{\Delta r}, \color{Red}{a}, t) - \color{Green}{c(t)} t$ -🌏-> $\underset{\color{Red}{a},t,s}{argmax} U(\color{Green}{\Delta c}, \color{Purple}{\Delta r}, \color{Red}{a}, t, s) - \color{Green}{c(s,t)} tlt;br><br>🌏⏰ST: $\underset{\color{Red}{a}}{argmax} U(\color{Green}{\Delta c}, \color{Purple}{\Delta r}, \color{Red}{a})$ -🌏-> $\underset{\color{Red}{a},s}{argmax} U(\color{Green}{\Delta c}, \color{Purple}{\Delta r}, \color{Red}{a}, s) - \color{Green}{c(s)} t$ -⏰-> $\underset{\color{Red}{a},s,t}{argmax} U(\color{Green}{\Delta c}, \color{Purple}{\Delta r}, \color{Red}{a}, s, t) - \color{Green}{c(s,t)} tlt;br> | 🎯Computation Target🔏 of each row should be solved with<br>- process in 💭Algorithm and Brick 🧱<br>- productized as 🔑Key | conditional on next row's 🎯 Target, and current row's 🔑Key in 🗣️language, 🎞️plots, 🗄️table, generate sentence that makes readers curious about the the next target, given key so far. |
| Section | Research Question | 🧱 Literature Brick | 🔑 Key message | figure |
| ----------------------------------------------------------------------------- | --------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | --------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
| 0. relation between <br>columns<br><br> | blueprint of 🔐lock and key guides producing 🔑key | 🧱brick are how author of the literature i picked thinks about 🔐lock and key | 🔑key is produced by synthesizing bricks🧱 | 🌉 bridge to new 🔐 explains comes at the end of each section, summarizing how 🔐lock and key solved using 🔑key from brick 🧱 and hints what will be solved in the next section. |
| 1. Venture's Cognitive Resource Investment as Vector | How can we formalize entrepreneurial cognitive resource allocation as a vector with direction (operations vs market) and magnitude (decision clock speed), considering how bounded rationality leads to globally optimal but approximate decisions? | [[📜Gershman15_comp_rationality]]: Provides unified computational theory showing how bounded rationality leads to approximate but globally optimal decisions, helping explain why entrepreneurs use simplified heuristics despite their biases. Shows how processing speed affects decision quality under resource constraints.<br><br>[[📜Bhui21_resource_rational_dm]]: Provides mathematical framework for modeling how cognitive limitations shape entrepreneurial attention allocation and decision making. Shows how resource constraints lead to reference-dependent preferences, stochastic choices, and perseveration in repeated decisions.<br> | Cognitive resource allocation can be represented as a vector:<br>- Has direction (operational vs market focus)<br>- Has magnitude (decision clock speed)<br>- Under resource constraints, approximate decisions can be globally optimal | |
| 2. Vector's Speed (σ)<br><br> 🏎️Reaction speed under pressure | What determines the optimal magnitude (speed) of the resource allocation vector in terms of sampling strategy (high-bar vs low-bar) and how does decision clockspeed speed (σ) change under time pressure?<br><br> | [[📜stern17_control_exe(ent, strategy)]] models the trade-off between control and execution strategies for entrepreneurs - showing that control requires upfront investment and delayed market entry but provides future protection, while execution enables faster market entry but requires ongoing reinvestment. This resembles a time-accuracy trade-off, where entrepreneurs must choose between spending time establishing control versus making faster but potentially less protected market moves.<br><br>[[📜gans23_choose(ent, exp)]] formally models how firms rationally choose experimental strategies based on their predispositions - showing that incumbent firms facing disruptive opportunities optimally choose "high-bar" experiments (minimizing false positives) while entrants choose "low-bar" experiments (minimizing false negatives). This links how experimental design choices reflect and reinforce the control vs execution trade-off in entrepreneurial strategy.<br><br>[[📜tenanbaum14_1sample(1decide)]] shows mathematically why few samples can be optimal for decisions under time pressure - people rationally take more samples only when stakes are high or time costs are low. Specifically, it proves that when samples are costly in terms of time, making many quick but locally suboptimal decisions based on few samples maximizes the global rate of return, compared to making fewer but more accurate decisions based on many samples. | σ determines sampling behavior:<br><br>- High σ = execution strategy (fast, noisy, low-bar experiments)<br><br>- Low σ = control strategy (slow, precise, high-bar experiments)<br><br>- Optimal strategy depends on firm position (incumbent vs entrant)"<br> | ![[Pasted image 20241105193618.png\|300]]<br><br>Sampling without replacement<br><br>longer time step = larger sigma<br><br>![[sterman00_woreplactee_longtimestep.png\|300]] |
| 3. Vector's Direction<br><br>🧭 Rational Environment Adaptation Direction<br> | How do environments shape optimal $\frac{p_c}{p_r}$ ratio?<br>How does this change when two actions are interdependent (e.g. sourcing and sales strategies) | [[📜fine17_e2e]] shows how operational decisions that seem modular (like sourcing vs. sales strategies) are actually interdependent - providing precedent for modeling cognitive resource allocation choices as an integrated system rather than separate components.<br> | Agent adapts to the chosen environment using signals generated with different ratio $\frac{p_c}{p_r}$. Environment types include:<br>- Product type (physical/digital)<br>- Supply chain position (tier3 vs tier1)<br>- Degree of strategy interdependence | ![[🎞️🧭.png\|300]] |
[[📝👻phantom rationalize meaning]]