🪵 BayesSD meeting Log

[hyunjimoon]

@tomfid
I will log our trials during cooking meeting here!

For Sep.W4:

• I wonder what each dimensions of prior_pred means (especially 4) 20 is length of timeaxis.
  

• 🎼 Master B major to control F major: simulation of, by, for Box-Flow #25

• If "Using net flows does solve the problem of autocorrelation of the measurements" this is the reason we need to invent a structure in generator, could statistical model structure that estimates parameter assuming first order autocorrelation embedded time series (gaussian process; Mike Betancourt's robust gaussian process estimation) be a viable solution?

[tomfid]

Not sure what I'm looking at here.

[hyunjimoon]

Sorry, I will use @ button more prudently, but the above is prior_pred from this line.

[hyunjimoon]

@tomfid mentioned kicking off cookbook so I drafted the title and its theme "Constructing Bayes x SD joint space". Construction of product measure is in fact based on rectangles.

Definition of Product Distribution Function. If $F_1, \cdots, F_d$ are distribution functions on $\mathbf{R}$, then $F(x) \equiv F\left(x_1, \cdots, x_d\right):=F_1\left(x_1\right) \cdots F_d\left(x_d\right), \quad x \in \mathbf{R}^d$ is a distribution function on $\mathbf{R}^d$, called the product distribution function of $F_1, \cdots, F_d$. If each $F_1, \cdots, F_d$ is a probability distribution function, then so is the product $F: \mathbf{R}^d \rightarrow$ $[0,1]$. For a given distribution function $F: \mathbf{R}^d \rightarrow \mathbf{R}$, we define $\nu((a, b]):=\Delta_{b-a} F(a)$ on the class $\mathcal{G}$ of rectangles of the form $(a, b]$ (From Ioannis karatzas' lecture note).

Finished ContinuousCode folder reconstructing its structure (I wish our book to revolve around codes) and described in
https://github.com/Data4DM/BayesSD/blob/master/ContinuousCode/def(ContinuousCode).md .

[hyunjimoon]

Oct W1,2 @tomfid
- how to implement prior_pred in vensim

• Inventory management casestudy #17 (comment)
• completed draws2data for inventory management model
• introducing python's labeled multidimensionxarray:

# Xarray is multi-dim labeled array which would allow temporal and cross-sectional comparison
# It use `coordinates` and `data variables`; the latter is equivalent to pandas' column name
# arviz (Bayes computation visualization library) used netCDF data structure as it need
# labels (posterior, prior, prior predictive, loglik) for each of the parameters each of which has 1,000 samples

• 📦 Zipping casual loop as lookup or inter-dependent fraction from coflow structure #38 (comment)
• developing our discussion on B-major vs F-major:

which can be found in: marginnote3app://note/DC95299A-D9FF-43F5-A84D-FE143701D782 (may only open in my comp) - hierarchical and dynamic aggregation: ""best cluster" is not necessarily the same as "clustering" - discussion on measuring linger time and wealth for (who-knows-how-much-inherent-heterogeneity) red and blue ppl: "After an exponentially-distributed time they leave. This is basically the classic first-order well mixed system". As long as we measure quantities red and blue differs (f(theta1, theta2); not f(theta)) it derails from poisson process with exponential delay and becomes compound and/or conditional poisson process. - three diagrams with four questions; strategy https://github.com//discussions/7#discussioncomment-3859648 - https://github.com//discussions/21#discussioncomment-3834648

[hyunjimoon]

@tomfid the above is from one of Nathaniel's lecture which could be helpful initial point of the lingo mapping table. I started in wiki but I became very certain in one day that wiki may not be a good place (doesn't have access in mobile app).

[tomfid]

I like it! I'd like to build the same thing for SD/Bayes.

[hyunjimoon]

10/21 agenda @tomfid

• stanify demo
• when do yo use timeshift in vensim?
• revisiting Tom's comment on poisson process (daily customer)
• could cumtime function in stanify make timestep dynamic?

[hyunjimoon]

Oct W4

1. I am certain I wish to do research on win-win strategy for platform business (Ed and Johan) using System data science (Hazhir and Tom).
2. how to break HMC but not break other samplers, degeneracy (colinearity, hierarchy) bad parameterization example in SBC
3. process noise as part of data structure? Stan cannot solve SDE; record in Can we include variation from noise into variation from prior? #70
4. advise on paper or literature; structure Jair suggested but I don't wish to constrain this to readsdr as it lack auto-translation
5. vensim plots; Tom has used seaborn

• existing plotting code?
• issue opening Vengine on Windows. We are getting this error: “reinstall vensim (error 977)”
• resolved by older version of Vengine, and so I wondered if this is an issue with the newer version of Vengine.

---

Detail:
1.
Tool-wise: Stan

• simulation-based checks: attached paper allows real data and test-quantity (e.g. lp__) consideration in SBC
• stanify: couples managerial with coupled statistics and computation which Stan-enabled. draws2data2draws directly applies my SBC paper above. Three stanfiles translated from inventory management dynamics diagram below (diff.equations as function block here) and I wish to push this to hierarchical settings. Compared to stan-team effort above, I am lonely doing this alone!

Target-wise: affordable analytics
will meet Geoffrey Parker, author of platform revolution, next week to propose how platform can prevent inequality from non-affordable analytics problem.

Problem:
- What justifies this paper?
- Who's the target (audience) of this paper?
- Benefits/drawbacks of Stan & Vensim/Stan

Hypothesis:
How readsdr could address such a problem?
Efficiency
Mitigation of errors

Analysis:
How many case studies & which?
Describe Bayesian workflow, where each step is supported by a function from readsdr
Model building
Synthetic data?
Exploratory optimisation?
Prior predictive checking
Sampling (Hamiltonian Monte Carlo & other algorithm?)
Posterior predictive checking & other expectations
Simulation-based calibration

Donella's wisdom:

Places to Intervene in a System (in increasing order of effectiveness)

12. Numbers: Constants and parameters such as subsidies, taxes, and standards

11. Buffers: The sizes of stabilizing stocks relative to their flows

10. Stock-and-Flow Structures: Physical systems and their nodes of intersection

9. Delays: The lengths of time relative to the rates of system changes

8. Balancing Feedback Loops: The strength of the feedbacks relative to the impacts they are trying to correct

7. Reinforcing Feedback Loops: The strength of the gain of driving loops

6. Information Flows: The structure of who does and does not have access to information

5. Rules: Incentives, punishments, constraints

4. Self-Organization: The power to add, change, or evolve system structure -> self-governance

3. Goals: The purpose of the system

2. Paradigms: The mind-set out of which the system—its goals, structure, rules, delays, parameters—arises

1. Transcending Paradigms

Guidelines for Living in a World of Systems

1. Get the beat of the system?

2. Expose your mental models to the light of day?

3. Honor, respect, and distribute information. 

4. Use language with care and enrich it with systems concepts.

5. Pay attention to what is important, not just what is quantifiable.

6. Make feedback policies for feedback systems.

7. Go for the good of the whole.

13. Defy the disciplines.

14. Expand the boundary of caring.

15. Don’t erode the goal of goodness.

Geoffrey's wisdom from platform revolution -> design data-sharing api (collaborative sharing e.g. maintenance)

## 6. monetization (capturing value created from network effects)
- well-managed platform create excess value in four ways: access to value creation, access to the market, access to
tools, and curation. Monetization is about capturing a portion of the excess value created.
- Techniques for monetizing a platform include charging a transaction fee, charging users for enhanced access, charging
third-party producers for access to a community, and charging a subscription fee for enhanced curation.
- One of the most crucial monetization choices is deciding whom to charge, since the difference in roles played by various
platform users means that charging them can have widely differing network effects.
- Given the complexity of the monetization challenge, platform managers should take potential monetization strategies into
account in every decision they make regarding platform design.

## 8. self-governance: increase value and enhance growth
- prevent and mitigate market failures caused by information asymmetry, externalities, monopoly power, and risk
- The basic tools for platform governance include laws, norms, architecture, and markets. 
- Each must be designed and implemented with care in order to encourage platform participants to engage in 
positive behaviors, incentivize good interactions, and discourage bad interactions.
- Self-governance is also crucial to effective platform management. Well-run platforms govern their own activities following the principles of transparency and participation.

[hyunjimoon]

I wish to use this Fri 3pm meeting as one milestone to couple our efforts; this blueprint is one initial point :)

The goal is to seek optimization + visualization support for dynamic modelers (especially pure Vensim users) who I think are the greatest market segment (@tomfid is there any statistics on user segmentation of each tool? for instance, from the following three cases, what would be the ratio of each three? + how much % of people are we leaving out?)

To be very careful but honest, removing the use of optimization algorithm without diagnosis is one option. I've heard good comments on Stan's design decision of not offering Metropolis Hastings option.

@jandraor could you please prepare the summary of rvar doc by Matthew (and SD-connection) as we discussed last week?

I am meeting with two Stan devs before - will share “arviz’s wrapper or multi-index for posterior group can streamline 2,3” process on Fri.

• meaning of from Multimodality reported in hierarchical prey-predator #89
  "I also added diffusion among the regions. This is potentially pretty interesting, because it means there are two ways to get similar behavior among the 4 regions:
  low variety in the vital rate parameters, and similar initial conditions, so the regions are on similar but independent trajectories
  significant migration among the regions - it doesn't take much to couple the regional trajectories fairly strongly
  Hopefully this creates an added opportunity for multimodality, where the variety and migration parameters are confounded."

[hyunjimoon]

#95

[tomfid]

A couple additional items from my view:

• getting Stan versions of pred/prey with 4-6+ parameters working
• priorities for further Vensim models

[hyunjimoon]

For NovW4

• Formulation of parameter upperbound #108

[hyunjimoon]

@tomfid
I wrote down the recipe for our seasonal cuisine. If you get the output of any of the following steps before next Thursday, please do share! It would be so meaningful to share this NEW data structure (using only vensim!) with seminar students. No pressure though, but I am tempting you with the beautiful plot for eight school model as below :)

step 1. vensim -> inferencedata structure .nc

• organize vensim's Multimodality reported in hierarchical prey-predator #89 output into .nc
• arviz example code from dict or dataframe to .nc
• .nc stanify produced

step 2. hierarchical in vensim -> inferencedata structure (adding coordinate):

• once we reach inferencedata data format, you could directly run .plot_trace() to get the following (e.g. for eight-school hierarchical model):

# test code for plot
import arviz as az
idata = az.load_arviz_data("centered_eight")
az.plot_trace(idata, var_names = "theta", coords = {"school": ["Choate", "Deerfield"],}, kind = "rank_bars", figsize= (20, 15), rank_kwargs={"bar_kwargs": {"alpha": 0.2}})

step 3. sbc_inferencedata structure sbc_nc

From step3, sbc and its plot are state-of-the-art, so let's write history together! Since this is new for arviz as well, we (arviz developers) are focused on designing sbc plots for the hierarchical model.

• arviz example code from stan output to .sbc_nc
• .sbc_nc stanify produced

[hyunjimoon]

updated the location with stanify's .nc!

[hyunjimoon]

12W1 agenda (Tom, Jair, Angie)

• three questions from scale, prevent outlier, remove process noise for speedup #124
• visualization with rvar, sbc @jandraor

[hyunjimoon]

12W3 (Tom, Jair, Angie)

• 2. Typify parameters in syntax-semantic with colors, units, bounds #7 (reply in thread)
• plot design for the hierarchical model, which variable to compare other than sigma (https://discourse.mc-stan.org/t/best-practices-for-simulation-based-calibration-with-hierarchical-models/19630)
• https://github.com/stefanradev93/BayesFlow/blob/master/docs/source/tutorial_notebooks/Linear%20ODE%20system.ipynb
• spatial and temporal hierarchy #120 + Pinocchio productization project here with better rendering here

• can you share model from inventory and hybrid modeling in Getting Ed's agreement on need for sparsity-inducing prior #88

[hyunjimoon]

• hazhir's policy optimization alternating mode example
• a model that compares top-down vs bottom up regulation for angel and devil system(naming on Reinforce and Balance-intervention?)
• trust_factory

[hyunjimoon]

• connecting stan function with enroads (possible)?

[hyunjimoon]

• difference btw capacity and capital from 🎼 Master B major to control F major: simulation of, by, for Box-Flow #25 (comment)

[hyunjimoon]

with Tom and Jair

• can spatial hierarchy in spatial and temporal hierarchy #120 treated as multi-echelon setting with I) backlogs and ii) order-fulfillment logic? can I, ii live without the other?

[hyunjimoon]

Dec20 Angie, Hazhir
a. process noise's effect on inventory model

• two - three estimated parameters is boundary for the nonhierarchical model
• I am using time_saveper = .5, but DT is not passed to stancode.

rather than using (flow-flowStock)/tau with tau=DT, perhaps it would make sense to use tau = something small, but fixed and >2*DT. That way if the RK45-auto routine is slicing DT into smaller intervals, the integration will be stable.

• stocked flow formulation: with process noise scale 10, we sometimes get negative stocked flow.

more experiments in here

b. fire-fighting covid model: hyunjimoon/VaccineMisinf#4 (comment)

• partial model calibration? + convention to ensure internal consistency?
• remedies to extend boundary when hit by one (~ may be solvable by non-uniform prior; something as increasing estimated parameter numbers just to fit time-series)

[hyunjimoon]

posterior:
JanW1 todo

1. plot process noise (and show it is auto-correlated enough)
2. capping process noise to 2 to prevent negative stock; this prevents negativity leaking to flow which explains negative stocked flow

DecW3 summary

1. process noise is being injected every saveper (as saveper now replace time_step) as it should be
2. kill stuck chains with an extremely low variance if loglik is smaller compared to the correct majority 
3. removing stuck chains would improve pair plots
4. one layer for hyperprior should be added to both forward and backward model. Using two parameter (location and scale) distribution. i.e. a. to b.
   a. theta_r ~ normal(mu_loc, mu_loc*.1) mu_loc = .25,
   b. theta_r ~ normal(loc_theta, scale_theta), scale_theta ~ exp(1), mu_loc = .25
   c. theta_r ~ normal(loc_theta, scale_theta), loc_theta ~ exp(mu_loc), scale_theta ~ exp(1), mu_loc = .25

[hyunjimoon]

review:

1. move away from max, min formulation, prevent negative stock value
   

2. lower plot process noise scale, but it increase (.1 -> .3:

pn_scale= 1 gives divergence, decreasing stocked_

3. plot obs
   

[hyunjimoon]

Hazhir Angie 1W1

1. objective function log posterior (accuracy)

# add loglik for "matter ~ form"
1. parameter_draw ~ prior_func(hyper_param)
X. target_sim = f(parameter_draw)
2. observed_data ~ likelihood_func(target_sim)

Q1. how to make use of components of logposterior? e.g. how can modelers improve their model, if they found:
i) contraction: likelihood_func is more informative, but consistent with prior_func
loglik_s + loglik_ss << loglik_prior

ii) containment:prior_func is more informative, but consistent with likelihood_func
loglik_s + loglik_ss >> loglik_prior

iii) compromise: tension btw information encoded in 'prior_func' and 'likliehood_func'
loglik_s + loglik_ss ~ loglik_prior

generated quantities{
    // Define observed vector (matching vector)
    array[N] vector[R] s_obs_post;
    array[N] vector[R] ss_obs_post;
    // Assign generated value to observed vector (matching vector)
    for (r in 1:R) s_obs_post[:, r]=lognormal_rng(log(s[:,r]),m_noise_scale);
    for (r in 1:R) ss_obs_post[:, r]=lognormal_rng(log(ss[:,r]),m_noise_scale);

    real loglik;
    real loglik_prior;
    real loglik_s;
    real loglik_ss;
    for (r in 1:R) loglik+=normal_lpdf(adj_frac1[r]|adj_frac1[r],0.025);
    for (r in 1:R) loglik_prior+=normal_lpdf(adj_frac1[r]|adj_frac1[r],0.025);
    for (r in 1:R) loglik+=lognormal_lpdf(s_obs[:, r]|log(s[:,r]),m_noise_scale);
    for (r in 1:R) loglik_s+=lognormal_lpdf(s_obs[:, r]|log(s[:,r]),m_noise_scale);
    for (r in 1:R) loglik+=lognormal_lpdf(ss_obs[:, r]|log(ss[:,r]),m_noise_scale);
    for (r in 1:R) loglik_ss+=lognormal_lpdf(ss_obs[:, r]|log(ss[:,r]),m_noise_scale);
}

2. optimization algorithm (speed)

• sharing metric and init
  

3. effect of process noise during forward, backward

Q. what hypothesis interests you? can it be a paper?

• closer the estimated parameter is to the process noise source (e.g. production adjustment time is closer to process noise than target delivery delay) uglier parameter recovery becomes. if we can construct a controlled setting well, this sounds interesting
• higher the multiplicative process noise is (e.g. process noise scale .1 vs 3), the uglier parameter recovery becomes
• compare results with process noise structure on and off (i.e. scale =0) hypothesis?

4. Could non-linearity of generator possibly affect rank uniformity in SBC?

5. literature review on simulation-based inference

Q. plots?

i) rejection sampling based on an analytic solution which gives nonuniform joint prior as attached, from their notebook estimation for ode notebook
Q. could we use this strategy? could we have prior knowledge on forbidden areas, and can we get access to this wit sda?

ii) plot design by prior location and scale, noise fraction as attached from their paper Detecting Model Misspecification in Amortized Bayesian Inference with Neural Networks

iii) feel for magnitude on prior location (5), scale (2.5), noise fraction (.5)
scale is half the location and noise is 1/10th (difference between scale and noise?)

[hyunjimoon]

Bayes-self-consistency: prior-data-posterior = posterior-data-prior

detail: https://discourse.mc-stan.org/t/need-help-on-sbc-uniform-proof/17777/14

[hyunjimoon]

Different opinion on adding measurement noise for posterior predictive check in synthetic data context + source of process noise.

Regarding our discussion on adding measurement noise to target simulated from the parameters' estimated value, I think adding sampling uncertainty might be what we want for our purpose: to compare with observed data to judge estimation quality. Here observed data is simulated in our case. wi measurement noise, we are ignoring sampling uncertainty from below. Detail in https://mc-stan.org/docs/stan-users-guide/sampling-from-the-posterior-predictive-distribution.html

Regarding what I asked you last Fall, "can process noise correspond to predictor uncertainty?" (which you answered, it might), "posterior predictive sampling for posterior predictive checks is different from usual posterior predictive sampling discussed in the chapter on posterior predictions in that the original predictors x  are used." from this official document following supports this. I know it is confusing... but thought you might be interested as you thankfully emphasize the importance of different uncertainty sources in Heterogeneity and Network Structure in the Dynamics of Diffusion: Comparing Agent-Based and Differential Equation Models paper.

[hyunjimoon]

Tom, Jair, Angie 1W1

• 🪵 BayesSD meeting Log #23 (reply in thread)
• progress on 'degenerating joker`: obsidian-rendered and code
• Michael's advice:

As for the "stuck" Markov chains you have to differentiate between bad adaptation/exploration and actual multimodality. The example used in https://github.com/betanalpha/mcmc_diagnostics demonstrates how one can use useful MCMC diagnostics to differentiate between the possibilities.

• @jandraor does sir model have example of stuck chain and can we write a story of "differentiating bad adaptation/exploration and actual multimodality" affecting policy?
• attacked uniformity and evolving policy

[hyunjimoon]

Tom Angie 0210

• Q from Tom's chronic waste disease talk #140

• multi time scale #141

• end of feb. notebook for andrew's blog

[hyunjimoon]

Q. is it correct that smoothing customer order to expected customer order doesn't prevent the amplification (inflow surge from exogenous surge)?

instead of shipping -> selling, grocercy store; expected deamdn goes up desired inventory goes up -> inventory goes up (housing market eg: constrcution time is long - build more house

[hyunjimoon]

HW: reformulate the purchase decision to accelerate the transition, and analyze the resulting overshoot in purchases.

ev_stockfrac.mdl.zip

[hyunjimoon]

Q3. simulat. loop which I thought could be handled in vensim causes error

[hyunjimoon]

Q4. connecting SEIR in EV context to bridge Jair's estimation workflow with policy workflow

key q:
4.1. does SEIR (not sir) have an interpretation for tech adoption? i.e. how to formulate as close as possible with SEIR (what are epidemic-specific assumptions, let alone the different goals of policy to increase vs decrease infected)
4.2. what to estimate in SEIR EV adoption model (market)

4.3. feedback on market_ev_v1:
--potential fleet buyer as ice user's initial value?
4.4. how to fill the gap btw market and evc_supply model? i.e. how to translate market.buying_ev into evc_supply.customer_order? (detail in #142)

Q5. how simulator can contribute and when? https://www.cartsmobility.com/blog/trenton-moves-interactive-graphic-visualization-new

market_ev_v1.mdl.zip

Jair's notebook:
Archive.zip

[hyunjimoon]

summary of Tom's comment

• there was a paper /observation study that instantaneous info sharing during beer game didn't help much; we concluded there were many other factors (perverse, low q info) that hindered the great use of this info
• john's paper on organization learning includes E stage (backlog of ppl waiting for the product)
• classic SEIR has: S is ICE user, E is ppl waiting for EV, ADV (market saturation) happens in SE, WOM happens in E-SE and I-SE.
• For Ev, EV haver is connected to ordering (=SE) but this is equal to setting EV haver as I. Remaining problem is what to set/name R.
• capacity is fully utilized (advertising is useless bc S << E)
• EV-EVC handshake failure would not be logged in EV, but in EVC
• who would have the record of balking effect (EV has long lead time and high price so ppl don't order)

• to create overshoot people should buy vehicle more than needed which is possible by the increased slope of market share or size which can be represented as stocked flow below

1. increase market share:

• make ICE harder to use (price or infrastructure)

2. increase market size: grow the total fleet)

• which unit (user vs vehicle) to use?: user with the assumption one user has one vehicle; if the unit becomes household, 1~4 for each household

• higher order has lower deviation

[hyunjimoon]

@jandraor and Angie 0217
For #76, @Dashadower and Angie prepared https://gist.github.com/Dashadower/face0ff696d83a13d0ea1624b70c7917, and below are some agenda for the meeting with Jair:

• feedback on the notebook
• final format for the notebook that we'll share with Andrew
  -- how to keep format differences (from different authors) less visible
• is due Feb. feasible (due date) or should we reschedule? when?

20 policies (mask, anti-virus, search strategies, vaccine (supply chain)) each of them are not mutually exclusive and have dependence
2: testing and isolation, testing capacity (deploy it + building up capacity)
4: contact tracing (who did you have contact with?, quarantine - ) + preparedness, mask, contact tracing)
social distance
incorporating process noise -> CTMC is easier than compartment model (2003, Rohani zotero; The Interplay between Determinism and Stochasticity in Childhood Diseases)

[hyunjimoon]

common purpose: smaller loop affect the bigger loop

1. add sbc to workflow

• determine the degree of hierarchy
• estimation (ncp, cp)
  

2. actionable item

• 20 policies (propagate uncertainty on policies)
• EV vs ICE

[hyunjimoon]

Jair' model is SEIR (modulate people to go from I to R)

• we agreed blog feedback can increase the audience of readsdr paper and agreed by next week:
  J. jair: show how to overcome non-identifiability: 1-2-3- nonuniform sbc (degeneracy would be captured) - 1-2-3 uniform sbc
  A. angie: based on uniform sbc - EV context

plans for A:

- SEIR: potential adopters: `S`, wom: `I` are telling `S` (internal demands), `I` infected for a while you move to `R` 
what are epidemiology-specifc assumptions? among the assumption Jair is making (infection to go up vs down)

key q:
1. does seir (not sir) have an interpretation for tech adoption?
2. how to fill the gap btw `market` and `evc_supply` model? i.e. how to translate `market.buying_ev` to `evc_supply.customer_order`?
3.  with evc supply chain? (`market.buying ev` becomes `evc_supply.customer order`(charging station)
4. feedback on `market_ev_v1`: 
--`potential fleet buyer` as `ice user`'s initial value?

[hyunjimoon]

charlie, angie

1. the faster industry clockspeed is, the more reliable "revenue" serves as proxy for quality evaluation measure

• in slow clock industry, airbus which focus more on airplane quality beat boeing (1/3 -> 2/3 market share)
• growth and revenue is more reliable evaluation measure in fast clock speed e.g. IT. of course not all evaluation is eligible as being an objective function. growth and revenue would have low construct validity

2. we can analogize big pharmaceutical company's manufacturing flexibility with startup's operational flexibility: both are nail-stage actions (as industry and as one firm) to hedge against productivity risk during scale-stage

• follow up with charlie with list of data we need regarding flexibility metrics for pill making machine that aims to be product-agnostic

3. we will use tom's supermind ideator as one implementation of our idea "quality evaluation and alignment process for bilateral information asymmetry e.g. CEO and investor" in collaborative relation (building construction analogy)

4. we will use simulation to extract risk (objective := aligned among members) from uncertainty (subjective) in need, solution, fulfillment

5. 6 lens, 3 bias, 4 test diagram is completed here

[hyunjimoon]

charlie, angie

1. scale o4s diagram

• Q1. using scott team's paper how to represent thickness of need-solution adoption? + merge scott's operations with our ten keyword

2. nail o4e diagram; examples of six path?

• Q2. correspond cases with six paths? + next time paper on five stake holders and connect with tom's system structure

shared use of the facility by
academics, corporate researchers and start-ups would simultaneously lower the basic and applied
research costs of all researchers (by allowing them to share the use of the facility), allow for richer
interchange of ideas and knowledge in this particular area, and potentially spark direct research
collaborations and technology transfer from more academic research to industrial application and
scale-up.

3. test three class, choose two:

• transportation (inverse RL in panel data; m3s unsupervised panel data decomposition for interpretability embed causality to enhance interpretability with T45 #179)
• eom
• algorithm and behavior VS tepe (Q3. how imminent use is finance i.e. nail VS scale?)

4. computation with creativity talk

[hyunjimoon]

tom, angie

• paper review on
  1.behind alphago: Silver, D., Huang, A., Maddison, C. J., Guez, A., Sifre, L., van den Driessche, G., Schrittwieser, J., Antonoglou, I., Panneershelvam, V., Lanctot, M., & Dieleman, S. (2016). Mastering the game of Go with deep neural networks and tree search. Nature

2. hazhir and john's heterogeneity and network structure marginnote3app://note/B7287211-4B36-4EBC-9C47-B9E62A546D04

• mapping example of research with 8 nodes

• entrep. function from 2 here
• mit-google workshop startup genome pj

[hyunjimoon]

tom, angie
case studies (don't have model) - superstitious cause (lots of them)
comparison, not indep. on competitive env, (imbalance, you )

rescue is to build (reductionist, smaller dots - integrative models - connect them)

nation has more governance problem than firm
appropriability:=get paid from your activity (pollution)
pharma mental health - use tech to help keep medication (noncompliance)

value create, value capture (pill with edible chip) - customer were healthcare providers had high turnover (not interest in long term)

healthcare (tangled up) -BCBCB -> BCBABC
subsets of health good (less regulated) - EV consumer electronics, building material (manufacturing) - customers are builders (fragmented, not heavily regulated, end customer are clueless (most ppl don't know energy efficiency)) (pin)

harmful gene (turned on)
if don't aid greece we might demise VS if we invest in
sales function (different incentive; percentage of; internal externalities (going for easy sales, new customers not supporting old customers) bad genes out in the market)

2021 angie can't imagine implication -> won't invst in ( TIME PREFERENCE: discount; utility now, )

laws of physics (stability of estimates); advice i get about appropriate response (good, diametric; reguritating; )

pearl (asking - if i quit smoking now vs if i quit smoking ten years back then)

choosable actions back then (taking past action; past control actions; reproduction ratio is 2 - stop growth of disease -> lower than 1 (50% greater transmission VS lowering 25% policy, still grows; sort out; go back and have ten times as many death))

heterogeneity, avoid areas where huge network effect (social media; first mover advantage), random, physical product in un (regulation )

instrument (manufacter; local vision; exogenous in regional income; who within the nontradable are )
income rises

[hyunjimoon]

tom , angie
business_evol.pdf

1. Explicit is evolution (gene) and implicit is revolution (meme)
2. Evaluate and assumption should evolve (macro, micro, nano)
3. "rather than asking what tools would be required to do analytical strategy development, we instead ask what we could do with the analytical tools we could actually build" (solution based on need VS need based on solution)
4. "none of these can resolve three core problems of omitted variables, interaction effects, intercorrelation as these problems are not at root result of unique shortcoming of regression as a method, but rather inherent to the phenomenon under study." (phenomenon, theory, mesarue) as (need, solve, approximate) triplet

[hyunjimoon]

peter, angie

• can we model technology as one agent? environment: market? (adoptive, adaptive, leapfrog)
• meaning of "inherent to the phenomenon under study"?
• product market fit or business model innovation; examples of each

[hyunjimoon]

tom, angie
covid model, state level (use NY posterior for MA prior), kalman filter it, a group + (1-a) specific (block diagonal) LKJ

4*4 (track covariance; 54 * 4 -> 200 by 200; covariance of one state in NY and California stat)

customer oriented action (state of corr)
define genotype with phenotype; assume node is not a constraint (professional CFO)

reduce to something simpler (just the interplay of two or three action)
C then further C (take a model, freeze state space - others respond to that)
isolate the effect of action; scatter plot matrix (row, col are state, pick state); implicit mental state,

initial state (deer - physical, interpretation; alive or dead; invetory model: importance of S, SS..)
resource alloc: 89 paper; alpha_sl (0,1), alpha_s (inverse of time), beer game (doesn't cost to order more beer (1,1000case) resource allocation)

stockmng model gives uc bottleneck breaker (among AGCT); causal loop further decompose them into C_professionalize using pair

[hyunjimoon]

planning agenda for jb, charlie meeting:

• BAE3. update or updated phi (environment): #192 (comment)
• BAE1. argmax_theta (y|theta, phi's atomness) industry effect #190 (comment)

[hyunjimoon]

scott, @miaomiaozhang20, angie

The conversation revolved around the importance of a well-designed idea evaluation process, with Speaker 1 emphasizing the significance of both human and AI evaluators. Unknown Speaker suggested that expensive and expert evaluators can be beneficial. Speaker 1 and Speaker 3 concurred that the process should provide early-stage feedback and response, with a need for further research on how different compositions of founders can respond to different types of feedback. The speakers also discussed the importance of a good theory and experimental design in conducting empirical research, with Scott sharing his experiences and insights on the challenges he faced during his graduate school years. Unknown Speaker and Speaker 3 offered alternative perspectives on the distribution of responsiveness in a two-player game, highlighting the complexity of the issue. Through these discussions, the speakers emphasized the importance of persistence, creativity, and feedback in overcoming the challenges of empirical research.

Transcript

https://otter.ai/u/R4DLc9niOccwAVtu4lGz8-sC9Nw?view=transcript

Action Items
[ ] Review literature on how entrepreneurs respond to early stage feedback on their ideas
[ ] Identify key papers and references on this topic from judges paper
[ ] Come back with a tightened framework on what we know and ways to get at the question
[ ] Formally write down how the model can identify differences between solos and duos based on evaluator feedback
[ ] Figure out potential field experiment design and data access
[ ] Get high quality expert evaluators, not just random crowd evaluations
Outline
Beliefs and decision-making in entrepreneurship.
Speaker 1 expresses skepticism towards Bayesian reasoning, arguing that people are more likely to update their beliefs when they find someone who shares their views.
Speaker 1 discusses data quality and the idea of "T star" with Speaker 2, using an analogy involving Angie to illustrate their point.
Speaker 1 discusses starting with a primitive distribution of beliefs and choosing whether to act on an idea.
Entrepreneur confidence and insider information.
Speaker 1 expresses high confidence in their entrepreneurial idea, citing insider information and industry expertise.
Speaker 1 argues that entrepreneurs must have a more positive view of their idea than they do, and that this view is grounded in their subjective expected value.
Speaker 3 agrees that a prior distribution should be started from, but that it should have nice properties and be updated based on the entrepreneur's own activities.
Independent invention evidence and risk skewing in a paper.
Speaker 1 argues that Tom Aster's paper on independent invention evidence of unrealistic optimism is one of his most cited and loved papers.
Speaker 1 expresses encouragement for the idea, but notes that a well-developed theory and a successful field experiment are crucial for the project's success.
Speaker 1 highlights two challenges to the project: developing a good theory and conducting a successful field experiment.
Conducting research and experimentation in economics.
Graduate student studied with Amos Tversky and Lee Ross, inspired by their intellect and kindness.
Speaker 1 struggled to conduct experiments in the 90s due to lack of money and resources, leading to a shift in their career focus.
Speaker 1 was skilled at creating datasets and using computers to produce tables with stars or not stars, which was a difficult task at the time.
Evaluators with expertise and credibility, like Bob Iger, are more valuable than random evaluators.
Entrepreneurship and AI feedback.
Unknown Speaker: AI evaluator is good, Ethan Moloch is an AI influencer.
Speaker 1: Entrepreneurs respond differently to early stage feedback, depending on their composition.
Speaker 1 suggests using a two-by-two design to compare solo and paired responses.
Speaker discusses the importance of identifying competing forces in a research project, with a focus on the role of data and experimentation.

[hyunjimoon]

shiying (S), yichen (Y), angie (A)

angie summary

• S recommended rock health database, therapeutics casestudies (https://hbsp.harvard.edu/search?N=&Nrpp=25&Ntt=gene+therapy&action=&searchLocation=header&activeTab=products), Four Success Stories in Gene Therapy (https://www.scientificamerican.com/article/four-success-stories-in-gene-therapy/)
• Y asked how to define partnership (as it appears in scale stage) so angie explained there seems to be different types of partnership. partnership starts with A but AC-partnership is more distribution channel level (increasing delivery capability) and AT-partnership is more increasing manufacturing capability or experimental design (platform companies)
• Y asked how to model evolving environment and A answered it's better to choose slow clockspeed industry and benchmark from it

otter summary

Action Items
[ ] Come up with 3-5 business hypotheses to test for the next meeting.
[ ] Figure out what existing data is available to test hypotheses.
[ ] Determine a focused problem/literature gap to target for research.
[ ] Consider division of labor for dissertation work vs separate project.
[ ] Collect comprehensive datasets needed to build the tool as a career goal.
[ ] Focus on graduating within the next 3-4 years.

Positioning technology for a cardio company.
Speaker 1 struggles to position cardio company as distinct from technology company, seeking partnerships to address supply chain challenges.
Partnership nature and technology development in biotech industry.
Speakers discuss the importance of considering the nature of partnerships and how they vary across industries.
Speaker 2 discusses how new technologies in biotech industries can lead to the creation of startups.
Speaker 1 expresses concerns about the fairness of the sample and the resources available to startups in the biotech industry.
Speaker 2 mentions that investors typically invest in platform companies with a set of technologies and needs, and that the platform acts as a matchmaker to bring people together.

Biotech industry homogeneity and partnerships.
Speakers discuss the complexity of gene therapy and how it affects the homogeneity of startups in the field.
Speaker 2 suggests focusing on specific, homogeneous industries for partnerships to avoid confusion and improve performance.
Speaker 1 agrees and suggests deep diving into industries like biotech or gene therapy to understand partnership dynamics and implications for startup performance.

Entrepreneurship, MVPs, and startup strategies.
Developing concrete cases for collaboration with dogleg dog to add industry consideration and learn from past startup cases.
Speaker 1 is confused about the term "minimum viable product" (MVP) in the context of healthcare, as it seems to be different from other industries.
Speaker 2 provides an example of a framework for customer discovery in biotech, but acknowledges that it may not be applicable to all industries.
Speaker 2 proposes a framework for turning ideas into products, combining customer discovery and startup tactics.
Speaker 2 emphasizes the importance of considering industry dynamics, such as supply and demand functions, in identifying customer needs and forming market strategies.

Using machine learning to generate business hypotheses.
Speaker 3 expresses hesitation in helping with gene therapy/biotech due to lack of expertise in the field.
Speaker 2 expresses interest in using machine learning to generate business hypotheses, while Speaker 3 highlights the need for domain knowledge in a specific industry like biotech.

Capturing domain knowledge for AI tool development.
Speaker 1 emphasizes the importance of domain knowledge in understanding what attributes are important for a digital biotech company, as machine learning models will build on this database.
Speaker 1 highlights the challenge of capturing diverse attributes across companies and industries, requiring a balance of domain knowledge and machine learning capabilities.
AI-powered tool for analyzing partnerships in different industries.
Speaker 3 expresses concern about the definitions of RSS and AGC T, suggesting that different industries may have different meanings for these terms.
Speaker 2 mentions iterative updates to the AGC T taxonomy based on new business case studies and collaboration with different industries.
Speaker 2 hopes having expertise in different industries will help with record-keeping and analysis.
Speaker 1 questions the focus on one industry vs. going broad and deep, and the potential impact on deliverables and publishability.
Analyzing audio transcript for insights.
Speaker 2 proposes using Bayesian learning to map the evolution of startups and industries.
Startups' actions, not attributes, determine industry similarity.
Speaker 2 discusses using a hierarchical model to analyze the relationship between startup age, industry, and performance, with the goal of understanding how cross-industry comparison can provide valuable insights.
Speaker 2 provides an example of how Hard Komodo can be expressed as a regression, with the effect of age and industry on startup performance coming from the age itself and the industry startup effect.

Analyzing audio transcripts and industry evolution.
Speaker 3 raises the point that industries may evolve differently than their successful cases suggest, as companies adapt to changing environments.
Speaker 2 adds that the optimal strategy may be following the herd, rather than trying to identify unique characteristics of the industry.
Speaker 1 discusses the importance of understanding how different industries behave in similar situations, and how successful firms tend to respond in those situations.
Speaker 2 asks for clarification on how to apply this concept to their own industry, and Speaker 1 suggests drawing inspirations from other industries and innovating to build a competitive advantage.
Startup growth strategies and opportunity evaluation.
Speaker 2 seeks funding for their Ph.D. program, hoping to secure a grant to stay in their program.
Speaker 1 and Speaker 2 discuss the potential for learning from different growth trajectories in various industries, including startups.
Speaker 2 is concerned about evaluating opportunities and deciding when to invest resources.
Speaker 1 suggests startups struggle with balancing exploration and exploitation, and how to evaluate low-hanging fruit.
Using machine learning to generate hypotheses.
Speakers discuss the Blitzscaling hypothesis and its limitations, with one speaker suggesting that it may be an abduction rather than a hypothesis.
Speaker 1 discusses the challenge of testing hypotheses in the absence of data, suggesting a targeted approach to collecting data for specific problems.
Speaker 2 agrees, emphasizing the importance of identifying a clear problem to test and collecting data specifically for that purpose.
Speaker 1 and Speaker 2 discuss their research project, including potential problems and solutions.
Speaker 2 suggests using a hypothesis generating machine as a spin-off project, while Speaker 1 prefers to go deep on one topic before expanding to other areas.

Industry changes and partnerships in biotech.
Speaker 3 expresses concern about external factors triggering changes in industries, such as fraud cases like Theranos, and the need to factor in this uncertainty in modeling.
Speaker 2 jokes about not receiving positive feedback, but agrees to consider external factors in their analysis.
Speaker 3 notes that even in the biotech industry, there is a need for medications or products 20 years ago, and this need has not changed.
Speaker 2 suggests that industry changes can be a factor in determining the need for resources and allocation, and that this need can be influenced by factors such as uncertainty and the potential for success.
Speaker 3 discusses opaque partnerships in biotech industry, seeking clarity on product and ecosystem levels.
Partnership types and their implications.
Speakers discuss partnership types, including horizontal and vertical expansion, and their impact on operational visibility and market discovery.
Speaker 3 explains that horizontal expansion can involve partnering with other companies within the same ecosystem to leverage each other's strengths, rather than expanding into new markets or industries.
Speaker 2 asks if there are any examples of horizontal partnerships within the same ecosystem, and Speaker 3 responds that it is common for companies to form alliances to use each other's synergies in terms of distributors or manufacturers.
Startup aims to disguise vertical integration as horizontal collaboration through resource sharing.

Industry analysis requires understanding industry characteristics to make informed decisions. Comparing successful companies across different industries can be challenging, but drawing inspirations from other industries can offer potential benefits. External environmental factors can trigger changes in industries, and allocating attention in the market while balancing uncertainty in technology and resource allocation is crucial. In the biotech industry, partnerships can provide significant benefits by leveraging each other's strengths and resources, expanding distribution channels, and gaining access to new customers. However, carefully evaluating and analyzing partnerships is essential to ensure their success, and differentiating between different types of partnerships, such as horizontal and vertical partnerships, is crucial at the ecosystem level.

https://otter.ai/u/oYN69wUCZ1s3xzUyf3Mwpk7OMxk?view=transcript

AI terminology and industry-specific definitions.
Speaker 1 and 2 discuss the importance of domain knowledge in defining key terms in AI analysis, with Speaker 2 sharing their experience of iteratively updating AGC T definitions based on new business case studies.
Research deliverables and publishability in a field.
Speaker 3 considers the trade-offs between depth and breadth in their research, prioritizing publishable results.
Using Bayesian methods to analyze startup data across industries.
Speaker 2 proposes using Bayesian learning to map the evolution of different startups and industries.
Inferring industry similarity through startup actions, not attributes.
Speaker 2 discusses using a hierarchical model to analyze the relationship between startup age, industry, and performance, with the goal of identifying the unique effect of each factor on startup success.
Speaker 2 explains that cross-industry comparison can provide insights into the sequence or interdependencies of factors that previous data may not capture, allowing for a more nuanced understanding of startup performance.
Analyzing industry evolution and successful strategies.
Speaker 1 suggests that comparing successful cases within an industry may not accurately reflect the industry's evolution, as the environment and industry itself may have changed over time.
Speaker 2 adds that the challenges observed may not be the optimal case, and that following the herd may be a better strategy, highlighting the importance of considering the cultural aspects of the industry.
Speaker 3 discusses the importance of understanding how different industries can behave similarly in response to similar environmental conditions, and how this can inform decision-making and innovation.
Speaker 3 suggests that drawing inspirations from other industries and companies that have faced similar situations can help create a competitive advantage and novel solutions.
Startup funding, opportunities, and growth strategies.
Speaker 2 discusses the importance of benchmarking across different industries, citing flies as an example, and how this can be applied to long-living mammals and birds.
Speaker 3 expresses concern about funding for Speaker 2's PhD program, with Speaker 2 mentioning that they may need to look for additional funding.
Speaker 2 is considering how to evaluate and prioritize new business opportunities, struggling with the trade-off between investing resources in multiple opportunities versus focusing on one at a time.
Speaker 3 provides insights on how startups can approach this challenge, including using a real options framework or prototyping multiple opportunities sequentially.
Startup hypothesis testing and data collection.
Speaker 3 expresses skepticism towards Blitzscaling hypothesis, suggesting it's more of a "soft knowledge" that can be challenged and criticized.
Speaker 3 and Speaker 2 discuss the importance of grounding hypotheses in literature and data, and how abduction can be used to update hypotheses as new information is observed.
Speaker 3 discusses the challenge of collecting data for a research paper, particularly when there is a lack of clear hypotheses to test.
Speaker 3 suggests a targeted approach to data collection, focusing on a specific problem to test and collecting additional data to inform unexpected hypotheses.
Speaker 2 suggests generating hypotheses for their dissertation project, with Speaker 3 offering feedback and suggestions.
Speaker 3 discusses their approach to research, which involves going deep on one topic and then broadening their scope to consider other related areas.
Industry changes and partnerships in biotech and IT.
Speaker 1 expresses frustration with lack of support from others on their project, while Speaker 2 jokes about not receiving positive feedback.
Speaker 1 highlights the importance of external environmental characteristics in triggering sequences, as industries are no longer stable and facing increasing uncertainty.
Factoring in actual uncertainty can make tool more robust, as industry changes may impact resource allocation.
Speaker 1 suggests that biotech companies may partner to leverage each other's distribution and manufacturing channels, rather than solely on product level.
Speaker 1 recommends looking at announcements or press releases to understand the details of partnerships between drug companies.
Partnership types and their implications.
Distributors classified as partnerships, not suppliers or manufacturers, for differentiated supply chain management.
Speaker 1 explains that expanding horizontally within the same ecosystem can involve partnering with other companies to leverage each other's strengths, such as distributors or manufacturers.
Speaker 2 asks if there are any examples of partnerships within the same ecosystem that are not classified as horizontal, and Speaker 1 responds that it is hard to tell without knowing the specific details of the partnership.
Startup and M&A partnerships may involve vertical or horizontal integration, depending on the goals and resources exchanged.

[hyunjimoon]

Prob.Comp X BayesEntrep	theory 🟩 (desirability)	tool 🔴 (tech fbl)	input	output
🧬 genome	statistics X probabilistic computing - Bayesian Learning Approach to Entrepreneurial Strategy (s7.1)	bayesDB - AI-assisted inferential statistics: “what genetic markers, if any, predict increased risk of suicide given a PTSD diagnosis? and how confident can we be in the amount of increase, given uncertainty due to statistical sampling and the large number of possible alternative explanations?”
🗺️ map	business - knowledge transfer/reconfiguration between industries (with different clockspeed and hardwareness, charlie's expertise - Camuffo, A., Gambardella, A. & Messinese, D. (2023b). “Bayesianism and Unforeseen Events: Empirical Evidence from Two Field Experiments”. Academy of Management Proceedings 2023 (1). - sec.5.2 (Experimenting With Alternative Theories): "Therefore, entrepreneurs run experiments against alternative theories. 10 If entrepreneurs have an alternative theory with expected value Q (for example, an alternative opportunity, idea or strategy), then important changes in V – such as those induced by updating priors µ Θ pθq or priors on priors ω – are more likely to imply V 1 ă Q and thus a switch to the alternative theory Q. Besides, experiments that generate unexpected (“surprising”) outcomes – for instance observations that represent unforeseen contingencies (Karni & Vierø, 2013 and 2017), or regarding low probabilities states (Ortoleva, 2012) – may nudge the entrepreneur to consider novel attributes and causal links and, hence, be conducive of novel theories. Therefore, not only do entrepreneurs use experiments to test their prior beliefs about a future state space, aiming at higher V through a more favorable µ Θ pθq or higher ω, but also as a potential source of “anomalous” observations that can give rise to alternative theories (Mullainathan & Rambachan, 2023)." - decomposition "founder is “optimistic” about this theory, i.e. they believe in it and set a high V Θ, defined by (3) and high V , defined by (5). The founder’s strong prior belief (”optimism”) derives from a high ppx h | θ h q. At the same time, they think they have good technology and are confident that it can be put to good use, so they expect ppx e | θ e q also to be high. The conditional probabilities associated with the causal links ppx d | θ des , x e , x s q and ppx s | θ sh , x h q are also high because the available information on GAI adoption converges on indicating that the parameters are positive"	- JB's help on version control tools + visualization + evolution
🧭 compass	simulation X entrepreneurial decision making - "entrepreneurs change their prior beliefs over (1) new states that may emerge when they obtain awareness about previously unforeseen actions or consequences and (2) over previously null states that had been previously considered as not possible. Reverse Bayesianism enables entrepreneurs to adjust the state space they envision (their theory/opportunity) for unknown events. This process is consistent with the methodic doubt of scientists who acknowledge there might be other explanations/theories to what they study, which they do not know (Camuffo et al., 2023b). We can think of “Reverse Bayesianism” as “forward-looking” Bayesianism. - Scientific Approach to Entrepreneurial Decision Making as Bayesian Learning (s7.2) - "Bayesian learning occurs for entrepreneurs adopting the scientific approach at two levels. First, “scientific” entrepreneurs form, test and update their beliefs about the success or value of their ideas. Second, they form, test and update their beliefs about the generating mechanisms underlying such ideas." - First part is forward looking, Second part is about past (DGP)	- gen-finance

• A, T should be more relatively defined (process:product e.g. apple manufacturing chain/iphone1-3)
• b2b, b2c may differ as b2b sells process

how to
komodos: health map company

• testing the belief that doctors as distributers matters and the value of visualization would rise

[hyunjimoon]

below table is most updated version after feedback on time to pivot from biotech (david), it (yichen), transportation (bhuvan) experts

feedback-reflected table

pivot and belief defined on i) market segment (desirability) and ii) value proposition (feasibility)	Parameter	Hypothesis on Time to Change Market Segment (T2CS) w.r.t. Parameter = $\frac{\text{Time to change implemented desirability}}{\text{Time to change implemented feasibility}}$	Example in Biotech	Example in IT Industry	Examples in Ride Hailing	Examples in EV Battery Manufacturing	priority, note
within one market segment, across changed value proposition	DB: Diffuseness of prior belief -.003	HIGHER DB Hypothesis DB2T2CS: The more diffuse the prior belief on the segment (i.e., higher uncertainty), the longer the startup will stick with the chosen market segment before pivoting. Rationale: With higher uncertainty, the startup will require more experiments to reduce the uncertainty and make a confident decision about pivoting.	LONGER T2CS A biotech startup focusing on developing a novel gene therapy for a rare genetic disorder (Dravet syndrome) may stick with this market segment longer if there is high uncertainty about the therapy's efficacy and market potential. They may conduct more preclinical to gather evidence and reduce uncertainty before considering a pivot to a different therapeutic area or indication.	LONGER T2CS An IT startup focusing on a new cloud computing technology may remain longer in its initial market segment if there is high uncertainty about the technology's adoption and market demand. They might engage in more market research and pilot testing to reduce uncertainty before considering a pivot to another technology or market.	LONGER T2CS A ride-hailing startup focusing on a new market segment, such as senior citizens, may stick with this segment longer if there is high uncertainty about the demand and user preferences. They may conduct more market research and pilot tests to gather data and reduce uncertainty before considering a pivot to a different customer segment.	SHORTER T2CS An EV battery startup targeting a specific market segment, such as high-performance sports cars, may remain committed to this segment for a longer period if there is significant uncertainty about the segment's technical requirements and market potential. The startup would need to conduct more experiments and gather data to reduce uncertainty before contemplating a pivot to a different vehicle segment or battery application.	1, high uncertainty should be separated to "leap of faith" in one aspect but less (there's sth there but i'd like to learn more - using experiments as to find evidence to justify) in other aspect
	DR: Market Instability, volatility .02	HIGHER DR Hypothesis DR2T2CS: The more unstable the market is, the time to pivot would be longer as the startup needs to account for noise in the data generating process. Rationale: In a stable market, the deterministic mapping allows for quicker decision-making, while in an unstable market, the added noise requires more time to gather accurate information.	LONGER T2CS In an unstable market, such as during a global pandemic or significant regulatory changes, a biotech startup developing a COVID-19 vaccine may take longer to pivot to a different product or market segment. The instability in the market may require the startup to gather more data, assess the changing landscape, and adapt their strategies before making a pivoting decision. In contrast, in a stable market, the startup may be able to make quicker decisions based on clear market signals and trends.	LONGER DR n an unstable IT market, possibly due to rapid technological changes or economic downturns, a startup developing a mobile app may take longer to pivot to a different market segment or application area. The instability requires the startup to gather more data and assess the evolving market landscape before making a pivot decision. In a stable market, the startup could make faster decisions based on clearer demand signals.	SHORTER DR (seeking opportunity e.g. via, Mclaren) In an unstable market, such as during a global economic crisis or regulatory changes in the transportation industry, a ride-hailing startup may take longer to pivot to a different market segment or service offering. The instability requires the startup to collect more data, assess the changing market dynamics, and adjust their strategies before deciding to pivot.	SHORTER DR An EV battery startup operating in an unstable market, possibly due to fluctuating raw material prices or policy changes, may require more time to pivot to a different battery chemistry or target market. The market instability necessitates gathering more data and evaluating the evolving landscape before making a pivoting decision.	6, DR affects DB, SR
	SR: Size of market segment -.09	HIGHER SR Hypothesis SR2T2CS: Time to pivot would be longer if in a bigger market. Rationale: - higher expected revenue (expected future revenue is higher for a bigger market) - lower cost (cheaper to get bigger sample from a bigger market) N is the number of customer a startup can get from unit cost of experiment	LONGER T2CS A biotech startup targeting a large market segment, such as developing a novel cancer immunotherapy, may take longer to pivot compared to a startup focusing on a smaller market segment, like a rare disease. The larger market size offers higher potential revenues and allows for larger sample sizes in clinical trials, which may encourage the startup to persist longer in this segment before considering a pivot.	LONGER T2CS An IT startup in a large market segment like mobile gaming apps may take longer to pivot compared to one in a smaller niche like VR apps for education. The larger mobile gaming market provides higher potential revenues and more extensive user feedback data, which could motivate the startup to keep iterating within this segment longer before considering a pivot to a different app category or market.	A ride-hailing startup operating in a large market segment, such as urban commuters, may take longer to pivot compared to a startup focusing on a smaller niche like after work cab ride (uber/lyft:30%). The larger market size offers higher potential revenues and allows for more extensive user data collection, which could encourage the startup to persist longer in this segment before considering a pivot.	An EV battery startup catering to a large market segment, such as mainstream passenger vehicles, may take longer to pivot compared to one focusing on a smaller segment like off-grid energy storage. The larger market provides higher potential sales volumes and more opportunities for customer feedback, which may motivate the startup to continue iterating within this segment for a longer period before contemplating a pivot to a different battery application or market.
across different market segment, across different value proposition (components other than customer in BM)	BR: belief and real gap .03	Hypothesis BR2T2CS: The larger the gap between belief and real market conditions, the longer the time to pivot. Rationale: When there is a significant discrepancy between the startup's beliefs and the actual market conditions, it may take longer for the startup to gather enough evidence to challenge their assumptions and make the decision to pivot. The larger the gap, the more experiments and data points are needed to bridge the difference and trigger a pivot.	LONGER T2CS A biotech startup may initially believe that there is a high demand for a new type of antibody-drug conjugate (ADC) for treating solid tumors. However, if the actual market demand is lower than expected or if the development of the ADC faces unexpected scientific challenges, there will be a gap between the startup's beliefs and reality. In this case, the startup may need more time and experiments to validate their assumptions, potentially leading to a longer time to pivot to a different approach or market segment.	LONGER T2CS An IT startup may initially believe there is high demand for a new project management SaaS tool. However, if actual customer adoption is slower than expected or development faces unforeseen challenges, a gap emerges between their assumptions and market realities. The startup would then need more time and customer validation to test their hypotheses, likely extending the time before deciding to pivot to a different product or market segment.	A ride-hailing startup may initially believe that there is high demand for luxury ride-hailing services in a specific city. However, if the actual market demand is lower than expected or if competitors are already well-established, there will be a gap between the startup's beliefs and reality. In this case, the startup may need more time and experiments to validate its assumptions, potentially leading to a longer time to pivot to a different service category or target market.	An EV battery startup may initially believe that there is strong demand for a new battery technology with higher energy density. However, if the actual market adoption is slower than anticipated or if the technology faces unexpected development hurdles, a gap emerges between their assumptions and market realities. The startup would then require more time and validation tests to reassess its hypotheses, which could extend the time before deciding to pivot to a different battery chemistry or application.
	ER: Experiment opportunity (function of capital) +.09	Hypothesis ER2T2CS: Higher capital would allow the startup to have patience and wait until its experiment passes the threshold. Rationale: With more capital, the startup can afford to conduct more experiments and gather more evidence before making a pivoting decision.	LONGER T2CS A well-funded biotech startup with a large capital base may have the resources to conduct more extensive preclinical and clinical trials for their lead drug candidate. This financial cushion allows them to gather more comprehensive data before deciding whether to pivot to a different drug candidate or therapeutic area. In contrast, a startup with limited capital may need to make pivoting decisions earlier based on a smaller set of experimental results.	LONGER T2CS A well-funded IT startup with significant capital may have the resources to extensively test and iterate on their MVP with a large user base before making a pivot decision. The ample funding allows them to run more experiments, gather extensive data, and deeply understand the market before committing to a pivot. In contrast, a startup with limited funding may have to make pivoting choices faster based on early signals.	A well-funded ride-hailing startup with substantial capital resources may have the means to conduct extensive market tests and pilot programs before deciding whether to pivot to a different market segment or service offering. The ample funding allows them to gather comprehensive data and insights before committing to a pivot. Conversely, a startup with limited capital may need to make pivoting decisions based on a smaller set of experiments and early market signals.	A well-capitalized EV battery startup may have the resources to invest in multiple R&D projects and conduct extensive testing before deciding to pivot to a different battery technology or market application. The sufficient funding enables them to explore various options, gather comprehensive data, and make well-informed pivoting decisions. In contrast, a startup with constrained capital may have to make pivoting choices more quickly based on initial results and market feedback.
	CT: capital threshold before pivot market segment -.06	Hypothesis CT2T2CS:** A higher capital threshold before pivoting to a new market segment will lead to a shorter time to pivot." Rationale: With a higher capital threshold, the startup will be more impatient and less tolerant of the current market segment's performance. They will pivot faster to a new segment when the current segment fails to meet the higher capital threshold. This aligns with the idea that startups with higher expectations and lower patience are more likely to pivot quickly.	?? (ip/exec-ctrl) A biotech startup with a high capital threshold (e.g., requiring a high return on investment or a large market size) may have a shorter time to pivot if their current market segment fails to meet these expectations. For example, if a startup is developing a cell therapy for a specific cancer indication but the early clinical trial results do not show a significant improvement over existing treatments, the startup may quickly pivot to a different cancer type or therapeutic modality to meet their high capital threshold and maintain the interest of their investors.	?? (INACCURATE) An IT startup with a high growth rate threshold for their current SaaS product may pivot quickly to a new product or market if the growth metrics fall short of expectations. For example, if a startup expects their user base to grow 20% month-over-month but only sees 5% growth, they may rapidly decide to pivot to a different application or market segment that shows more promise in achieving their desired growth targets and securing the next funding round. The high expectations lead to less patience with the current path and faster pivoting decisions.	?? having ip would make it longer A ride-hailing startup with a high capital threshold (e.g., requiring a high growth rate or market share) may have a shorter time to pivot if its current market segment does not meet these expectations. For instance, if a startup aims to achieve a 30% month-over-month growth in ride volume but only sees 10% growth, it may quickly decide to pivot to a different customer segment or geographic market that demonstrates more potential in reaching their aggressive growth targets and securing future funding rounds.	An EV battery startup with a high profitability threshold for its current battery product may pivot rapidly to a new battery design or target market if the financial performance falls short of expectations. For example, if a startup expects a 20% profit margin but only achieves 5%, it may promptly choose to pivot to a different battery chemistry or customer segment that exhibits more promise in attaining their desired profitability levels and attracting investors. The high expectations result in less tolerance for the current approach and quicker pivoting decisions.	IT pivot example isn't accurate - expected growth rate based on info collected from experiments is ok
note		- Pivot is changing target desirability or feasibility during nail phase - Feasibility distribution is fixed to Logistic(0, 1) - Time to pivot implemented feasibility is uniformized to 1

feedback

DB (1)

• opportunity is discovered vs created (difference between; )
• competitor
• longer for sw, shorter for hw (as they may have chosen the segment with high advantage)
• via pivoted from b2c to b2b (shuttle)

DR (1)

• more test (keep other parameter same - concurrently)
• economic downturn (affect capital); more pessimic (=unstable)

SR

• size of market is a function of DR (if new tech invented)

BR

• how confident you are ()
• overestimating BR (optimism)
• overconfidence (me as the object) is more on reaction () wrong datapoint (confidence -> optimistic)
• good signal ()

ER

• experiment oportunity,
• pivoting too much is not good - more capital you can wait (persistent) - learn enough
• Impatience X Overconfidence = Doom (more money - less dedicated - survive for a longer time; less determine) change example from quantative

CT

• risk appetite
• high uncertainty (feature, testing) - unless its path dependent
• growth metric != (financial exp = revenue); expanding while loosing money
• benchmark metric (to fulfill)
• how much you're willing to spend in one market (risk apitite 10% (.9threshold) = run away from the market; risk apitite 50% = .5)

Yichen

🚜startup mining (NUS-MIT)

In the conversation, participants discussed various strategies for navigating market uncertainty and pivoting when necessary. Speaker 1 emphasized the importance of experimentation and data-driven decision-making, while Speaker 2 highlighted the role of risk appetite and market stability. Unknown Speaker added that overconfidence can hinder market learning, and that persistence is key. Later, Speaker 1 and Speaker 2 discussed the challenges of navigating market dynamics for digital startups, emphasizing the importance of iteratively updating products based on customer feedback and optimizing capital allocation. They also explored the interplay between optimism and overconfidence in entrepreneurship, with Speaker 1 highlighting the gap between belief and reality and Speaker 2 explaining that optimism is necessary but must be balanced with a realistic assessment of the market and one's own abilities. Finally, Speaker 1 and Speaker 2 discussed the relationship between funding and overconfidence, with Speaker 2 suggesting that more money can lead to less determination and more pivoting, while Speaker 1 argues that a lack of money can make startups more geared and less overconfident.

Transcript

https://otter.ai/u/lGTACMwN5N7QZwfj1Z3vRoMzOGo?view=transcript

Action Items
[ ] Update the example used in the hypothesis about the size of the market affecting pivot time, as quantum computing is not a good example.
[ ] Consider ways to quantify how learning about a market affects the time to pivot, and look for papers that model this process.
[ ] Discuss ways to theoretically justify when parallel exploration is better than sequential exploration in entrepreneurship. Potentially explore setting up metrics or ratios to guide this.
Outline
Experimentation and market beliefs in startup development.
Jen and Speaker 2 discuss implementing ideas through simulation models, with Jen sharing a hypothesis and seeking feedback.
Startup founder seeks feedback on testable hypotheses for market acceptance and product visibility.
Product development strategies in digital startups.
Speaker 2 mentions that intrapreneurs often update their products based on customer feedback, using an iterative process.
Speaker 1 agrees that confidence in a product can influence the likelihood of making changes to the market versus the product itself.
Parallel experiments are common in digital startups, not just biotech.
Speaker 2 explains that parallel testing is encouraged in biotech and therapeutics when time is of the essence and there's high uncertainty in the market or feature, as it's less costly and can save time.
Speaker 2 also notes that while parallel entrepreneurship is common in both biotech and digital industries, the cost involved in doing parallel testing is relatively low compared to starting from scratch in digital industries.
Entrepreneurship and risk tolerance.
Speaker 2 suggests a guiding ratio for parallelization, but most papers take an inductive approach with no clear guidance.
Nubar FRB, CEO of Moderna, discusses parallel entrepreneurship and its potential benefits.
Speaker discusses their beliefs on market fit, technology, and capital in the context of product development.
Market knowledge and opportunity creation.
Speaker 1 suggests that the more diffused one's prior beliefs are on a market, the longer it takes to pivot.
Speaker 2 argues that knowledge of a market is not always proportional to competitive advantage, and that there can be a gap between perceived knowledge and reality.
Speaker 2: Market feedback reveals gap between expected and actual needs.
Market stability and time to pivot in startup development.
Speaker 2 expresses uncertainty about appropriate term for initial state in entering foreign market.
Speaker 1 and 2 discuss the importance of gathering data and validating experiments in an unstable market.
Speakers discuss the impact of market uncertainty on priorities and decision-making.
Entrepreneurial risk appetite and market stability.
Entrepreneurs' risk appetite affects their decisions, including market size and stability.
Markets with larger sizes and higher revenue potential may take longer to pivot due to lower tooling costs and higher expectations.
Startups may take longer to pivot due to overconfidence and optimism, with larger gaps between beliefs and reality requiring more experiments to bridge the difference.
Entrepreneurship, optimism, and overconfidence.
Entrepreneurs may overestimate market potential due to optimism, leading to erroneous entries.
Overconfidence in one's abilities can lead to poor decision-making.
Speakers discuss optimism and overconfidence in decision-making, with a focus on the role of competence and judgment.
Experimentation costs and market patience in startup development.
Overconfident founders are less willing to pivot, but higher capital allows for more experimentation and data gathering before deciding on a direction.
Speaker 1 discusses the importance of patience and capital in experimentation, suggesting that having a lot of capital can help companies stick with one market for a long time and gain valuable information about it.
Speaker 2 expresses skepticism about using quantum computing as an example, citing the high cost of building such a system.
Market learning and overconfidence in startups.
Speaker 2 suggests qualitative papers on learning the market, but no quantitative modeling work has been done.
Speaker 1 suggests that overconfidence and impatience can lead to poor market decisions.
Startup funding and pivoting strategies.
Speaker 2 mentions that having more money can make one less overconfident, as they have more chances to ignore negative feedback and keep trying until something proves their beliefs.
Speaker 1 suggests experimentation as a way to overcome overconfidence, as it allows for feedback from multiple sources to update one's beliefs.
Speaker 2 suggests that having more money can lead to less determination and more pivoting in a startup.
Speaker 1 discusses a decision rule for pivoting in a market, with a higher capital threshold leading to faster pivots.
Startup growth and capital investment.
Speakers discuss the limitations of using growth matrix as a sole metric for evaluating startup success.
Unknown speaker suggests threshold for experimentation is $100, while Speaker 1 prefers to spend more time in one market to gather data.
Unknown Speaker: "Just how much you're willing to spend in one market" (1:12:38)
Speaker 2: "Risk appetite is how much risk you're willing to take, as a financial term" (1:14:20)

Bhuvan

Speaker 1 and Speaker 2 discussed the importance of identifying and targeting niche markets, with Speaker 1 emphasizing the value of focusing on a specific geography or segment to increase revenue and profitability. Speaker 2 agreed and added that startups could consider pivoting to other markets if the current market is not viable. Later, Speaker 2 and Speaker 3 discussed their efforts to analyze and improve transportation management through data analysis, with Speaker 3 providing an in-depth analysis of traffic signal intersection data. Speaker 2 and Speaker 3 also discussed their plans for data analysis and research collaboration, with Speaker 3 expressing interest in using the data for their research.

Transcript

https://otter.ai/u/JKHPpfmDYOIiSePVyPSYnLXdrWI?view=transcript

Outline
Pivoting a startup based on market research and cost analysis.
Company may pivot to other markets if profitability per trip is questionable due to high costs and variability in demand.
Speaker 1 mentions that their startup is considering pivoting to a more premium market, such as airport drops, to increase revenue due to high costs.
Speaker 2 expresses concern about pivoting too frequently, citing recent research that suggests it can negatively impact market learning.
Uncertainty in market and product requirements leads to experimentation and R&D investment in manufacturing.
Pivoting in uncertain markets with a focus on hardware and software.
Speaker 1 observes that hardware startups invest more time before starting a company, which can make it harder to pivot to adjacent markets.
Speaker 2 suggests that market stability can affect the time it takes to pivot, with unstable markets like during COVID leading to more diffuse customer preferences and longer times to pivot.
Speaker 1 discusses how uncertainty during COVID led to pivots in business strategies, with examples from McLaren and friends' game rooms.
Speaker 2 highlights the importance of moving fast in a volatile market to seize opportunities, with a focus on hardware vs software pivots.
Speaker 1 mentions McLaren Automotive and their pivot to delivering breathing devices during the pandemic.
Speaker 2 discusses the dynamic nature of the EV market and how startups are pivoting faster in response to regulations and market demands.
Entrepreneurship, market segments, and hardware/software differences.
Examining the cost-benefit analysis of entrepreneurship, speakers discuss the importance of market size and willingness to spend time in startups.
Speakers discuss market size and hardware/software differences in transportation industry.
Tesla started in a niche market with the Tesla Roadster, then expanded into the mainstream market with the Model S and Model 3.
Entrepreneurship and market expansion strategies.
Tesla's expansion and pivoting strategies discussed, with examples from the automotive industry.
Tesla Roadster launch was a "halo product" to gain attention and funding, which led to a pivot to B2B and B2C services.
Market segmentation and product development.
Speaker 2 is confused about the term "b2b to b2c" and wonders if it has different connotations.
Speaker 1 explains that the term "b2b" refers to selling to other businesses, while "b2c" refers to selling to individual consumers, and that the market for b2b can be larger than b2c.
Speaker 1: Gap between belief and reality determines speed of pivoting (33:57)
Speaker 2: Longer gap, faster pivot; shorter gap, slower pivot (35:18)
Entrepreneurial pivoting strategies.
Experimentation and capital are key to success in business, as seen in Tesla's ability to survive mistakes through Elon Musk's financial backing.
Entrepreneurs face challenges in both finding new markets and improving their products.
Pivoting in startups based on cost and feasibility.
Speaker 2 compares cost of market change vs cost of product change, finding that market change is more expensive due to longer time frames.
Speaker 1 agrees, noting that well-funded ride-hailing companies can extend markets faster at an early stage.
Speaker 1 suggests that startups with limited capital may need to delay pivoting decisions due to the small set of experiments, leading to more doubt and lower desirability to change.
Speaker 2 proposes an alternative approach to pivoting, defining it as exiting the previous month and choosing the next month, potentially resulting in slower decision-making.
Startups with more risk appetite pivot faster, while those with less risk appetite experiment more.
Pivoting in startups, IP, and market dynamics.
Speaker 1: Pivoting fast and pivoting more may have a correlation, but forced pivoting can be detrimental (e.g., investor pressure).
Speaker 2: Hardware pivots may be challenging due to lack of means to convert, while software pivots may differ in phases and market failure.
Speaker 1 discusses testing in a sea port and later meets with Speaker 2.
Speaker 2 mentions that IP can make it harder to pivot to other markets, citing Uber Black as an example (1:00:05)
Speaker 1 agrees that IP can be a barrier to change, but notes that it can also provide a competitive advantage (0:59:56)
Uber may pivot to a new customer segment or product line with more optimism and capital.
Transportation data analysis and simulation modeling.
Speaker 3 expresses hesitation about collaborating on a project due to concerns about data quality and ease of analysis.
Speaker 2 expresses lack of ambition and concern about data processing, while Speaker 3 offers to help with data collection.
Speaker 3 explains their process data is not about bus arrival times, but rather how long the bus can wait at an intersection.
Speaker 3 explains traffic signal delay analysis using histograms and simulations.
Analyzing traffic data for policy recommendations.
Speaker 2 seeks data on delay types to support argument for different policymaking approaches.
Speaker 3 is researching ways to reduce traffic delays by analyzing data on bus drivers' fatigue, signal priority, and dwell times.
Speaker 2 is designing AI for startup operations, recommending investments based on data-driven insights for each individual startup.
Speaker 3 suggests using parameter fitting for HLA type to improve prediction accuracy (1000 SATs).
Speaker 2 proposes separating chronic and acute heart disease data to develop policy prescriptions (4-5 years of interest).
Data analysis and simulation modeling.
Speaker 2 suggests analyzing data to write a paper on the impact of manager intervention on project delays.
Speakers discuss potential collaboration on a cost project, with one speaker interested in using a simulation model and another speaker sharing data for further research.
Speaker 3 seeks help analyzing data for proposal, struggles with obtaining relevant information.
Traffic data analysis and visualization.
Speaker 2 seeks data to identify most heterogeneous delay distribution, focusing on northbound and southbound delays.
Collaborator suggests sharing post-processing data to help with project.

[hyunjimoon]

tom, angie

classification of people (atomized bit)

• a person is an atomized bit, modularized thoughts that are mostly consistent
• I decipher signals (e.g. writings and voice) to learn their bit/soul
• based on the following label of person's function into one of the four quadrants
• each person I label people's function into one of the four quadrants

bit is flow, atom is stock, Q. role of energy to fasten bit2atom?
example of bit: technology, less situated, higher rate of success for observer
example of atom: asset, more situated

1. label $bit_t$

BE stands for bayesian entrepreneur, PC stands for probabilistic computing

past\future	observer 👀	actor 🤜
effect of feature	describe [[def(BE_⬇️)]], [[def(PC_⬇️)]]	explain [[def(BE_⬆️)]]
prediction	predict [[def(PC_⬆️)]]	synthesize [[def(PC_🔃)]]

• ⬆️, ⬇️, 🔃 is ability to act, think, think-then-act
• 🔃 requires time stone, which increases act-think clockspeed through simulation (detail in [[W11_Stones and Gauntlet]])
• [[def(PC_⬇️)]] desires [[def(PC_🔃)]] and with easy-to-use tool, persuasion success rate would be over 50%

2. label $atom_t(bit_t)$

• [[def(BE_⬇️)]] is observer or actor using non-computerized human language (non-scalable approximation) to explain certain effect of feature

• [[def(BE_⬆️)]] is actor choice of adapting before theorizing (industry mentors)

• [[def(PC_⬇️)]] is observer choice of using measurement, econometrics, causal inference to explain effect of feature

• [[def(PC_🔃)]] s

3. persuade $atom_t(bit_t) \rightarrow atom_{t+1}(bit_{t})$

• order fastest to slowest clockspeed action and plan out change

4. persuade $atom_{t+1}(bit_t) \rightarrow atom_{t+1}(bit_{t+1})$

---

5. label $bit_{t+1}$

social science testing with genome

• https://github.com/orgs/Data4DM/projects/3/views/2?pane=issue&itemId=57482640
• marginnote3app://note/5F80E6CE-B741-430D-8B66-149DA0D1C8E7
• bayesdb (IS THERE STH truly causal? my hypothesis example) in knowledge production system

transfer learning and hierarchical bayes in transportation

• predicting startup growth with hierarchical gaussian process #200
• spatial and temporal hierarchy #120
• connection with ford and boland paper on fast-slow dynamics modeling
  Bloand17_cont_servnet_design.pdf
  Ford18_fast_slow_sim.pdf

[hyunjimoon]

gal, angie

pre-meeting

given our goal dynamic canvas.pdf
q.

• how to intuitively deliver three axis (direction, speed, accelaration) to startups?
  

post-meeting

[hyunjimoon]

miaomiao, angie

prior

1. AM's prior that's in the best form to be evolved to SMA prior (S:Scott, M:Miaomiao, A:Angie)
2. syncing A and M's prior on labeling for RAG

just in general, supply side (feasibility) - A - M - S - demand side (desirability)

likelihood

In the first part of the conversation, Speaker 1 and Speaker 2 discussed entropy reduction mechanisms in GPT prompts, with Speaker 2 emphasizing the importance of experimentation to find a product-market fit. In the second part, Speakers 1 and 2 discussed their concerns about a research project they are working on with Scott, weighing potential risks against rewards. Speaker 1 proposed an alternative idea and sought feedback on their other ideas, including the role of VCs and the difference between entrepreneurs and founders.

Transcript

https://otter.ai/u/s4_sbCL4NR1s-rfnM5JVf1RtZXU?view=transcript

Action Items
[ ] Have another meeting with Professor Scott to provide a final update before potentially discontinuing the project.
[ ] Prepare to simulate their conversation with Professor Scott to determine the best sequence and topics.
[ ] Speaker 2 to present their initial simulation model on type 1 vs type 2 errors in belief updating at the next meeting.
[ ] Schedule regular monthly meetings to continue discussions and work towards presenting at a future conference.
Outline
Entropy reduction mechanisms for information theory.
Unknown Speaker discusses entropy reduction mechanism and information theory, seeking to integrate ideas from different sources.
Speaker struggles to connect ideas and find a cohesive theoretical framework, awaiting feedback from Chris.
Improving AI models' noise tolerance through experimentation.
Unknown Speaker discusses experiment on adding noise to GPT prompts, hoping to observe differences in response between solo and dual groups.
Speaker wonders if tighter noise distribution in solo group might lead to less varied responses, while dual group's ability to detect signal from noise could be advantageous.
Experiment design and theory validation.
Unknown Speaker discusses experiment and solo, confusion between theory and experiment.
Speaker questions why solo is titled solo, and if experiment can show tighter or getting tighter.
Model development and uncertainty, with disagreement on approach.
Unknown Speaker expresses uncertainty about proposed model, disagrees with Scott's wording.
Modeling project with conflicting opinions.
Unknown Speaker discusses potential risks and rewards of a project with Scott, expressing uncertainty about the project's direction.
Speaker mentions Daniel's strengths in organizational model building, but acknowledges their own limitations in this area.
Experimentation and entrepreneurship with a focus on modeling and entropy.
Unknown Speaker discusses ideas for experiment with Jackie, prioritizing their needs.
Speaker seeks feedback on theoretical angle, experimental setup from Scott.
A project with a colleague, including potential next steps and concerns about detachment from reality.
Unknown Speaker suggests meeting with Scott to discuss project and prioritize tasks (0:17:52)
Unknown Speaker worries about detaching from reality if project doesn't move forward (0:19:40)
Entrepreneurship, quality, and risk preferences.
Unknown Speaker discusses entrepreneurship, quality, and belief updating process.
Speaker wants feedback on theory about entrepreneurial firms and risk preference.
Simulation project, type one and type two models, and confidence.
Unknown Speaker discusses simulation project with colleague, focusing on type one and type two models.
Speaker seeks colleague's input on how to present research at an entrepreneurship conference.
Product and market changes based on experimentation and data analysis.
Unknown Speaker discusses solo and duel actions in product-market fit, with solo actions focusing on changing the product and duel actions challenging the model itself.
Speaker suggests that duel actions may be more effective in questioning the market, as individuals with more experience in different markets may be more willing to do so.
Experiment progress and goals for a Bayesian conference.
Unknown Speaker discusses experiment goals and progress with Scott, including sharing their goal for presenting at a conference.
Unknown Speaker and Scott discuss the importance of clear communication and setting aside dedicated time for their conversation.

posterior

schedule for 30min meeting:

1. M2’s progress so far

• table's type1,2
• hard & not high expertise and interest in econ. modeling
• interest in type1-2
  

2. A's type1-2
3. next step