Dr. Vivienne Ming

edtech flooded with well intentioned people making no impact whatsoever

get out of mindset that only you have the solution

people solve problems for themselves (Coursera predicated on idea that billions of people will

spend time watching ML videos for hours on end)

those who come already have a degree

everyone in the Bay thinks they’re a social entrepreneur(do good+get rich)

at some point someone (board) will make you choose

be clear to everyone when pivot is made

goal: to make better people

built tech. with google glass to help autistic people reach facial expressions

problem w/ GG was obsession w/ making it lightweight, social, and fun

treating diabetes, ticking boxes led to one time treatment pump

impossible to build an unbiased robot

approach is just in time prompts that  help people take right action

artificial pancreas like thermostat, at best reactive, added deep learning for better predictions

bipolar disorder costliest disorder in world

what if you knew a month ahead, and key people are contacted?

building superpowers

first step in data science project is read the literature or find an expert

no new problems, have been looked for last 100 years

massive tax on being different

37% of biased hiring companies out of business in NYC in 5 years

more you step out of streotypes less people will stereotype you (trad. models don’t apply)

zero demographic hiring useless, trust is most important in hiring teams

more you emphasize merticoracy, less meritocratic it actually is

best performes have endogenous incentives, don’t care about external, do it because it matters

predictive value for were you’ll be 10 years from now, careers, directions, etc.

overcame desire to please parents, impress friends, etc

do you have a larger purpose?

future of work: creative adaptive problem solver

SV talent wars will be on global scale, ruthless

Muse: concerned not w/ high grades, but relevant ed, long life, healthy and creative

Shiftgig: built sequence of work experiences that help on demand workers move out of economy

AI marketplace to connect enterprise shifts with workforce

patents are roadblock to helping people, have multiple cos. solving challenging problems

The Tax on Being Different


Biotech Investing Lessons

Any good investor worth their salt has a list of rules and lessons derived from their experience in the markets. They don’t necessarily need to slavishly follow them, but it provides a good set of parameters for making decisions, especially during rocky and tumultuous times. Below is a list that I will be constantly updating, with examples, some original, some borrowed, but all useful.

  1. Scale into positions when betting on binary outcomes, whether it be FDA approval or trial results (e.g. burned on ADMP, bought entire ~$500 position in one go, but more sensible to make 3-4 buys, especially with zero transaction costs thanks to Robinhood)
  2. Note main drug in pipeline and the disease that it is targeting; certain types of drugs (biologics vs. NMEs) have different rates of approval
  3. Ladder out of successful positions by selling 10% at 50% gain, another 20% on a 100% gain, and another 20% on a 200% gain (from original position). The remaining 50% rides on house money, and you take profits while ensuring you don’t miss out on next blockbuster company (not Blockbuster). Credit to Bret Jensen for this one.
  4. Sell at least half of stake after significant setback, whether it be restructuring, trial failure, or CRL, assuming no other drugs in pipeline. Double down very rarely. This would have been helpful on stocks such DVAX, CEMP, among others. The bounce after failures is sharp and sudden and most usually miss it. One is better off taking their lumps as there are far too many opportunities out there to analyze.
  5. The corollary to the above is if the loss is greater than 50%, consider other assets the company has. Sharp selloffs are usually overdone and you can usually wait a bit and get a slightly better price. A good example of this is PTLA; after a CRL that shocked many, it recovered to it pre-setback price in a matter of a few months.
  6. Have exit price when getting in on a stock based on technicals. This lesson was not heeded on GALE, NVAX, among others. Usually when playing a bounce. ignore the fundamentals, have a profit target in mind, and when it hits, get out.
  7. ACCOCHIAS: Analyst reports, Catalyst, Chart, Option Chains, History, Insider Activity, SEC filings. Credit to Kyle Dennis
  8. Adam Feurstein: How to Spot Red Flags in Clinical Trial Data
    • single arm trials more difficult to implement because they lack a comparator arm
      • especially with progression free survival or overall survival endpoints
    • when press release hits the tapes:
      • read entire thing, especially if headlines says “positive,” or other superlative, as problems likely hidden in waning paragraphs
      • look at design of trial, # of patients enrolled, primary and secondary endpoints
        • reassuring words: intent to treat analysis, statistical significance, achieve primary endpoint
        • not reassuring: per protocol, retrospective analysis, responder analysis, subgroups, modified intent to treat, trend, grade 5 toxicity
          • credibility of subgroup analysis requires additional questions and research
        • changes made to design, including endpoints? double check info on clinicialtrials.gov
          • patients cannot disappear! look for Ns

10 Attributes of Fundamental Investors

Via Michael Mauboussin

  1. Be numerate (understand accounting)
    • focus on PV of FCF
    • path of ROIC (how is value created)
  2. Understand value (PV of FCF)
    • life cycle of industry, competitive position, barriers to entry, economies of business, management skill
  3.  Properly assess strategy
    • drivers of profitability and competitive advantage
  4. Compare effectively (expectations vs. fundamentals)
    • thinking by analogies fails to reflect full reference class (e.g. turnaround story similar to last one, or all ones?)
    • limit comparisons to attributes and you therefore miss essential insights
  5. Think probabilistically (few sure things)
    • constantly consider probabilities of various outcomes
    • subjective probabilities problematic with black swans
  6. Update views effectively (beliefs are hypotheses to be tested, not treasures to be protected)
    • actively open minded, offsets confirmation bias
    • avoid false causations while looking for info that does not appear to be causal
      • all views held are tenuous
  7. Beware of behavioral biases
    • what is your rationality quotient?
      • epidemic rationality (degree to which beliefs map accurately to your world)
        • essential RQ ingredient
      • sidestepping biases = disposition + training + environment
  8. Know the difference between information and influence
    • prices go from source of info to source of influence
      • great investors don’t get sucked into vortex of influence
  9. Position Sizing
    • specifying a goal, identifying opportunity set, considering constraints
  10. Read and keep an open mind

The Mid Life Valley (A U-Bend)

Medical Innovation Manifesto

The FDA is broken. One can look at the saga of Sarepta Pharmateucials and its quest to get eteplirsen (a drug for Duchenne’s Muscular Dystrophy, a terrible disease) approved. In Innovation Breakdown: How the FDA and Wall Street Cripple Medical Advances, Joseph Gulfo lays out exactly how to fix the FDA.

Gulfo brought multiple companies into the public markets, shepherding them through messy financings, lawyers, and of course, an inept and inconsistent FDA. I’ll leave you to read his actual story (it’s well worth it), but he concludes his book with the Medical Innovation Manifesto, summarized below.

  1. Encourage leadership: director of the FDA oversees 10,000 full time employees, messy leadership leads to messy delays that delay innovation
  2. Clarify priorities of the FDA: prioritize health (does the drug get the job done or not?), rather than comparative standards like price which are not part of the FDA’s mission
  3. Define FDA review guidelines: Innovation Impact Guidelines
  4. Ombudsman should report to commissioner
  5. Refine Advisory Panel process: industry should pick a third of panel, relax conflict of interest rules, and always vote on approvability
  6. Enforce binding protocol process: concrete goalposts make everyone’s lives easier
  7. Hold FDA accountable when it fails to meet review guidelines: reimburse companies with less than $100M in annual revenue for delays that are the FDA’s fault
  8. End “major” and “minor” labels for deficiencies during review
  9. End transparency initiative: doesn’t actually do what it purports to do
  10. End interactive review initiative: see above
  11. End the innovation pathway: “breakthrough,” “priority” and “accelerate approval” designation lead to perverse incentives, “expedited” and “binding protocol” paths are enough
  12. Remove excessive regulatory requirements from growth companies: eliminate dark pools, shorting for certain stocks, Sarbanes Oxley, and enforce existing laws regarding stock manipulation
  13. Eliminate unnecessary shareholder lawsuits
  14. Eliminate programs that demand evidence based medicine
  15. Expand patients ability to pay for new products
  16. Establish Innovation Impact Index
  17. Replace redundant indications with Continuous Improvement Program
  18. Encourage feedback

These are good starting points. Hopefully Trump’s appointee is listening.

Universal Basic Income

Notes from a PRI x Lincoln Labs Event on Dec. 12, 2016 in San Francisco 

  • UBI has rare level of bipartisan support—similar to to earned income tax credit
    • previous experiments derailed for political purposes
      • headlines from 70s: people stopped working
        • subheadline: to go back to school, find a better job, take care of kids at home
      • YC has only done experiment for 2 mos., don’t expect research results from Oakland pilot
        • main group: $1,500/mo, other group, $500/mo
        • city of Oakland and state onboard, ensuring not interfering w/ other benefits like housing
          • everbody from poverty level to $90k in annual income
        • measuring how people spend time, economic well being, subjective happiness
  • don’t think in terms of robot replacement, look for lessons to learn in mean time
  • emerging market governments have begun to adopt cash transfers
    • lazy welfare recipient definitely a myth in EM, less evidence in developed countries
      • hard for every develpoment pgogram to give cash, people think there is no free lunch
        • most transfers conditional like Brazil’s Bolsa Familia
    • SS largest cash transfer program
    • GiveWell did $30M for 80,000 households in Kenya, comparing disbursing over 2 years vs. immediately
    • is it affordable? depends tax base, e.g. Scandinavia has larger tax base than other countries
      • $0.70 enough in Kenya, $80B enough to get everybody above poverty line in world
        • in comparison, we spenf $130B in foreign direct investment every year
  • looming structural unemployment, what is the growth industry for good, high paying jobs? [health care?]
    • can’t we invest in better education rather than UBI?
      • US government annual budget for AI research: $1B
    • moving to society with capital and time surpluses
  • existing system of welfare is BROKEN, UBI needs to replace it, can’t have both
    • hard for people to imagine different welfare state
      • UBI still dystopian, think of optics of Valley billionaires doling out checks to destitute
        • 15th century Europe: surplus of educated young men led to revolution
          • surplus of uneducated absorbed into domestic service
  • free trade has destroyed far more jobs than technology
    • estimated arrival dates for technology keeps getting pulled forward, AI attacking all domains
      • what if Nvidia is the reverse Tesla??
    • every tech innovation said to presage unemployment but always ends up augementing
      • why is this time different?
      • gay marriage didn’t break through until it was framed as equality issues, similar change needed for UBI
    • protectionism works to protect nascent industries, not declining ones

Tony Xu- Doordash

                                                      Tony Xu

  • thought he was going to be academic, loved applied math, found out program would take 12 years
  • had career in mom’s small business
    • delivered for small/large businesses, understood last mile problem
      • helped sell first 50 restaurants on DD himself
  • two reasons something can’t be free: market or lack of system
  • 10AM-2AM every day @ YC
    • creating driver, merchant, and customer facing products
  • find everybody’s superpower
    • used to think 10x person existed in each industry, but now thinks only applies in specific context
    • able to believe in someone before they believe in themselves
  • want for convenience didn’t start with smartphone, started in 195o
    • 20% of households were dual income, now 65%
  • biggest problem: how to increase quality as you increase quantity
    • natural state a decay function
    • technology quickly dithers away without maintenance
      • largest network of merchants by 5-6 times
  • as context changes, content changes, different needs
    • obsess over recruiting right people (still hires executives)
    • hard to predict how people react to conext change
  • think as far out as you’re alive = 20 years out
    • private drone delivery 2 years ago
      • 1st and last 10 feet problem with drones, so 7-10 years out rather than 2-3 years
  • culture defined by incentives, best version of you
    • made all info free, scaring investors, empowers employees to make own decisions
      • managers have final say, ego can’t win, diversity most important
        • fear that Valley has become Wall Street in promoting non diverse thinking
    • foster entrepreneurship if people want to become entrepreneurs
  • Valley needs to better value person you are today, not what you read on TechCrunch
  • book recommendation: What I Wish I Knew When I Was 20

LAUNCH Scale Day

Tom Patterson- Tommy John

  • normal online brands usually only innovate on price, TJ innovated on product
  • $100 to make prototype, no experience, background in strategic selling
    • called Neiman, said “I’ll be in Dallas next week” was a lie
      • knew women didn’t know shit about men, sent product to all men in office
  • shifted to ecommerce in 2012 to get repeat customers, no wasteful promos
    • spent 6 mos. of ad budget for 1 month on Howard Stern, paid for itself in 36 hours
      • got exposed to very diverse audience
  • every man can relate to ad
  • expect to reach $100M in sales in 2018

Alex Few- How to get 100K users before Launch, Calm

  • apps that went viral, feeder into real product
    • low probability, high reward, allowing for many failed experiments
    • tap in zeitgeist (Trends, memes, news, be opportunistic)
    • generate ideas (gather inputs, meditate, brainstorm/probe)
    • execute (day or 2 to build, simple design)
  • all businesses have time of day, but need right timing

Jed Katz- Getting to Series B, Javelin Ventures

  • Series B crunch, investors want less risk, scalability
    • make decision easy for them before running out of cash
    • 12 month goals: brand, recruit, roadmap, metrics
  • antagonize board when they deliver value, access their brains
  • create a presentation for new recruits, realistic timeline
    • what KPIs attract the next investor?
      • can you prove the size of the opportunity?
    • no red flags, high growth rates,  investors reupping, little churn
  • goal of 1st meeting to get 2nd meeting

Pejman Nozad- Coming to America

  • went from being journalist to car washer, sold $8M worth of carpets
  • failed at selling duct tape to shop owners about to be bombed
  • billion dollar mistake: couldn’t come to agreement on lease with Facebook ($50k would have $1.5B)

Glenn Kelman- Top 10 Growth Moments, Redfin

  • first time real estate agents have been on my side, love company
    • 5% market share in major markets
  • adding features add to surface area to product
    • weary to make tiny decisions (black v. white font)
  • Microsoft invested one third of engineering resources into building better software factory, make it easier to build product
  • don’t let anyone trap you in a mistake, own the error
  • you are a work in progress; look back/forward five years
  • ask for the $$
  • get to the good part: job is to be an optimist
    • always felt that if he failed/gave up someone else would solve problem in $60B market
  • name a great brand built by someone rational, gave away $100M in savings because customers would have used product even without savings
  • real estate cyclical as are company valuations
  • take the red pill, tried automating before hiring real estate agents
  • 100 VCs passed before 1 took the bait
  • be nicer to people, forget the P/L sometimes
  • worried about 12 year backlog in IPOs

Josh Elman- 5 Types of Virality, Greylock Partners

  • word of mouth virality: easy to spell, describe, remember
  • incentivized reality: give some, get some
  • demonstration reality: Musical.ly, just by using the product people show it off
    • easy  mechanisms for sharing creations, evoke a “how did you do that” reaction
  • infectious reality: people will work to get those people on because product better for both
    • for Nextdoor, sent customer postcards instead of email
    • invitations are key, natural network effect
  • outbreak virality: FlappyBird
    • hard to initiate combo of word of mouth and demonstration

Stacy Ferrera- 3 Hacks for Hiring in Competitive Landscape, Forge

  • start with network, takes 1-2 hours to go through 500-1K connections
  • 5-15 min phone calls before 30 min interview
  • focus on on site experience
  • do hard work on site
  • transparent about equity and process



LAUNCH Scale Day 1

Jason Lemkin- How to Get to $1M-$10M in ARR

  • people have bought shit they didn’t need
  • try to get to $10M ASAP (<4 years after getting to $1M)
    • even $1M is a huge accomplishment
    • many stall out at $4-5M ARR w/o right VPs
      • recruit sequentially
        • VP of sales all about recruiting
        • VP Marketing = lead boost
          • most with this title do corporate marketing (pretty slides, etc.) not lead gen
        • VP Product needs to manage 10x-100x roadmap
          • great product heads usually come only from Salesforce, Facebook, Google, etc.
  • no one feels like growing too slowly in SaaS
    • don’t enter segment where you have no customers (will be tempted to)
    • head of customer success team drives up NPS
      • guard your mini brand (if you’re doing it right, people will know you in your niche)
        • need to be more professional with 100+ customers

Cyan Bannister- Founders Fund

  • personality of company takes shape with or without you
  • company’s mission can be constantly re-evaluated
  • don’t fire someone on the fly
    • get ahead of employees’ needs, surprise and delight
  • metric driven milestone/roadmap
  • Simon Fuller: Valley doesn’t understand psychology and human motivation
    • what does the data actually mean?
  • too many reports = death
  • can’t copy paste one company’s template, pick and choose instead
    • Salesforce has excellent culture
    • with 1099 based business, need to be more company centric (where we’re all headed)
      • 6 year option grants keep people locked in

David Sacks- Zenefits

  • Thiel made ultimate contrarian move with Trump endorsement
    • contrarians may not be right even half the time, but when they are makes up for all the misses
  • hard to bring polarization into company
  • Facebook first to use .edu growth hack
    • consumerization of the enterprise
  • open sources licenses fix response to compliance
    • proactive with engaging in regulators
      • don’t defy, deny, and double down, produce of PR industry that benefits from damage control
        • acknowledge, fix, and repeat, as all about transparency
  • PayPal first in viral products (give and get, need email, etc.)
  • new product is app store for HR
  • reupped past investors at lower (!) valuations
    • employee ownersho[ doubled since Sacks took over
    • sales cannot be in charge of own analytics

Technology of Tomorrow

31 October 2016


  • how to find a technical cofounder: go to a nerd meetup, Galvanize is a good one
  • what is the next wave: AI, robots, and data science were this wave, have to anticipate what’s next
    • first movers in industry get the biggest exists (Nest in IoT, Oculus in VR)


  • high switching costs between processes (gaming v. watching movies v. something else)
    • takes time to get into experience (10 min, so experience needs to be at least 2o min.)
      • biggest opportunities in BLM for AR
    • airlines will be severely affected, why fly when you can feel like you’re there?
  • social aspect of women shopping exists w/ Taobao
  • AR for communication (tutoring)/dating needs to be presented in right way


  • most AI just saves stuff in database and refers back to it (>50% of apps)
  • investors have herd mentality
  • journalism analyzes public companies, human in the loop needed in most advanced AI
    • Facebook fucked up by allowing shitty sources


  • impossible to get cleantech funding now
    • Chinese markets sometimes flip side of US, funding cleantech/media like crazy, while SaaS still emerging


  • Form Labs makes materials based on lasers rather than extrusions, better resolution and still high end
  • print out batches of specialized cells
  • healing circuits: self repairing, check out DARPA


  • industry makes things that are otherwise not possible


  • used to not be able to power them inside body, now they create electric field around body, can steer nanofish
    • nanobot engineer skeptical about circulating potential, body very good at detecting foreign objects (liver, etc. are all elaborate filters)
      • more excited about semiautonomous bots with sensors
  • smart dust interest from big semiconductor space (Texas, ADI) who want to get out of sensor business due to margins constantly being pushed down
    • $870k prototype not including materials, with materials multimillion dollar endeavor
      • very hard to have proprietary tech.
    • can we afford to have everyone making massive margins?
    • governments and big corporations can afford to make long term investments, startups can take advatnage of off the shelf technology
      • paradigm has shifted to some extent at the moment
  • hard for consumers to handle ingestible robots, depends on how big of a pain point health problem is
    • technology not enough for those on middle of adoption bell curve
      • #1 reason to do stuff: our peers (pre-suasion)
  • 5D nanostorage holds enough data for 14B years

Brain chips

  • help with neurological disorders like Alzheimer’s, help people like Stephen Hawking
  • eventually will be able to record dreams
    • expose brain to image, see which parts light up, ask person to think of image, same parts should light up
      • problem: infinite combinations possible, a car to one person might be a Honda v. a Ferrari to other
        • neural cartography
          • still don’t understand how unconscious mind works

DNA editing

  • Intelligenomics using it to create new drugs
  • Venter growing human organs inside of a pig
  • China creating designer babies on the low to send to Mars (ultimate A/B testing)