Suggest edit — 19-05-01 Research Failure Stories

Title

Name

Note

---
title: "19-05-01 Research Failure Stories"
visibility: public
---

# 19-05-01 Research Failure Stories

Category: [[idea-lists-upon-request|Idea Lists (Upon Request)]]

[Read the original document](https://docs.google.com/document/d/1nhiqv_CEYkwYZh5Lsl8SXZWGpfGAnRYItgfq4zBX3A4/edit?usp=drivesdk&sa=D&ust=1596495076795000&usg=AOvVaw3fCtShmTthuNV2_UEskPff)

---

Aspects of a valuable ‘failure in research’ event:
1. 1 on 1 personal failure stories - partner up, talking to your person about an example where you took on a research project that failed in some sense
2. Speakers with quality stories of struggling in research share the story with an audience.
3. Interview / conversation, where a junior scientist talks to a senior scientist (or something like it)
4. Round of insights, where there’s a sweep through everyone’s most valueable
5. Round of questions, questions people are most curious to have answered by the event
6. Contrarian truth, 
7. In 1 on 1s, tell story of the failure that had the most positive impact        
   1. Ex., from which the most was learned
   2. Ex., from which you bonded with the person who would become a great influence
8. In 1 on 1s, tell story of the failure that had the most negative impact        
   1. Ex., from which your reputation never recovered
   2. Ex., which you lost years of your life to
9. Split into small groups, each with a person with many stories worth telling. Engage in that context with some interview / prompts / free form conversation, and then switch to another set of small groups.
10. My vulnerable truth about research is…
11. How much do you work on research you believe in, relative to research that advances career / status / income?

Do we need a re-frame from failure?
What forms of failure exist?
1. An idea you expected to work did not work.
2. A system you thought could be implemented could not be implemented.
3. A project you thought was accomplishable in a given span of time actually took orders of magnitude more time.
4. Nothing was learned.

Why is this interesting?
1. Ever present example of violated intuition. ‘This should have worked but didn’t’ points to a breakdown in our model of how a system works, and guarantees an update in our beliefs if the reason our expectation was violated can be discovered.
2. This is information that’s not communicated, due to shame, reputation, selection bias and more. It’s taboo and rare and so full of richness.

1. 30-40 minutes of warm 1 on 1, opening with vulnerable stories / uncomfortable truths.
   1. Want a bunch of high quality people to talk - limit it to 10-20m after.
2. 4 stories, 5 minutes each.
3. 20m of unstructured conversations.
4. Circles of 5 with uncomfortable truths. Can also do introductions.
   1. 25-40m. Writing everything on a board?
   2. Each group presents 2 truths.
5. End the event, and let people chat.
6. Talk into the night.

Invitations:
Turns out we just need 4 high quality stories. Everything else is group work or 1 on 1.

Who to invite?

Jeremy:
1. Ilya Sutskever
2. Samy Bengio
3. Ian Goodfellow
4. Ray Kurzweil
5. Stuart Russell
6. Andrew Ng
7. Greg Brockman
Laura:
1. Saul Villeda
2. Nathaniel David
3. Cynthia Kenyon
4. Jasper Rine
5. Daphne Koller
6. Vijay Pande
7. Andy Dillin

Setup logistics:
* Email invitations
   * Paperless Press
* Overall invite list
   * Composition?
      * Fraction of machine learners / CS researchers vs. Biology / Longevity researchers vs other (ex., physicists)
         * ⅓ ⅓ ⅓ ?
   * Speakers
      * Invite 3 each - LD + JN
* Fb Event?
* Event Description / Motivation
   * Come hear and describe the most uncomfortable and contrarian truths in your field and hear inside stories of research failures and victories from a handful of phenomenal scientists. You’ll all be extremely interesting people.

Order of events:
1. June 25th. Laura and Jeremy sent out 3 invitations each. In the face of a rejection, send out a new invite.
2. June ~27th. Send out invitations to ~25 people, ⅓ ⅓ ⅓ ML / Bio / other.
   1. Send out an invitation that gets a response / commitment quickly, so we can invite more people in the face of conflicts / rejections.
3. July 7th - Final logistics setup / low level decisions about exactly which prompts to use and when to follow when.
4. July 8th [Day Of] - Food/Snacks - LD
5. July 8th [Day Of] - Manning Door / Greeter - LD / JN / NS

---

*Source: [Original Google Doc](https://docs.google.com/document/d/1nhiqv_CEYkwYZh5Lsl8SXZWGpfGAnRYItgfq4zBX3A4/edit?usp=drivesdk&sa=D&ust=1596495076795000&usg=AOvVaw3fCtShmTthuNV2_UEskPff)*