Idea Machines

In this episode I talk to Jason Crawford about his work on the history of progress, funding and incentivizing inventions, ideas behind their time, and more. Jason is the author of the Roots of Progress blog, where he focuses on telling the story of human progress in an amazingly accessible way.

Key Takeaways

• Funding *structures* are understudied as a progress-enabling mechanism
• *Why* inventions happen is not so straightforward as we might think
• Culture may matter more than we think for building the future and there are concrete things we can do to build a culture of progress

Links

Roots of Progress Posts

• Smallpox - The history of smallpox & the origins of vaccines
• Charting progress
• Six threads of technology
• Arsenic as a pesticide

Other Jason Appearances

• Palladium Podcast with Jason that touches on the philosophy behind progress studies

Random

• Anki and memorizing - Augmenting Long-term Memory
• Ideas behind their time - Ideas Behind Their Time - Marginal REVOLUTION
• Tyler Cowen talking about bricks - My Conversation with Mark Zuckerberg and Patrick Collison - Marginal REVOLUTION

In this episode I talk to Eli Velasquez about creating startup ecosystems, commercializing research, especially when it's not necessarily venture-backable, and how the US government thinks about startups.

Eli is the head of Venture Development at VentureWell - a non profit organization that funds and trains faculty and student innovators to create businesses. VentureWell helps run I-corps, which talked to Errol Arkilic about in Episode 15. Currently, Eli runs all over the world helping create fertile ground of startup ecosystems and in the past he's worked with intellectual property both in industry at Boeing and Academia at Texas Tech. Basically he's working on meta-meta innovation: creating new ways to make places where it's easier for people to create new things.

Major Takeaways

1. Too much government aid can turn companies into zombies because their customer becomes the grant-giver instead of money-paying customers.
2. At the end of the day ecosystems happen because people's mindsets change
3. On average bringing a technology to market on average takes more than five years.

Resources

Venturewell

In this episode I talk to Bill Janeway about previous eras of venture capital and startups, how bubbles drive innovation, the role of government in innovation. Bill describes himself as "theorist-practitioner": he did a PhD in Economics, was a successful venture capitalist in the 80's and 90's with the firm Warburg Pincus and is now an affiliated faculty member at Cambridge and the member of several boards.

Key Takeaways

1. Bubbles have arguably been the key enabler of infrastructure-heavy technology.

2. Venture capital may be structurally set up to only be useful for computing and biotech.

3. Most technology that venture capital invested in was subsidized at first by the government in one way or another.

Resources

Doing Capitalism in the Innovation Economy

VC: An American History

Wikipedia article on Bill

NYT Article on Fred Adler from 1981

Bill's Website

Bill on Twitter

In this episode I talk to Mark Hammond about how Deep Science Ventures works, why the linear commercialization model leaves a lot on the table, and the idea of venture-focused research. Mark is the founder of Deep Science Ventures, an organization with a fascinating model for launching science-based companies. Mark has many crisply articulated theses about holes in the current system by which research becomes useful innovations and what we might do to fill them.

Key Takeaways:

1. There are many places where innovation is slow and incremental because everybody is focused on individual pieces: batteries are a great example here.

2. The perception that deep/frontier/hard tech companies are riskier and take longer to provide returns may in fact be more grounded in popular perception than fact

3. The factors that make translational research so expensive may not be inherent but instead driven by administrative overhead and the fact that much of it is pointed in the wrong direction.

Resources

Deep Science Ventures

Mark on Twitter (@iammarkhammond)

Systematised 'quant' venture in the sciences.

LifeSciVC on biotech returns

Alexey Guzey is an independent researcher focusing on how to systemically increase the rate of biology discoveries and the idea that reviving the patronage system may be a way to do that. We spend most of our time talking about the project he's been working on for the past year but also touch on some of his thinking around connecting with people, which he's written about extensively.

Key Takeaways

1. Most people doing biology research are embedded in a system that incentivizes incremental consensus steps and divides researcher time

2. There are some institutions that stand at least partially outside of that system - Calico and Janelia being two examples

3. Maybe we should be supporting more crackpots

Resources

Alexey's Essay: Reviving Patronage and Revolutionary Industrial Research

Followup: How Life Sciences Actually Work: Findings of a Year-Long Investigation

Alexey on Twitter:@alexeyguzey

Alexey's Website

HHMI Janelia

Calico

Andrew York

Ronin Institute

Emergent Ventures

Phillip Gibbs - crackpots who turned out to be right

Cindy Wu and Denny Luan are the founders of experiment.com - a platform that allows anybody to request funding for a science project and anybody to fund them. It's fascinating because it stands completely outside of the grant funding and publication system that drives most science today. In this podcast we discuss how the current system prevents the creating of new fields, why science communication may be even more important that science funding, and new models for company governance.

Key Takeaways

• The incentives built into the grant system make it hard for new fields to emerge
• Arguably, changing how science is communicated might have the biggest impact on our knowledge creation system.
• The concept of ownership and governance of companies being two separate axes that need to be considered separately

Resources

Experiment.com

The Science of Science Funding

DIY biohackers trying to see infrared with vitamin A

Innocentive

Public benefit corporation

Purpose Trusts

Wellcome Trust/Foundation

Employee Owned Breweries

Topics

• Consolidation and risk aversion in science
• Hard to fund research outside of funding buckets
• Field politics
• Hard for younger scientists to get funding
• NIH budget stayed the same, proposals have doubled
• Government funds what's popular
• CERN is a consortium of companies doing funding
• Only real solution is disseminating knowledge
• DIY biohackers trying to see infrared with vitamin A
• Digging up dinosaurs
• No money to prepare dinosaur bones
• Incentives for science
• Brewery example of employee owned corporation
• New models for funding businesses
• Ownership and Governence Axes
• Making scientists stakeholders in
• Danger of masking philanthropy as investment and vice versa
• Would VCs ever fund something that's not purely for profit
• New Company structures

In this episode I talk to Errol Arkilic about different systems involved in turning research into companies.

Errol has been helping research make the jump from the lab to the market for more than fifteen years: he was a program manager at the National Science Foundation or NSF, Small Business Innovation Research or SBIR program, where he awarded grants to hundreds of companies commercializing research. He started the NSF Innovation Corps, a program that gives researchers the tools they need to make the transition to running a successful business. Currently he is a partner at M34 capital where he focuses exclusively on projects that are being spun out of labs. Seeing the often rocky tech transition from so many sides has given him a nuanced view of the whole system.

• While there are some best practices around commercializing research, like business model canvases, many pieces like assembling a team and finding complementary technologies are still completely bespoke.
• The commercial value of research is a tricky thing. Some is valuable, but not quite valuable enough to form an organization around. Other research could be incredibly valuable if the world were in a slightly different state. Different approaches are needed in each situation.
• The mental model of MIST vs TIMS - market in search of technology and technology in search of market.

Links

M34 Capital

The SBIR Program

Business Model Canvases

Errol on How the NSF Works

Pasteur's Quadrant

NSF Innovation Corps

Topics

What is the pathway to commercialization

How do you have an iterative process when people don't know what they want

What do the best researchers do to pull out core problems to work on?

How do you address the tension of people wanting to apply their hammers?

What are examples of people who have applied very specific technologies?

How do you assemble a team around a technology?

How do you systemitize assembling teams?

How do you systemitize finding technologies that can plug a technological hole?

What do you think about patents?

Patents, trade screts,

Technology that isn't venture fundable

Valuable ideas that aren't valuable enough to pursue

Systemitizing finding whether value could be harvested

Where is the role of SBIRs in today's world

SBIR decision making process

Lengendary SBIR successes

Push vs. Pull out of lab

How do you find MIST projects

Are there labs in unintuitive programs

Next steps outside of local ecosystems?

Does any new innovation need a champion?

What should people be thinking about that they're not?

TISM vs MIST

In this conversation Sam Arbesman and I talk about unlocking cross-disciplinary innovations, long term organizations, combinatorial creativity and much more. As you might expect from someone with Generalist Thinking as a main area of interest, Sam has out-of-the-box insights in a ton of domains and he's amazing at capturing them in tight concepts like "knowledge mining" and "jargon barriers."

By day Sam is the Scientist in Residence at Lux Capital. Don't cite me on it but I think he may be the only person with that job title in the world. In the past he's done research in complexity science and history and the two of them combined, written books, and worked in non profits.

1. The concept of knowledge mining - recombining existing knowledge to create new knowledge.
2. Unintuitively, Video games may secretly be some of the most powerful cross-disciplinary research labs.
3. There are tactics you can use to generate cross-disciplinary creativity by cultivating a bit of randomness in your life.

T-Shaped Individuals

Sam on Twitter

Sam's Website

Small World Networks

Complexity

Undiscovered Public Knowledge (and a 10-year update)

Spore

Kongō Gumi - the 1400 year company

The Red Queen Hypothesis

Other content from Sam:

https://fs.blog/samuel-arbesman/

https://25iq.com/2016/03/12/richard-feynman-and-charlie-munger-expert-generalists/

Topics

Favorite examples of combinations of ideas via generalists

Ref: Small world networks paper

T shaped individuals

Attempts towards systemic cross-discipline idea sharing

Don Swanson - undiscovered public knowledge

Jargon Barriers

Jefferson West Uwash - topographical map of fields

Combinatorial creativity

Systems for increasing the rewards for broad thinking vs. specialized thinking

Need to define complexity science

Computer games as a place that rewards generalist research

Meta portfolio for generalist institution

Self-sustaining insitutions and criteria for them

Reinventing selves

Or provide something people always want

Japanese construction company that lasted 1500 years

IBM original machines

The Red Queen Hypothesis wrt Organizations

Model that you need massive innovations to sustain growth (look up professor)

Does the VC funding research paradigm constrain what can exist?

Wired magazine researcher - "everyone loves the big idea that changes the world, but what about the ones that make a difference?"

The importance of different approaches to making things exist

How do you know if small ideas and tweaks in complex systems have intended effects?

Promoting randomness and optionality

What are tactics for increasing randomness and optionality?

Randomly reminding about books

Go to crazy different conferences

In this episode I speak to Matt Clifford about talent investing, how big long term projects can start small, and financial innovations.

Matt is the CEO and co-founder of Entrepreneur First. Entrepreneur First, abbreviated as EF, is a fascinating system. It starts with cohorts of around fifty to a hundred ambitious, talented people who want to start companies but might not even have an idea to build around.

1. The mental model of predictable vs. unpredictable value.
2. The idea that hypothesis testing speed predicts success even in projects where you won't see real results any time soon.
3. The idea of money as a commodity that fuels innovations

Background on EF (context for some of the podcast)

EF then helps cohort members pair up into teams and get companies off the ground. Matt and Alice Bentinck started EF in 2011 and the history is kind of a crazy story: it started as a non-profit and now has raised a massive fund from LPs. One of the highlights in the story that really put EF on the map was a company named Magic Pony that sold to Twitter for an unconfirmed 150 million dollars eighteen months after starting at EF. There are links to Matt talking more about both the structure of EF and EF's history in the show notes. EF is a fascinating innovation system because it challenges many ideas that have basically become gospel in the startup world - everything from "if someone isn't willing to start a company in a garage with no income they don't have what it takes" to "only founding teams with a long working relationship can succeed."

Matt on Twitter (@matthewclifford)

Matt's weekly newsletter

EF on Wikipedia

Magic Pony exit referenced in podcast

Matt speaking at Startup Grind about how EF works

Ideas

Capital as a resource like any other

Adverse selection

The best CEO of a deep tech business often doesn't know the best CTO of that business

Predictable value vs Unpredictable value

Predictable market does not necessarily mean existing markets

Basically logic-able innovations

Job as founder is to lay out 18 month roadmaps

Think of VC as a financial product

Providing optionality to the founder

Income sharing, with optionality

The power of finance innovations

Misalignment of incentive between VCs and entrepreneurs because VCs have a portfolio

In this episode I talk to Evan Miyazono about tackling metaresearch questions, how novel physical phenomena go from "oh that's cool" to devices that harness cutting edge physics, and how we could better incentivize the creators of innovations where traditionally it's hard to capture value, like open-source software and early-stage research.

Evan is a research scientist at Protocol Labs where he helps lead their research efforts - coordinating researchers both inside and outside the company. Protocol labs is best known for Filecoin: a blockchain application for distributed storage. At the same time they also have a much larger mission that we get into in the podcast. Before joining Protocol Labs, Evan did his PhD at Caltech where he worked on turning crazy physics into practical devices for cryptography.

• There might be ways to demystify both intuition and "big H Hard" research research in order to improve our systems for breakthrough discoveries. It's still super speculative but worth thinking about.
• Observations about physical phenomena and the world are at the core of many innovations, but the most of the process is driven from the top down by the problem, rather than bottom-up by the solution. On top of that, the process of solving the problem can actually feed back and increase our understanding of the underlying phenomena.
• Finally, there might also be new legal structures we could put in place to encourage more open-source development and fundamental research by allowing people to access more of the value they create in those activities.

Protocol Labs

Evan on Twitter

A quick talk on Protocol Labs research

Metascience

Cloud Seeding - From the abstract: "The intent of glaciogenic seeding of orographic clouds is to introduce aerosol into a cloud to alter the natural development of cloud particles and enhance wintertime precipitation in a targeted region. ... Despite numerous experiments spanning several decades, no direct observations of this process exist."

SourceCred - a tool to help open source contributors capture the value of their contributions.

Evan on Google Scholar if you want to go really deep. Try saying "Coupling of erbium dopants to yttrium orthosilicate photonic crystal cavities for on-chip optical quantum memories" three times fast.