Graduate students are the labor in academic science and engineering

People giving out grants try to derisk them as much as possible

Manufacturing is important and often ignored

The word ‘novel’ is used as an idea bludgeon in academia

All the incentives in research drive towards conservatism

Peer review doesn’t catch duplications

Bell Labs is one of the few examples of an innovation org that was aligned with its money factory

A new professor needs about $1.5m to get going

Academic culture prizes individual recognition

Younger professors have to spend more time getting money

Academics are incentivized to do peer-reviewed publishable work

Things that are not paper-worthy enough for academia and not product-focused enough for a startup

Pay-it-forward tithing could resolve the tension between the potential public-good and potential profit in research.

The idea that academia is for crazy exploration and companies are for low risk exploration is bullshit

Three myths about scientific peer review

A PI needs to provide evidence that they have access to a lab facility

Different axes of where research gets done in the US

What was the real incentive for people to work on the DARPA Grand Challenge?

It is easier to have a multipoint mutation in technology than in evolution

NS Notes on Officers Techniques

Necrobotics illustrates how academia forces engineering

The boundaries between industry government and academia are fuzzy

Startups and Academia are bad at addressing structural constraints

The Architected Innov􏰂ation Institute - Engineering Scientific Revolutions

There are very few senior individual contributors in academia

Is it more impactful to find and train undiscovered diamonds or to unlock well-positioned but constrained researchers?

Wolfgang Merkt conversation 12 Oct 2021

An institution riffing on the ARPA Model could enable paradigm shifts that are not happening

Academic administrators and professors are playing very different games

An Institute for Creative Technology - Riffing on ARPA

Funding structures lead to several layers of indirection on what even gets proposed

Is it possible that we’re actually getting less research because we have more researchers?

Running list of relevant science news?

# §Academia Constraints
- [[Academia is the game where you gain status by getting attention for new knowledge]]

- [[Universities have multiple conflicting missions]]
- [[Universities have been pushed more towards applied research since the end of WWII]]
- [[Academic credit is a first-to-the post system]]
-  [[Graduate students are the labor in academic science and engineering]]
- [[People giving out grants try to derisk them as much as possible]]
- [[Professors do three or more jobs]]
- [[Successful professors are not necessarily the best researchers]]
- [[University professorships are the highest status jobs in academia]]
- [[National labs and other government research centers are a research alternative to universities]]
- [[The game in national labs is roughly the same as academia but with different constraints]]
- [[In academia more grant money leads to more power]]
- [[Academic research is often just the concept half of creating a heuretic]]
- [[Professors are constantly doing the calculation of whether a grant is worth the time to apply for it]]
- [[Most grant money is tagged so unrestricted money is like gold for professors]]
- [[Universities can take more than half of grant money as overhead]]
- [[A PI needs to provide evidence that they have access to a lab facility]]
- [[Physical experiments are expensive]]
- [[Younger professors have to spend more time getting money]]
- [[The government grant process depends on politics and committees]]
- [[Any technology that comes out of a university needs to go through their tech transfer office]]
- [[Months of grant writing is one of the things that makes people think twice about academia]]
- [[Academia incentivizes novelty, not focus]]
- [[Academia is not a good environment for systems engineering]]
- [[Academics are incentivized to do peer-reviewed publishable work]]
- [[Going outside of the approved process can be a black mark against people in academia]]
- [[People proposing work that they have already done is a sign that feedback loops are broken]]
- [[Necrobotics illustrates how academia forces engineering]]
- [[Labs implicitly judge research through the sense of dollars per paper]]

### Related
- [[§Startup Constraints]]
- [[§Corporate R&D Constraints]]
- [[§Philanthropy and Non-Profit Constraints]]
- [[Market Failures in Science]]

[Web URL for this note](http://notes.benjaminreinhardt.com/§Academia_Constraints)

[Comment on this note](http://via.hypothes.is/http://notes.benjaminreinhardt.com/§Academia_Constraints)






How does money work in PARPA?

“The System” of AT&T

AI research is an exception to the decline of corporate R+D

National labs and other government research centers are a research alternative to universities

Problems that require system-level solutions often require massive teams

Why try to emulate DARPA and not Bell Labs

R+D labs require companies that can win based on Technology

# §Corporate R+D Constraints
 Follow innovation through bell labs
	* Run into problem with “the system” - can’t fit enough calls into a single cable. More cables are expensive. Want people to make more, better, calls. Problems constantly coming down from the top and 
	* 
* Follow Innovation through GoFaAmSoft
* Politics 
	- [[Internal incentive structures within a large organization work against high quality R&D labs]]
	* [[R&D orgs are always on the side of an org chart - The lab is always separated from the mothership]]
	* [[Rewards in a big company operate on one timescale and R&D happens on another timescale]]
* The most valuable research for a company tends to be systems-level improvements. But [[Problems that require system-level solutions often require massive teams]] while [[Research tends to work best when it can be done by small teams]].
* Risk aversion
	* [[Asymmetric career risk]] 
	* [[Asymmetric career risk impacts corporate R+D]] 
* Incentives
	* [[The incentives inherent to corporate R&D mean that researchers motivated by money leave and those who are motivated by glory are not the best in the world]]
	* [[At the organizational level corporate R&D labs either cause brain drain or don’t have the best people]] 
* [[Corporate R+D orgs are expensive]]
* [[R+D orgs are answerable to shareholders both in scope and timescale]]
* [[Companies want R+D to be off of their balance sheets]]
* [[The siloization that tends to happen in large corporations hamstrings applicability and serendipity]]
* [[Big companies cannot start small things]] 
* [[Expensive research needs to address an existential threat eventually at an organizational level to maintain support]]
* [[There are several macro-economic analyses of both why and whether there has been a decline in corporate R+D]]
* [[Corporate R+D orgs experience a tension between Research and Development]]
* [[Corporate R+D orgs have a lot of momentum]]
* [[Healthy corporate R+D requires a trifecta of conditions]]
* Trends
[[Corporate Innovation has four main options: partner, R+D, corporate venture capital, and venture building]]
<draw 2x2 box>
	* Externalize R&D through startups and universities 
	* Internal startups via Red Box at Swisscomm or Area120 at Google
	* “Innovation Arms”
	* Grants
## Related
- [[§Startup Constraints]]

                            

[Web URL for this note](http://notes.benjaminreinhardt.com/§Corporate_R+D_Constraints)

[Comment on this note](http://via.hypothes.is/http://notes.benjaminreinhardt.com/§Corporate_R+D_Constraints)




§Academia Constraints

Decoupling from market discipline is like cave diving

Current value capture mechanisms are crude

When should an idea that smells like research be a startup?

§Corporate R+D Constraints

Startups are good at point changes

Systemic decay theory of stagnation

Ideally corporate R+D enables work on a general purpose technology before it’s specialized

There are many tight feedback loops built into the ARPA model

One function of a private ARPA is to fill in the gap once filled by corporate R+D

There are many important innovations that are just bad investments

Startup Constraints - essay

The market in for-profit organizations is fairly efficient

Profit-maximizing organizational structures can hamstring technologies whose impact depends on them

PARPA could walk the line between profit and non profit

Startups  need to spend effort on sales, marketing, and productization

# §Startup Constraints
- [[Startups need to make a lot of money or have a lot of users]]

- [[Startups need to conform to investor timescales]]
- [[Venture capital doesn’t work for things that only realize their value on long timescales]]
- [[Startups need an exit strategy]]
- [[Startups need a great story]]
- [[Startups need to produce a scalable product]] 
- [[Startups need to break development into correctly sized stages]]
- [[Startups need specific growth curves]]
- [[If a project has massive possible upside and low enough risk it is venture fundable]]
- [[Scalability and venture fundability happens when you turn a process innovation into a tool innovation]]
- [[Startups are good at point changes]]

- [[Startup Constraints - essay]]

[Web URL for this note](http://notes.benjaminreinhardt.com/§Startup_Constraints)

[Comment on this note](http://via.hypothes.is/http://notes.benjaminreinhardt.com/§Startup_Constraints)






Innovation orgs need to be aligned with their money factory

Bell Labs

Alignment requires existential threats

The incentives inherent to corporate R&D mean that researchers motivated by money leave and those who are motivated by glory are not the best in the world

Bell Labs was only profitable in conjunction with AT+T

Why was Bell Labs Different?

§Bell Labs Model

Labs Document Outline

# Bell Labs is one of the few examples of an innovation org that was aligned with its money factory
 [[Innovation orgs need to be aligned with their money factory]] and arguably one of the reasons behind [[Bell Labs]]’s outlier success was that it managed to have a broad scope and long time scale that were both *still aligned* with AT&T.

At the end of the day AT&T’s goal was to maintain its monopoly and increase its business. In order to do that, it needed to enable people to make more, better calls. 

[[“The System” of AT&T]]  had such an expansive scope that it actually benefitted from the whole spectrum of activity happening at Bell Labs. Everything from information theory to novel materials directly fed back into AT&Ts core business. More cables are expensive so information theory actually helps people make more calls. The faster you can switch, the more calls you can route so semiconductor switchers *directly* help with the goals. 

Additionally, [[Bell Labs and AT&T had timescale alignment]] thanks at least in part to AT&T’s monopoly. The monopoly enabled Bell Labs to plan on multi-decade timescales because they didn’t have to worry about a competitor coming along tomorrow. 

AT&T maintained their monopoly only with tacit approval from the government. This approval rested heavily on Bell Labs’ output - both continuous improvements to A&+Ts offerings and open licenses from the lab. [[Regulation drove many of Bell Labs’ actions]]. 

Between the cost reductions and monopoly maintenance, Bell Labs was actually addressing existential risks to AT&T. [[Alignment requires existential threats]]. Even R&D labs that are building new product lines within the scope of a company’s actives rarely address existential threats to that extent.

This alignment in both timescale and focus-scope between an innovation org an its money factory. See [[§Startup Constraints]], [[§Corporate R+D Constraints]], [[§Philanthropy and Non-Profit Constraints]], and [[§Academia Constraints]] for all the lovely ways the misalignment can happen.

## Related
- [[gertnerIdeaFactoryBell2012]]



[Web URL for this note](http://notes.benjaminreinhardt.com/Bell_Labs_is_one_of_the_few_examples_of_an_innovation_org_that_was_aligned_with_its_money_factory)

[Comment on this note](http://via.hypothes.is/http://notes.benjaminreinhardt.com/Bell_Labs_is_one_of_the_few_examples_of_an_innovation_org_that_was_aligned_with_its_money_factory)




There is a significant class of innovations that would create drastically less value for the world if their value had been captured by their creators

It is hard to capture value from research

Most tech transfer offices are not profitable

Value capture does not always need to happen through IP

Creations whose impact depends on being public generate uncertain value through not-particularly modular combination with many other things and have a large combinatorial surface area

It is nebulous a priori which research should produce public goods vs private goods

A normal for-profit structure would hamstring an ARPA-riff

Better value capture mechanisms

Profit is important for both practical and ideological reasons

PARPAs ideal legal structure comprises both a nonprofit and a for profit

There are many questions and issues with the idea of pay-it-forward tithing

ICT Distillation

ICT  Reviewers

ICT Outline

Omniscience is not enough

# Current value capture mechanisms are crude
The direct ways to capture the value from research are to patent[^1] it or build a product around it (and leverage trade secrets.) Other possibilities are more indirect and ‘credit based’: consulting, speaking fees, or leveraging reputation to get more lucrative positions later.

### Patents
Patents require an expensive application that can take years to process, lawsuits to enforce, and effort to monitor for violation, and lead to an extremely illiquid market. On top of that, there is tons of valuable work that just isn’t patentable.

Capturing value with a patent is often inefficient. In an ideal situation, someone building a product that incorporates the patented process comes to you, the two of you negotiate a licensing deal, and they send you a check every year. The world is not ideal. Often, you need to *find*people who are violating your patent and threaten to sue them if they don’t pay up. Standard lawyer advice when building something is to *not* look for prior art because it decreases your liability in case you happen to reinvent something. Even if someone *does* come to license your patent, the negotiation over licensing terms can be brutal - there are no standard terms and especially for a startup it can create a drag chute on a product’s profitability that makes an already hard job even harder. Once the licensing agreement is signed, you still need to make sure that the company is paying you the royalties you’re due and sue if they’re skimping you. [[Many inventors have spent a lot of their lives in court over patents and sometimes wind up dying in the gutter in spite of that]]. 

It’s worth noting two historical facts about patents that shed some light on their limitations.  [[Patents have become more BS over time]]. On top of all of that, patents weren’t even created as a value-capture tool for discovery or invention.  Instead, patents started as a way for monarchs to increase their income without going through Parliament by granting temporary monopolies.[^2] Of course, the modern patent system is explicitly meant to incentivize innovation so this might be nothing more than a cool anecdote. 

Additionally, [[Patents inherently prevent a piece of IP from being combined with other knowledge]].  The combination-preventing nature of patents is particularly problematic if you want to use them to try to fund work to address gaps in work on processes and systems (which naturally require technological combinations). <[[In the past, corporate labs filled a particular niche in the innovation ecosystem]]>

### Products
The first reason products are a crude way to capture value from research is that most modern products are an amalgamation of different technologies and designs. So while a piece of research may have vastly improved a component, that doesn’t necessarily translate to a vastly superior product. 

Second, products are inherently coupled to companies, so the success of a product (from a value capture standpoint) is inherently tied to the success of a company. This coupling means that there are many other factors that affect the product’s success besides the product itself (let alone the research-based piece of it). Markets, sales channels, margins, fundraising, perception, etc. all make or break companies independent of their products.  [[The best technologies don’t always win]] is an adage for a reason.

“But unique new technology creates a moat that will make the company successful!” Jerry Neuman’s article,[[Productive Uncertainty]] argues convincingly against pure-technology with the idea of “excess value” an invention would produce for a startup. 
`If a moat exists prior to the startup being founded (say, a patent) then, absent uncertainty, this patent could be sold for at least as much as the startup could garner from it.` 

More broadly, value capture through products force innovations to be subject to either [[§Startup Constraints]] or [[§Corporate R&D Constraints]]. None of this is to say that products are a *bad* way to capture value, only that they are subject to many constraints and contingent requirements that mean that the correspondence between valuable technologies and technologies whose value can be captured by a product is weak. 

### Related
- [[Michael nielsen re david d value of creative work twitter]]
- [[Patents held back aviation]]

[^1]: I’m going to just talk about patents but the same arguments apply to both
[^2]:Anton Howes <[[Age of Invention- How to Build a State]] >



[Web URL for this note](http://notes.benjaminreinhardt.com/Current_value_capture_mechanisms_are_crude)

[Comment on this note](http://via.hypothes.is/http://notes.benjaminreinhardt.com/Current_value_capture_mechanisms_are_crude)




Phenomena-based cycles

Systems must take performance hits to get out of local optima

§ARPA model

DARPA has a mixed record on transitioning

Program design is like creating the lego instruction book

Someone needs to buy manufactured technology eventually in a monetized economy

The first step of roadmapping is frontier mapping

DARPA works where markets dont

Many of ARPAs outlier results were things that nobody was asking for

Mark McGranaghan conversation 19 Jun 2020

What is a market failure?

Bottlenecks Nanotech Talk 2021

Conversation with Ryan 22 Jun 2020

# Decoupling from market discipline is like cave diving
*Note that I have never gone cave diving. I am not that cool.*

Following the aphorism “make something people want” is a good heuristic for adhering to market discipline even before making money. Paradoxically, there are many counter-aphorisms like “if I asked people what they wanted they would have said a ‘a faster horse.’” The resolution to this paradox is “Make something people want at some point in the future.” In the long run, obviously you want to make something people want.   [[Someone needs to buy manufactured technology eventually in a monetized economy]].  However, the whole time you’re *not* making something people want, you are burning resources. The question then becomes “what is the thing people want in the future and can the technology or idea survive long enough to get there?” 

Cave diving feels like a great analogy for this process. While you’re going through the cave, you’re using oxygen and if you run out before you reach the next air pocket, you’re dead. Working on a product you can always test against the market is like normal diving - you can always come up from air and maybe you need to get the timing right so you don’t get the bends. 

Perhaps it’s torturing the analogy but there are some interesting concepts to pull back from cave diving. The longer the path, the better you need to plan. If it’s a short hop and you can see where you’re aiming you don’t need to plan as much. If you know that there are lots of possible air pockets, you have the freedom to explore more. Similarly if the technology has lots of market ‘exit ramps’ you can do more exploration. Software is uniquely repurposable which might offer an explanation for why agile product development makes sense for software but perhaps not in other situations.  Having a precise idea of where you’re going is very important. Even if you don’t know the exact way, you need to keep your bearings towards where you’re going. 

### Related
- [[DARPA works where markets dont]]
- [[Many of ARPAs outlier results were things that nobody was asking for]]
- [[Market failures are actually failures of the existing institutions to allow markets to work]]
- [[Market failures are coupled to externalities]]
- [[Applicable markets have band gaps]]
- [[§Startup Constraints]]
- [[DARPA has a mixed record on transitioning]]
- [[Advanced Market Commitment Thread]]
- [[odlyzkoDeclineUnfetteredResearch1995]] 
	- People can have long timescales if output is clear
- [[Is it possible that base rate growth in the market is actually a bad thing for innovation?]]



[Web URL for this note](http://notes.benjaminreinhardt.com/Decoupling_from_market_discipline_is_like_cave_diving)

[Comment on this note](http://via.hypothes.is/http://notes.benjaminreinhardt.com/Decoupling_from_market_discipline_is_like_cave_diving)




How Tech Loses Out over at Companies, Countries and Continents

In the past, corporate labs filled a particular niche in the innovation ecosystem

Could FROs fill the gap left by corporate R+D?

Magic Leap is a good case study of a potentially amazing technology killed by startup constraints

25 Oct 2020

General-purpose technologies need pitchfork development

The NSF could partner with philanthropy to unlock new general purpose technology

# Ideally corporate R+D enables work on a general purpose technology before it’s specialized 
# Ideally corporate R+D enables work on a general purpose technology before it’s specialized 

Specialization has a complicated relationship with new technologies. New technology will never[^1] be as good as good as old technology along every dimension. This failure of Pareto-superiority is why [[Frontier technologies need to start in niche markets where they are especially valuable]]. It’s rare that a technology can do well in a niche market without *some* specialization. If you imagine the straightest path that a technology can take to its most powerful or general use[^2], the specialization work to fit into niches can require larger or smaller diversions from this path. These diversions take the form of everything from technology development, to marketing, to building sales channels for a company. Going from niche to niche is essential, but you can also get stuck in a niche or jump onto a different development path altogether. [[Elon is the master of charting paths between niches]]. More time and resources ([[Slack - concept]]) can enable you to pick an “optimal” sequence of niches or search around for yet-unknown niches instead of being forced to hop into the nearest or most obvious ones. In other words, if [[Decoupling from market discipline is like cave diving]], corporate R&D can act like extra oxygen tanks. 

I don’t know how often I’ve seen a presentation from a grad student or professor that abstractly goes like “we’ve invented this incredible technology that could potentially do <amazing thing> however it will take a lot of work to get there, so we’re starting a company to build a product to do <kind of pathetic thing that will still require a lot of specialization and company-building> in <domain totally unrelated to <amazing thing>>.  While I’m usually sure it’s doomed to failure, it’s unfortunately the best move given history and [[§Startup Constraints]]. First, many successful startups *do* look like they started in a tiny scoffable niche  - Tesla (weird rich people), NVIDIA (gamers), PayPal (Beanie Babies on eBay?), Apple (hobbyists), Twitch (life live-streaming), Amazon (books) … the list goes on. However, each of these weird niches didn’t require huge diversions from the critical path.[^3] Second, ‘start in a niche’ is folk startup wisdom *for a reason*. History is littered with the corpses of startups that were going to be the next platform or [[General purpose technologies]] as soon as they launched. General Magic, NeXT, Quibi, Atrium, Rethink Robotics, and of course Magic Leap [[Magic Leap is a good case study of a potentially amazing technology killed by startup constraints]]. Many others that started with grand ambitions were forced to crawl into a niche to survive - Kindred Robotics comes to mind. At the same time, startup investors want to see *massive* potential returns. ‘Find a compelling story about how a niche becomes a billion dollar market’ puts a technology between a rock and a hard place.[^4] 

In most cases, it wouldn’t have been feasible to do the necessary work to find or develop for a better niche in an academic setting either. While that work still requires “research” it also requires focus and systems engineering, both things that academia does not support for a slew of reasons. [[Academia is not a good environment for systems engineering]]. [[Academia incentivizes novelty, not focus]]. 

It’s of course possible that I and others just overestimated the potential of many of these technologies. However, the examples of technologies that were massively impactful only after “hanging out” in a corporate research labs for many years would seem to argue otherwise.  The [[Transistor]], [[Public key cryptography]] and  the [[Graphical User Interface]] (and personal computing in general) come to mind. Ideally corporate research labs provide the environment for to do the work to get a technology to a point where it’s viable for a ‘useful’ niche and enable it to find that niche in the first place. 

In addition to simply having longer time scales, less existential risk, and larger budgets than most startups, there are several specific ways that corporate R&D labs (used to) help technologies find good niches that startups and academia didn’t provide.

When you’re still trying to figure a technology out, it’s not clear which skillsets you want in the room. Corporate R&D allows there to be people floating between different projects loosely creating and breaking collaborations. [[Bell Labs enabled free radicals]].  To get the same effect at a Startup, you need to either find a magical individual who has gone deep on multiple areas (M shaped individual?) or hire people who might ultimately turn out to be useful. 

[[Bell labs asked people to do useful things but left them room to say huh that’s funny]] - often that moment of “huh that’s funny” can lead to the useful niche. 

 [[Healthy R+D orgs feel like they have very smooth project rampups with high ceilings]] - the discontinuous rampup implicit in a high-growth startup can force premature committal to a specific niche and the ceiling imposed by academia can prevent you from doing the work to accurately test a niche in the first place. 

Of course, this is all talking about an idealized corporate R&D organization, which arguably has become exceedingly rare. The need to produce demos for corporate higher-ups may serve the same ‘specialization forcing function’ as startups. One way to look at this would be to ask ‘how has the *management* of corporate R&D changed over time? On the hypothesis that [[People who have done a thing should be in charge of a thing]].  On the other hand, The [[Mother of All Demos]] was in fact a demo. Same with SpaceX live fire/other Elon musk things. So the upshot might not be that “demos necessarily force specialization” but that [[Demos are like afterburners]] - applied strategically along a planned path they can allow you to jump gaps you would otherwise be unable to cross, but applied whenever you feel pressure will leave you without any fuel.

### Related
-  [[Ideally corporate R&D enables targeted dicking around]]. 

[^1]: I will say this very aggressively in the hopes that I am wrong. [[Cunningham's Law]]
[^2]: This idea might be a fallacy. Arguably there is no platonic “ultimate stage” of a technology and whatever a technology evolves into is indelibly touched by the different specialized forms it needed to pass through. 
[^3]:There are a number of just-so stories I could tell about why these university spin-offs are different than the list of successes - most of those on the list were not deeply new technology, their founders were brilliant and plotted the exact path to success, it’s easier to go into new niches with software and chips, they got lucky, I’m a moron etc. However, the empirical evidence is that so many niche-seeking startups built around new (genuinely cool) technologies fail to help that technology realize its potential. 
[^4]: I’m *not* trying to say VC and startups suck and are broken! There are many amazing things that *have* come out of the system. What I am asserting is the incorrectness of the idea that “if it wasn’t a viable startup it would never have been able to be amazing for the world in the long run”

[Web URL for this note](http://notes.benjaminreinhardt.com/Ideally_corporate_R+D_enables_work_on_a_general_purpose_technology_before_it’s_specialized)

[Comment on this note](http://via.hypothes.is/http://notes.benjaminreinhardt.com/Ideally_corporate_R+D_enables_work_on_a_general_purpose_technology_before_it’s_specialized)






§Startup Constraints

# If a project has massive possible upside and low enough risk it is venture fundable
Software companies basically have infinite possible upside and you can’t get higher than infinity.



[Web URL for this note](http://notes.benjaminreinhardt.com/If_a_project_has_massive_possible_upside_and_low_enough_risk_it_is_venture_fundable)

[Comment on this note](http://via.hypothes.is/http://notes.benjaminreinhardt.com/If_a_project_has_massive_possible_upside_and_low_enough_risk_it_is_venture_fundable)




Brains are wombs for heuretics

# It is easier to have a multipoint mutation in technology than in evolution
A multipoint mutation is analogous to [[Systems must take performance hits to get out of local optima]] - in the tech tree model this means having one or more low impact precursor nodes.  

Almost by definition you need multipoint mutations to fix [[Structural Constraints]] because structural constraints mean that there are two or more interlocking dependencies. 

In biology, it’s (impossible?) to have a precursor with a single mutation that renders it inviable give rise to a precursor that has a second mutation that addresses issues from the first and builds on it. We’ve used genetic engineering and mutation breeding to get around this fact. ([[What does technological genetic engineering look like?]] and [[What does technological mutation breeding look like?]])

In technology a non-viable (non impactful) precursor can be kept alive in a lab or in an obscure paper somewhere. These are [[Unexploited Leonardo Ideas]]! Which may give a hint that the process of finding unexploited Leonardo ideas ([[Leonardo Ideas can be systematically identified]]) may be part of the answer to [[What does technological genetic engineering look like?]]

 - It is hard for startups to address multipoint mutations because it takes a long time and its hard to sell as a product [[§Startup Constraints]]
 - It is hard for corporate R&D to address multipoint mutations because people often have vested interests in the old system. [[§Corporate R+D Constraints]]
 - Academic research often creates proposals for multipoint mutations but because [[Academics are incentivized to do peer-reviewed publishable work]] they have no incentive to do implementation. [[§Academia Constraints]].
 - Why does 

### Related
- 



[Web URL for this note](http://notes.benjaminreinhardt.com/It_is_easier_to_have_a_multipoint_mutation_in_technology_than_in_evolution)

[Comment on this note](http://via.hypothes.is/http://notes.benjaminreinhardt.com/It_is_easier_to_have_a_multipoint_mutation_in_technology_than_in_evolution)




Week of 18 Jan 2021

Startups have taken over a chunk of corporate R+Ds niche

# One function of a private ARPA is to fill in the gap once filled by corporate R+D
[[In the past, corporate labs filled a particular niche in the innovation ecosystem]].  [[We need new institutional structures to fill the role that corporate labs once occupied]].

In order to do that, lets walk through the pieces of the niche one by one and examine how a private ARPA could make sure that it’s structurally set up to cover each piece.

1. [[Ideally corporate R&D enables work on a general purpose technology before it’s specialized]]

This will require explicitly and uncomfortably pushing back against the (accurate) common startup wisdom to “focus on a niche.” The more you follow startup wisdom, the more you’ll be bound by [[§Startup Constraints]].  At the same time any technology *does* need to find an initial niche eventually and a program should have some precise hypotheses about an eventual application. Counterintuitively, [this tension between considering how a technology will be used and potential local-optimum traps]<[[Tensions are the core theme of the PARPA writeup]]> may be one of the barometers for a healthy organization, ensuring that it doesn’t lean too far one way or another. Maintaining this unstable balance will require a lot of work to frame goals correctly. While cliche, the development of the transistor is a good example - it had a precise goal “find a solid-state replacement for the vacuum tube that we think will be useful in telephone repeaters and switches” but didn’t constrain the work to only “make more efficient repeaters and switches.” 

2. [[Ideally corporate R+D enables targeted dicking around]] - specifically with equipment and collaborators who would not otherwise be accessible. 

The structure of an ARPA-riff can enable targeted dicking around through low-stakes seedling programs, having programs exist under a single organizational umbrella, and having some number of ‘free radicals.’

An ARPA-riff that copies DARPA’s use of [low-stakes seedling programs]<[[DARPA PMs use seedling programs to ‘acid test’ the riskiest pieces of a program idea]]> can explicitly go into them with the attitude of “let’s see what will happen.” It’s important to acknowledge that of course, everyone would rather the experiments point to a promising program than be a nothingburger, and the way to prevent that preference from morphing into pressure is to explicitly make sure that negative results don’t have lasting effects. Specifically, this means that negative results can’t affect a PM’s ability to run other experiments or affect a performer’s relationship with the organization as a whole.[^1]

An organizational umbrella over programs can prevent serendipitous results from evaporating which can make targeted dicking around beneficial over the long term.[^2] If each program was a separate organization, pursuing any serendipitous discovery that didn’t help it move towards its goals would be a distraction. Someone might keep track of them but there would be relatively little incentive to do so.[^3] An umbrella structure would enable serendipitous potential side-quests discovered during the main body of a program to be fed back to the beginning of the “pipeline” and possibly become its own program. 

An ARPA-riff could also enable targeted dicking around by explicitly having people whose job is to do targeted dicking around.  These “free radicals” could be competent individuals who aren’t attached to any specific project. They could help PMs design programs, do experiments, or research areas that nobody in the organization has even thought about. You could imagine this role as having the flavor of a “pre-PM X-in-residence” where it is like an internship in that the person isn’t doing anything mission-critical and could potentially lead to them becoming a PM without the expectation of that outcome.  It would be unlike an internship because it would involve much more independence and the people in the role would be expected to already be skilled so there would be less of a teaching component. This sort of role is anathema to efficiency and could easily be a waste of resources, but rhymes with how [Bell Labs had researchers who were ‘between projects’ and would help out other researchers or poke at possibilities.]<[[Bell Labs enabled free radicals]]>.[^4]  

3. Corporate labs enabled high-collaboration research work among larger and more diverse groups of people than in academia or startups. 

The first way a private ARPA could fill this niche is through interactions between program managers. Program managers should be drawn from a wide range of technical experience. This diversity combined with setting up cultural expectations are set up correctly (riffing off of [[Bell Labs]]’ open-door policy[^5] for example) could go a long way towards fulfilling this part of the niche. 

It would seem like programs that initially revolve around externalized research might present a challenge to creating high-collaboration environments. However, DARPA PMs seem to do a good (emulatable) job addressing these challenges by holding regular closed-door meetings (so people can talk about real problems instead of just showing off good results) and [making sure different performers know each other.]<[[A large part of a DARPA program manager’s job is focused network building]]>  [^6] The shift from externalized research to internalized over the course of a [[Exploratory Program Organization]] will make the program further look like parts of a corporate lab. 

Realistically, a private ARPA won’t be able to fulfill this role as well as a golden-age corporate lab, but it will fill this part of the niche better than academia or startups do today. 

4. Corporate labs enabled very smooth transitions of technologies along the [[Technology readiness level - TRL]] curve / corporate labs cared both about newness and scale. 

Seedling programs, encouraging serendipity, and free radicals (see ::put link to #2::) as well as [running multiple programs with different levels of maturity]<[[A private DARPA riff both can and must do multiple experiments at once]]> could go a long way towards enabling smooth transitions from early TRLs to middle TRLs.  Admittedly, smoothing the transition from middle TRLs to later TRLs is tricky to structurally build into a private ARPA. One potential smoothing mechanism is how [[Exploratory Program Organization]]s can become independent organizations whose purpose is to get the technology that they developed out into the world, either by selling it as a startup or helping people adopt it as a non-profit. Of course, many technologies are more effective if they're absorbed by existing organizations — especially things like new processes.[^7] Building a consistent system to encourage existing organizations to integrate new technology is incredibly hard — even [[DARPA has a mixed record on transitioning]]. Despite that difficulty (and perhaps because of it) [it’s something that an ARPA-riff needs to strive for from day one.]<[[An ARPA-Riff could establish sales channels for frontier tech]]>::This feels like an awkward link::

5. Corporate labs provided a default customer for a lot of process improvements and default scale for products,

Similar to the transitions to existing organizations mentioned in #4 above, this is a challenge for an ARPA-riff compared to a corporate lab. [There are several experiments that the organization can and should try in order to establish default customers]<[[An ARPA-Riff could establish sales channels for frontier tech]]> but they are all contingent on relationships, reputation, and specific circumstances. Realizing that positioning technology to fail is a [potential failure point]<[[Reasons why an arpa riff would succeed or fail]]> will hopefully make it into a self-defeating prophecy but at the same time, too much focus on adoptability can push programs towards boring safe decisions.  
::Feels awkward::

6. Corporate labs provided a precise set of problems and feedback loops about whether they were actually solved

Not being officially attached to a parent company makes an ARPA-riff by default worse at this than a corporate lab so it’s especially important to explicitly set up these feedback loops wherever we can. However, feedback loops are a piece of the golden-age corporate niche where an umbrella organization over a number of programs is more useful than a set of stand-alone programs[^8].  Consistent cycling of PMs with a diverse set of industry experiences provides one feedback loop — it suggests that  ‘industry experience’[^9] should be something to look for in PMs. 

There are a number of other (hand-wavy, but worth noting) ways to set up feedback loops between an ARPA-riff and real problems. Like [[Ink and Switch]], the organization could do consulting work to help external organizations implement new tools/processes. In addition o feedback this sort of consulting would both help address #5 above and provide an alternative revenue source. An ARPA-riff could set up a consortia similar to the [[Media Lab]], which would give the organization a default reason to talk to other organizations. 

7. Corporate labs provided a first-class alternative to [academia where people could still participate in the scientific enterprise]<[[Different institutions have created monopolies on various enterprises]]>

Especially early in their lives, a private ARPA’s programs will still lean heavily on academic work, so the organization won’t address this part of the niche as thoroughly as corporate labs or Focused Research Organizations. However, if the idea of ‘gradual internalization’ works, [private ARPA programs]<[[Exploratory Program Organization]]> could become an alternative to academia for some people, some of the time. 

This is a place where organizational discipline is important. The temptation to go full “we’re going to fix research!” is strong but I suspect that attempting to provide a first-class alternative to academia would undermine many of the things that make the ARPA Model work in the first place. Instead, the organization should support *other* alternatives to academia by not biasing towards researchers with academic affiliations. While a single organization is probably unable to offer a first-class alternative to academia, one could dream *very* big and imagine that an alternative could emerge from an ecosystem of ARPA-riffs and other new types of research organizations. This is yet another reason why [[PARPA should create a replicable institutional model]].

8. Corporate labs enabled continuous work on a project on a 6+ year timescale

One might expect that it would be hard to support long-term projects in the ARPA model because PMs with relatively short tours of duty is a core feature. However, DARPA has a strong precedent of programs outlasting the tenure of PMs who started them. There’s no reason a private ARPA couldn’t do the same thing. Arguably, PM-PM handoffs can maintain continuity better than many academic labs where most of the work is done by grad students with ~5-6 year tenures with very little incentive for smooth handoffs[^10] or startups, where employees are notoriously mercenary and it can be surprising if someone stays more than three years.  

However, it’s also important to consider constraints that the organization will face early in its life that will hopefully relax over time. While an ARPA-riff is structurally set up to tackle long-term projects, it’s probably not a good idea for the first projects to take 6+ years to show results. The reality of the situation is that a new organization will have no track record and so people will (rationally) have less patience with it. Organizational reputation matters. If funders, potential employees, and collaborators start to see the organization as unable to get things done[^11] it can become a self-fulfilling prophecy. 

9.  Corporate labs enabled work in [[stokesPasteurQuadrantBasic1997]] - explorations into natural phenomena with eye towards exploiting those phenomena for human flourishing (and profit.)

There’s not much to say here besides this is the tightrope that an ARPA-riff should try to walk: things that are too research-y for a startup but not paper-focused enough for academia. 

[^1]: To paint the counterfactual of explicitly making sure experiments have no negative consequences: One could imagine that after a failed seedling experiment a PM could implicitly or explicitly need to ‘redeem’ themselves by going above and beyond. Similarly, the organizational attitude towards a performer who ran a failed experiment might implicitly shift towards “well they’re good people but a bit incompetent.” 
[^2]: As opposed to something like flossing where people are like “yeah I really should be doing that … later.”
[^3]: This is an example of the more general principle that [[Optimizing individual components of a system is often at odds with the system itself]].
[^4]: These are the ‘fellows’ in Phase Two of [[Pathways towards PARPA]].
[^5]:<sidenote>`They were not to work with their doors closed. They were not to refuse help to a colleague, regardless of his rank or department, when it might be necessary.` From [[gertnerIdeaFactoryBell2012]]. 
[^6]:[[Focused network building is a valuable early step for riffing on ARPA]]
[^7]:<sidenote> See [[fillerFundamentalManufacturingProcess2020]] to expand on this point.
[^8]:  <sidenote> See [[Adam Marblestone]] and [[Sam Rodriques]]’s [[Focused Research Organization]] proposal. 
[^9]:As suitcase handle-y as that term is. 
[^10]:”I’m busy writing my thesis. Here’s my lab computer - it has everything on it. Good luck!”
[^11]: People absolutely run out of patience even if you set expectations up front.  https://twitter.com/Ben_Reinhardt/status/1350507766381744128



[Web URL for this note](http://notes.benjaminreinhardt.com/One_function_of_a_private_ARPA_is_to_fill_in_the_gap_once_filled_by_corporate_R+D)

[Comment on this note](http://via.hypothes.is/http://notes.benjaminreinhardt.com/One_function_of_a_private_ARPA_is_to_fill_in_the_gap_once_filled_by_corporate_R+D)




It is hard for a private ARPA to capture value without hamstringing impact

# Profit-maximizing organizational structures can hamstring technologies whose impact depends on them
Profit maximizing organizations obviously have incentives to try to capture as much of the value they create as possible. Of course, profit maximizing companies are not mindless greed-monsters. In many situations, giving up short term profit can actually lead to longer term revenue. Ideally, these organizations would be able to sit at some point on the tradeoff curve that maximized value *by* maximizing impact. Keep in mind that profit-maximizing organizational structures can also be the best way to support a technology! [[tensions]]. 

However, there are four mechanisms that can lead to good-acting but for-profit companies hamstringing technological advancement.

*Specialization*.Generally, it’s a good idea for a for-profit business to start with a single product and a specific market or niche. This good business practice means that short of a large company with an R&D department, most of the development work should be focused on productizing a technology for a specific task. Startups who want to *start* by building a platform or doing all the things tend to go down in glory-seeking flames. Specializing technology for a specific task is not inherently a bad thing! It’s also possible to thread a path of increasing generality, starting from a small niche and jumping to bigger and bigger applications over time. Elon Musk is the master of this. However, it’s just as possible to get stuck in one of those niches, especially if it’s not profitable enough to support R&D on the more general technology. This situation happens all the time with university spin-off technology, where it really needs more general development to reach a certain performance threshold, but it gets stuck in a niche in order to build a company.  <Need examples.>

*Investor Pressures* For-profit companies working on technology generally need investors because they won’t turn a profit for at least several years after getting started. These investors have their own incentives, and maximizing the impact of a technology generally is not the most important one. Investors can have specific timelines for an exit, incentives to show growth in order to raise a new round, or just opinions about where the technology should go that conflict with maximizing impact. (Of course, any source of money comes along with its own set of pressures!)  

*Capturing value* For-profit companies need to capture some value that they create! [[tensions]], but [[There is a significant class of innovations that would create drastically less value for the world if their value had been captured by their creators]]. 

*Speed* An organization could realize all of the above traps and forgo trying to maximize profit, not take investment, and make most of its technology work public, pulling in revenue through non-exclusive license fees and consulting work. Unfortunately, this strategy can hamstring the technology you’re working on in a different way, slowing general-purpose development to the work that can be done when you have particularly lucrative contracts. Many aerospace technologies seem stuck in this particular mechanism. 

It is rare that an organization can both forgo significant value by making key parts of a technology public *and* make money hand over fist at a rate that satisfies investors and can fund the development of general-purpose technologies. 

### Related
- [[It is hard for a private ARPA to capture value without hamstringing impact]]
- [[§Startup Constraints]]
- [[§Corporate R&D Constraints]]
- [[It is hard to capture value from research]]



[Web URL for this note](http://notes.benjaminreinhardt.com/Profit-maximizing_organizational_structures_can_hamstring_technologies_whose_impact_depends_on_them)

[Comment on this note](http://via.hypothes.is/http://notes.benjaminreinhardt.com/Profit-maximizing_organizational_structures_can_hamstring_technologies_whose_impact_depends_on_them)




Xora has unique structural advantages that could allow it to push the technological frontier

# Startup Constraints - essay
- [[Startups need to make a lot of money or have a lot of users]]
- [[Startups need to conform to investor timescales]]
- [[Startups need an exit strategy]]
- [[Startups need a great story]]
- [[Startups need to produce a scalable product]] 
- [[Startups need to break development into correctly sized stages]]
- [[Startups need specific growth curves]]
- 
### Published
10/2018
## Questions
## Thoughts
## Related
[[§Startup Constraints]]
## Authors
- [[Benjamin Reinhardt]]
## Link
[Startup Constraints](https://benjaminreinhardt.com/startup-constraints/)

[Web URL for this note](http://notes.benjaminreinhardt.com/Startup_Constraints_-_essay)

[Comment on this note](http://via.hypothes.is/http://notes.benjaminreinhardt.com/Startup_Constraints_-_essay)






Addressing structural constraints often requires drawing on more than one discipline

# Startups and Academia are bad at addressing structural constraints 
Both systems encourage making one single clear contribution at a time. 
- Papers are supposed to be discrete building blocks that add one brick to the wall that is ‘the literature’
	- [[Academics are incentivized to do peer-reviewed publishable work]]
- It is common wisdom that startups need to focus on changing one thing at a time.

In each case, there are rational good reasons for these constraints but it makes it hard for either system to address [[Structural Constraints]].

Additionally,  [[Problems that require system-level solutions often require massive teams]] and both startups and single labs do not have those teams. Interestingly, the [[§ARPA model]] may be able to cheat around this constraint by using externalized research. [[R+D orgs lie on a spectrum between externalized and internalized execution and benefits]]

### Related
- [[§Startup Constraints]]
- [[§Academia Constraints]]



[Web URL for this note](http://notes.benjaminreinhardt.com/Startups_and_Academia_are_bad_at_addressing_structural_constraints)

[Comment on this note](http://via.hypothes.is/http://notes.benjaminreinhardt.com/Startups_and_Academia_are_bad_at_addressing_structural_constraints)




# Startups need a great story


[Web URL for this note](http://notes.benjaminreinhardt.com/Startups_need_a_great_story)

[Comment on this note](http://via.hypothes.is/http://notes.benjaminreinhardt.com/Startups_need_a_great_story)




# Startups need an exit strategy


[Web URL for this note](http://notes.benjaminreinhardt.com/Startups_need_an_exit_strategy)

[Comment on this note](http://via.hypothes.is/http://notes.benjaminreinhardt.com/Startups_need_an_exit_strategy)




Funding is path dependent and funding curves matter

Inverting the normal relationship between an R+D organization and is money factory may enable new classes of activities

# Startups need specific growth curves
* The Valley of Death must be small. Most VCs will not fund a company that spends too long with low revenue or few users. This is another manifestation of the timescale constraint.
* Growth must be monotonic (always going up and to the right.) In rare circumstances a startup can survive flat or declining growth for a period if its management had the foresight or luck to raise funds before the lackluster growth that last until the growth resumes up-and-to-the-rightness.


[Web URL for this note](http://notes.benjaminreinhardt.com/Startups_need_specific_growth_curves)

[Comment on this note](http://via.hypothes.is/http://notes.benjaminreinhardt.com/Startups_need_specific_growth_curves)




# Startups need to break development into correctly sized stages
A startup needs the ability to hit progress/derisking milestones with the amount of money typically raised in each venture round. VC funding happens in fairly standard round sizes: The first round is almost always <$3m. If you need $8m or $30m before your first major milestone, it’s hard to fund the project with VC money.

![](../BearImages/E8BE953F-5BA8-46C3-BD7D-796F35625F15-621-0003D44334B93FF0/startup-growth-curve.png)

[Web URL for this note](http://notes.benjaminreinhardt.com/Startups_need_to_break_development_into_correctly_sized_stages)

[Comment on this note](http://via.hypothes.is/http://notes.benjaminreinhardt.com/Startups_need_to_break_development_into_correctly_sized_stages)




In order to beat the stock market on average a 20 year project needs to return 4x

# Startups need to conform to investor timescales
[[Venture capitalists need to fund things that increase valuation in a few years and return in 5-7 years]]



[Web URL for this note](http://notes.benjaminreinhardt.com/Startups_need_to_conform_to_investor_timescales)

[Comment on this note](http://via.hypothes.is/http://notes.benjaminreinhardt.com/Startups_need_to_conform_to_investor_timescales)




# Startups need to make a lot of money or have a lot of users


[Web URL for this note](http://notes.benjaminreinhardt.com/Startups_need_to_make_a_lot_of_money_or_have_a_lot_of_users)

[Comment on this note](http://via.hypothes.is/http://notes.benjaminreinhardt.com/Startups_need_to_make_a_lot_of_money_or_have_a_lot_of_users)




One way to think about scalability is as the number of operations that happen without human intervention

# Startups need to produce a scalable product


[Web URL for this note](http://notes.benjaminreinhardt.com/Startups_need_to_produce_a_scalable_product)

[Comment on this note](http://via.hypothes.is/http://notes.benjaminreinhardt.com/Startups_need_to_produce_a_scalable_product)




Low Hanging fruit theory of stagnation

There are three main theories of stagnation

Human capital theory of stagnation

Regulations have increased by an order of magnitude

# Systemic decay theory of stagnation
The systemic decay theory of stagnation is a loose collection of assertions that changes in institutions - government, universities, culture - are responsible for [[Stagnation]].

[[storrshallWhereMyFlying2021]] points to ever-increasing regulation as the knife in the throat of innovation. Regulation doesn’t need to explicitly ban innovations - it just needs to make the hurdles they have to clear to be adopted so high that they never make it out of the lab/prototype stage.  [[Regulations create paradigm lock-in]], and outside of economic metrics ([[collisonScienceGettingLess2018]]) the nebulous but real evidence for stagnation is that we are not seeing as many paradigm shifts in atom-based technology. Additionally, there are several technologies we can point to that are not commonplace primarily because of regulations (nuclear power, most genetically modified things, many helpful drugs) so how many *don’t* we know about. [[Arguably cars wouldn’t exist under the current regulatory regime]]. 

Some people blame the fact that [[Universities have been pushed more towards applied research since the end of WWII]] at least in part because of the tight coupling between government funding and academia. Most politicians see government spending on research as an investment  and applied research has a much clearer ROI. I find this argument a little weak because lots of non-applied research still happens in universities. Yes, a smaller fraction is unfettered discovery, but I haven’t seen any compelling reason to believe that the fraction matters.

While the composition of university research might not be to blame, there are a lot of theories that center around the coupled dependence on and mismanagement of government funding for research. Everything from risk aversion ([[Evaluating transformative research programmes: A case study of the NSF Small Grants for Exploratory Research programme]]) to the bureaucratic apparatus that makes it extremely painful to apply for grants ([[Giving Away Money Costs More Than You Think]]), for the most part only gives small or absolutely huge grants. [[storrshallWhereMyFlying2021]] goes so far as to point out that GDP growth is in fact inversely correlated with non-military government R&D funding. Most foundations are not much better. 

One contributing factor might be the death of corporate R&D. We can point to so many impactful innovations that came out of corporate R&D in the middle of the 20th century. [[aroraChangingStructureAmerican2020]] notes that incumbent companies have shifted towards expanding markets by mergers and acquisitions.[^1] This strategy shift means that most innovations need to either work within all the constraints of a startup ([[§Startup Constraints]]). 

There’s a mini-nebula of theories around [[The (idea) valley of death]].  What would ideally be a communication-rich channel going both ways between people trying to innovate and improve on existing technologies and the people who use and sell them is instead a high friction, leaky pipeline with very little feedback. [[Academic research is often just the concept half of creating a heuretic]], but a proof of concept is often not enough to derisk a for-profit company.  [[drexlerRadicalAbundanceHow2013]]  argues that many innovations don’t happen because  [[Academia is not a good environment for systems engineering]], which is necessary and underrated for building actual functional products.  There is another class of theories that the valley of death has widened because globalization and other forces have separated manufacturing from research. [[Definite Optimism as Human Capital]]. 

The discussions around The Valley of Death invariably touch on the patent system. These theories go that [[The relationship of government, academia, and technology fundamentally changed in 1980 with the Bayh-Dole act]]. Now technology transfer offices are all trying to grab their share of the profit (despite the fact that [[Most tech transfer offices are not profitable]]), making it hard to combine patents into impactful innovations, demanding a large enough share in companies that it makes it hard for some to get funded, and generally increasing the friction to turn university research into useful products. [[Any technology that comes out of a university needs to go through their tech transfer office]]. Corporations also abuse the patent system - taking out defensive patents that mean paradigm shifts that require several new pieces are all potential lawsuits. 

The patent issues point to a more [[Nebulous]] systemic problem: declining [[Slack - concept]] across institutions: Academia, Incumbent Corporations, and Startups. This idea is a massive rabbit hole so I will touch on it only briefly, but in general [[Systems must take performance hits to get out of local optima]] but without enough slack, any system (or individual) that takes a performance hit will get competed out of existence. The oversaturation of academia with people, potential returns to investing in software, patent obsession, regulations, and decline of R&D labs may all have contributed to this declining slack.[^2] 

Declining slack is coupled to an *even more* [[Nebulous]] issue: culture. In part, the world is so competitive because the pie doesn’t feel like it’s growing anymore in the first place, so it’s unclear which way the causality runs between [[Stagnation]] and culture. Peter Thiel’s critique is mostly a cultural one ([[Back to the Future (Thiel)]]) that we don’t make and support big, precise plans ([[Definiteness means that you have a long-term plan]] and generally don’t have a definite optimistic stance, embracing indefinite optimism which leads to incremental improvements and short-term optimization. [[To make new things you need to take risks]] and there’s a sense that we are just more risk averse (see the previous paragraph about regulation.) Perhaps related is the phenomena of increasing leader age  across institutions. And there’s a sense that with the decline of religion and nationalism we aren’t striving for anything bigger than ourselves or our lifetimes. 

### Relationship to other stagnation arguments

The systemic decay theory is almost the dual of the [[Human capital theory of stagnation]], which stands to reason: systems both are composed of individuals and drive their incentives. The difference may be in the flavor of solutions that each theory suggests - whether you focus on shifting individual incentives or institution building/reform. Even then, the lines are blurry.

All of these varied causes could be *why* [[Low Hanging fruit theory of stagnation]] is correct. Perhaps there is a limit to how coordinated humans can be and we’ve picked most of the fruit that can be easily grasped within those constraints. Similarly, all of these effects could be *because* we’ve run out magical growth pixie dust. 

### Counterarguments
I need to do more work to understand the counterarguments more deeply. They boil down to the fact that as a society we *are* spending a lot of money and manpower on research. It’s not like *nothing* new is being created, and we’ve tried a lot of things that just haven’t worked. 

### Related
- [[Notes on the changing structure of American Innovation]]
- [[What changed from the high-quality corporate labs in the past?]]
- [[Any technology that comes out of a university needs to go through their tech transfer office]]
- [[Low hanging fruit implicitly assumes a linear development model]]
- [[Stagnation]]

### Works not cited in the body
- [[aroraChangingNatureCorporate2015]]

[^1]: [[AI research is an exception to the decline of corporate R+D]]
[^2]: Note that I’m not saying competition is bad. Simply that more competition is not always better. [[alexanderStudiesSlack2020]].



[Web URL for this note](http://notes.benjaminreinhardt.com/Systemic_decay_theory_of_stagnation)

[Comment on this note](http://via.hypothes.is/http://notes.benjaminreinhardt.com/Systemic_decay_theory_of_stagnation)




Good roadmaps work backwards from a goal

PARPAs second evolutionary stage is an architecting institute

PARPAs third evolutionary stage is a full private ARPA

# The Architected Innov􏰂ation Institute - Engineering Scientific Revolutions
- The core idea is that there is an organization that hosts “Architects” 
- Architect may be a better term for PMs
- Principles
![](BearImages/4B427308-60F0-4089-945D-B3CA72B607DF-98199-000AA867B4581C46/DB10724F-8118-40FA-A73E-9E4ACDE2B613.png)
	- Architects must be omnidisciplinary
		- Architects need to know multiple fields at multiple levels of abstraction
		- Need to be able to seek out detailed technical consulting from larger groups of experts 
	- Architects must care only for enduring Impact
		- ::How do you se up incentives to do this?::
	- Roadmap problems systematically and comprehensively
		- The typical notion of interdisciplinary research is to optimize for serendipity 
			- ::Extension Need to engineer who is in the room?::
		- A key to roadmapping is to evaluate all possible solutions in a maximally unbiased way - working backwards from the problem
	- Avoid being a fixed set of people
		- The very notion of a finite institution saving a complete problem on its own might be an instance of push thinking
	- Be agnostic to where collaborators come from
		- No matter how great your institute, most of the relevant people are not going to pack up and move to it
	- Create and leverage trusted networks
		- How to create trusted networks
			- Pairwise interactions between architects and potential team members
			- Pay attention to who works well together, public vs private information
			- Architects need ‘unimpeachable’ creditability
	- Design around problems rather than debugging through them
		- Use existing arts whenever possible - it forces you to find things
		- Outsource everything possible
		- View skills as commodities
- “Latent Architects”
![](BearImages/E13B472B-E081-43FD-95A7-FB2D9878E815-98199-000AA965F7D5A0BB/05D67568-400A-4513-AB45-F8C42F335049.png)
![](BearImages/52FD904E-E434-4BAD-9465-A61A5FE33EF8-98199-000AA96A63A54C2A/590A6437-3B30-4069-8BBB-78BD7E13DDEA.png)
![](BearImages/70C1AE8A-99B7-46A3-9256-834F07EBF423-98199-000AA96D947A9814/93B0DB03-6E32-4D07-A5BB-700920836F6A.png)
- [[§Academia Constraints]]
- 
![](BearImages/3179F3EE-50B2-49A0-9EFF-ADEB2C91BA20-98199-000AA9769BCC520C/7CCBD296-3331-44DF-91F6-5116A47892EC.png)
- [[§Startup Constraints]]

- Features of the Architected Innovation Institution
	1. Act catalytically
		- Estimate 5000 sqft for 10 architects and $25m/year
		- Dynamic funding is important
	2. Generate strategies and act flexibly to execute on strategies
	Possible moves
![](BearImages/49AE1095-2973-4DB4-A797-610557C08B66-98199-000AB6FAD696D488/C55B9441-8EAA-4B58-82CD-EC724225D6E9.png)
	3. Creates a core capability in ground truth biological analysis
	4. Core services to facilitate efforts 
		- Need to have technical people on staff
	5. Provides flexible mechanisms of support for exceptional individuals who do not fit in elsewhere
		- Maintain “fellows” with no projects or goals
- Punts on funding question
![](BearImages/067A6658-AD76-47DD-8D11-3D24958AE9EA-98199-000AB75C67154EBA/148E4D13-324E-49CB-8C1D-E82A692EEFA3.png)
### Published
?
## Questions
## Thoughts
- Possibly a great intermediate step for [[§PARPA]] - at $25m/year level
- Think of PARPA planning in terms of money 
## Related
## Authors
- [[Adam Marblestone]]
## Link
https://www.dropbox.com/s/1a3aqw05ieeaebw/the%20architected%20innovation%20institute_2020-09-18%2016.20.35.pdf.pdf?dl=0



[Web URL for this note](http://notes.benjaminreinhardt.com/The_Architected_Innov􏰂ation_Institute_-_Engineering_Scientific_Revolutions)

[Comment on this note](http://via.hypothes.is/http://notes.benjaminreinhardt.com/The_Architected_Innov􏰂ation_Institute_-_Engineering_Scientific_Revolutions)




# The market in for-profit organizations is fairly efficient
The barriers to creating a for-profit organization boil down to initial costs, talent, and long term funding. All of these barriers have become low enough that even crazy ideas that are good investments should tend to get funded.

The costs to starting a company are low - even if you need a lot of equipment because it is relatively easy to borrow lab space or equipment time. The startup ideology has pervaded society to such an extent that even [[People who should not be starting companies are starting companies]].

Socially, entrepreneurs are revered and people have been so indoctrinated that ‘stupid’ ideas can become billion dollar companies that there isn’t much inhibition the talent front.

On the funding side, there are more early-stage VCs than ever and despite the fact that technically [[Venture capital doesn’t work for things that only realize their value on long timescales]], the sexy success of SpaceX has encouraged at least some of them to bet on long term projects that could be wildly valuable. [[If a project has massive possible upside and low enough risk it is venture fundable]]. Additionally, angel investing has become a status symbol in some circles, pushing capital’s risk tolerance even farther. 

In combination these factors mean that the for-profit ventures that would be good investments but are not being started or funded are rare and on the margin. Of course, [[There are many important innovations that are just bad investments]].

Perhaps the place where they are inefficient is actually at the super capital-intensive end. The [[Long Term Stock Exchange (LTSE)]] will test this. 

### Counterargument
The counterargument to this assertion is that the success stories of impactful companies do seem to be incredibly contingent. Success depends on the right people being in the right place at the right time and being willing to doggedly persevere. Are there *systemic* ways to increase the rate of these contingent events?

### Related
- [[§Startup Constraints]]
- [[It is hard for a private ARPA to capture value without hamstringing impact]]
- [[There are many important innovations that are just bad investments]]



[Web URL for this note](http://notes.benjaminreinhardt.com/The_market_in_for-profit_organizations_is_fairly_efficient)

[Comment on this note](http://via.hypothes.is/http://notes.benjaminreinhardt.com/The_market_in_for-profit_organizations_is_fairly_efficient)




Efficiency biases systems towards false negatives

Malcolm Handley conversation 21 Sep 2021

# There are many important innovations that are just bad investments
Important innovations[^1] can be bad investments in both an absolute sense because they can’t capture much of the value they create and in a relative sense compared to other possible investments. 

Many innovations are bad investments because you just can’t capture that much of their value. These come in two categories. 
1. Innovations whose value is just inherently hard to capture. Research often falls into this category ([[It is hard to capture value from research]]) especially research that is spread between multiple organizations ([[It is especially hard to capture value from externalized research]].)  Innovations that resemble [[Problems where it is easy to know whether an answer is correct but hard to get to that answer]] are especially hard to capture from - these can be design innovations, just-right-recipes, best practices, etc. 
2. Innovations that would be hamstrung by capturing their value. [[There is a significant class of innovations that would create drastically less value for the world if their value had been captured by their creators]]

There are also innovations that are bad investments relative to other options. Low liquidity, high uncertainty, and long time horizons are all undesirable characteristics of investments and also characteristics of investments in potentially-paradigm-shifting innovations. When the stock market is providing >10% year-over-year returns that are very liquid and (perceived as relatively) low risk you need to have a convincing story that you can do *much* better than that with a low-liquidity, high-uncertainty project that won’t pay out for more than a decade. Limiting ourselves just to people to technology venture investments, you’re up against software companies that have incredible margins because they have low capital costs and often have known sales channels and little true technology risk. VC portfolios are already constructed so that any company they invest in needs to have a chance of a near-infinite valuation in order to pay for all the failure. It’s hard to claim a potential payoff higher than the one they already expect, so you’re effectively asking them to accept higher risk for the same potential reward.  [[If you need to have the possibility of near infinite returns on VC investments how is it possible to invest in higher risk things and expect a return?]]Unfortunately, the attractiveness of software as an investment renders many other innovations “bad investments” by comparison.   

Often the situation is a combination of the two situations - uncertain or even good-but-not-mind-blowing captureable value combines with capital costs, margins, liquidity, or timeline to make an innovation a bad investment for rational investors. 

### Related
- [[§Startup Constraints]]
- [[Can you get VC-level returns by funding research?]]
- [[Is it possible that base rate growth in the market is actually a bad thing for innovation?]]
- [[The market in for-profit organizations is fairly efficient]]

[^1]: Obviously the definition of an “important innovation” is nebulous. 



[Web URL for this note](http://notes.benjaminreinhardt.com/There_are_many_important_innovations_that_are_just_bad_investments)

[Comment on this note](http://via.hypothes.is/http://notes.benjaminreinhardt.com/There_are_many_important_innovations_that_are_just_bad_investments)




DARPA PMs use seedling programs to ‘acid test’ the riskiest pieces of a program idea

The DARPA execution framework boils down to showing that thing is not impossible, showing that thing is possible, and then making that thing possible

The DARPA Model for Transformative Technologies

The Value of Vision in Radical Technological Innovation

PARPA needs to be obsessed with PMs

NSF and DARPA as Models for Research Funding - An Institutional Analysis

# There are many tight feedback loops built into the ARPA model
PM hiring is part of a feedback loop between the organization and the outside world. PMs are hired to explore a specific area that DARPA should be exploring but isn’t yet. Even though it works on disruptive technology, DARPA needs to be responsive to changes in the world - whether it’s new possibilities (technology push) or new needs (technology pull.) [[Innovation orgs need to be aligned with their money factory]]  This buffering between the outside world and hiring is one of the ways [[DARPAs director is important]]. 

From day one there is a feedback loop between a new program manager and the other program managers. ‘Onboarding’ at DARPA is a super informal process that involves mostly oral traditions and shadowing. Informal onboarding is normally *terrible* best practice but most organizations don’t limit themselves to Dunbar’s number. [[DARPA is relatively tiny and flat]]. Program managers get significant informal feedback from other PMs as they hone their program design. This feedback includes whether something is ‘in scope’ for the organization ([[There are DARPA-style ideas that DARPA doesn’t pursue]]) and whether the design is precise enough. High quality informal feedback depends on other PMs actually caring, so this is another reason [[DARPA PMs need to think for themselves, be curious, and have low ego]]. 

There are multiple levels of feedback loop between the PM and the research community during program design. (steps 1&2 of [[The DARPA execution framework boils down to showing that thing is not impossible, showing that thing is possible, and then making that thing possible]])  The tightest loop is informal conversations with individual researchers and groups about where they see blockers and possibilities in the area that the PM can then incorporate into their program design. The PM then uses workshops to set up feedback loops between those possibilities - an idea might be good but an enthusiastic proponent could gloss over a major hole that needs to be filled. Finally, small projects test the riskiest parts of those possibilities against the world before the PM incorporates them into the program design. [[DARPA PMs use seedling programs to ‘acid test’ the riskiest pieces of a program idea]]. The workshops also set up feedback loops between previously unconnected groups of researchers focused on the program area that hopefully create new ideas and (ideally) new research communities.

PMs also set up feedback loops between different parts of an industry - academia, startups, and big companies.  [[A large part of a DARPA program manager’s job is focused network building]].  ::new:: My personal experience suggests that the feedback loop is not as tight between DARPA and the startup and VC world. DARPA created a specialized position just to try to help commercialize technologies coming out of their programs. However, they don’t seem to have a strong grasp on [[§Startup Constraints]].  

Feedback loops enable the PM to adjust, kill, or start projects as necessary during the execution of the program. Frequent contact with performers working on a project enables both the performer and the PM to incorporate new information into decisions on whether they should make small adjustments to the project’s goals and how that propagates into the rest of the program design.  PMs also use this information to decide whether to kill the project completely or start a new one. [[It is relatively easy for DARPA PMs to re-deploy funding]].  Failed projects are also part of a feedback loop: `The failure (of a performer) triggers a discussion in which the first question is whether the goals were correctly specified and how they might be redefined in the light of the research that has already taken place`

The ARPA model also uses long-term feedback loops to maintain high quality PMs and research collaborations.  [[A players hire A players and B players hire C players]] so high quality program managers make other high quality program managers want to join.  Low friction interactions with DARPA and money that enables performers to do things they wouldn’t be able to do otherwise keeps ideas coming from the community.

From [[The Value of Vision in Radical Technological Innovation]]
![](../BearImages/429CCF41-4757-4E30-9001-C133FBFC9DDC-314-00003F2B31A41FF5/649AF9B7-29B6-456E-AB44-A89F98506FBE.png)

### Related
- [[Feedback loops and play are important for breakthroughs]]

[Web URL for this note](http://notes.benjaminreinhardt.com/There_are_many_tight_feedback_loops_built_into_the_ARPA_model)

[Comment on this note](http://via.hypothes.is/http://notes.benjaminreinhardt.com/There_are_many_tight_feedback_loops_built_into_the_ARPA_model)






The way to get around journals is to focus on non-paper artifacts

PARPA needs a constitution with checks and balances

Mark Johnson the Elder Conversation 4 Jan 2021

Possible PARPA Missions

# Things that are not paper-worthy enough for academia and not product-focused enough for a startup
There are many heuretics (discoveries/inventions/innovations/ideas etc.) that don’t fall within the constraints of the institutions we expect to support them. “Things that are not paper-worthy enough for academia and not-product focused enough for a startup” is shorthand for one class of these heuretics. Note that there are many others! 

First, let’s address some misconceptions about what falls into this category.  Wacky ideas and long-term projects are two things that many people assume don’t get supported in universities or startups. This is not true.
	- Long-term projects get support *if* they hit consistent milestones *and* they’ll be able to capture a lot of the value they create if they succeed. 
	- Wacky ideas get support *as long* as they don’t threaten paradigms too badly *or* someone thinks they can make money off of them. 
	- Research projects that require a lot of people or coordination *as long as* the final product is a paper they can all put their names on. See the LHC. 
	- Research projects that require some engineering effort *as long as* the engineering effort resembles something that has been done before: building a bigger telescope/particle collider/linear accelerator/space probe, etc.

The reasons that ideas are not paper-worthy or product-worthy *make sense* given the incentive frameworks in their relative institutions.  [[Most people are not malicious or stupid]]. 

### What makes something not paper-worthy?
‘Paper-worthy’ is meant as shorthand for ’things academics are incentivized to do.’ So things that are not paper worthy are those that don’t obey [[§Academia Constraints]].  It’s important to explicitly call those implications out because that call-out can shine a light on where to dig for hidden gold. 

1. **An idea isn’t paper worthy if not considered “novel enough” or “not a contribution to the literature” by peer reviewers.** [[The word ‘novel’ is used as an idea bludgeon in academia]] and weeds out many classes of ideas. 

A non-exhaustive list of what gets shot down by peer review: 
- New approaches to old “solved” problems, 
- Better implementations of an existing idea that’s already been implemented, 
- Efficiency improvements along a different metric than the ones the field normally cares about, 
- Scaling existing techniques, 
- Creating something for a very specific application as opposed to a general principle/purpose, 
- Further work on things that have already gotten to proofs of concept
- System integration of several already-existing ideas.

For more see:  [[Academia is not a good environment for systems engineering]]. [[Academia incentivizes novelty, not focus]].  [[Academic research is often just the concept half of creating a heuretic]]. 

2. **An idea generally isn’t worth writing a paper about if it’s not expected to get a lot of citations.** Even if it’s published, low-citation ideas languish in obscurity. It’s valid to ask ‘are there things that are valuable that don’t get a lot of citations?’ 

I would argue yes, valuable projects that wouldn’t get many citations include:
	- Projects that are intermediate steps to a larger vision without a strong conclusion besides “we did a thing!”
	- Engineering work in a theory-heavy field or theory work in an engineering-heavy field.  More generally, work in a field that the field doesn’t think is valuable. 
	- Ideas that are far enough out of mainstream won’t get many citations because people don’t have the bandwidth to dig in and see if it’s worthwhile and the default answer is “no” (see #3 below) 
	
3. **It’s hard to publish ideas that are too outside of the approved way of doing things.** [[Paradigms outline a set of questions and the rules around answering them]].  Ideas can go too far outside of accepted paradigms when when you’re asking different questions, or measures success differently than standard ways. [[Antedisciplinary Science]] often falls into this trap as well, when it’s too much like discipline X to be published in a journal for discipline Y but it’s too much like discipline Y to be published in a journal for discipline X.  This situation is of course tricky because many (most?) ideas that go outside of paradigms are legitimate crackpottery. [[Anyone on the knowledge frontier is a bit of a crackpot]]. 

4.  **It’s hard to write a paper if you can’t even get the money to do the research to write the paper.**
The requirement to align with current funding priorities is particularly pernicious because [[The government grant process depends on politics and committees]]. Additionally, one reason to write papers is to continue playing the game, so if a paper doesn’t hint at your continued ability to do work that aligns with funding priorities, you’re incentivized not to work on it. [[People giving out grants try to derisk them as much as possible]].

5.  **A project is unlikely to lead to a paper if it’s not within scope for a PhD student or untenured professor.** 
[[Graduate students are the labor in academic science and engineering]]. In exchange for cheap labor, grad students expect to produce research within a 4-7 year time frame that can help them move on to the next rung of the academic career ladder. (Despite the fact that [[There is an increasingly wide gap between the number of graduate students training for PhDs and academic positions for them to fill]].) While a tenured professor might be willing to work on something crazy that will take 10 years and might be a complete dud, no grad student would rationally work on it.  The incentive for PhD students to produce high-impact papers for the sake of their careers serve to amplify all the other constraints, and add the timescale constraint. Additionally, PhD students in science get most of their funding via research grants, so research they work on is especially susceptible to #4. 

6. **Projects that produce a lot of structured data or reusable code don’t led themselves to papers.** Academic papers are built around the paradigm of scientific inquiry, where collecting data and writing code are only in service to producing an abstract model or theory - the more generalizable the better. There is little reward to producing high quality, reusable datasets or code, despite the value they can create. 

7. **A project is unlikely to lead to a paper if it involves a lot of coordination.** Academia emphasizes individual agency. This is a good thing, but if a project cannot be effectively modularized it will end up requiring a lot of grungy coordination work that doesn’t count as a contribution unless you’re in charge of the project. A big reason to go into academia in the first place is to avoid having a boss and because you like doing your own thing, your own way. Obviously there are exceptions like the LHC, but that is a situation where the experimental particle physicists are locked into a paradigm and have no other option. 

Coordination aversion hits hard projects that don’t fit into a particular disciplinary bucket or require integration between components built in different labs. Projects of a particular scale require coordination by their very nature. See[[A new structure for scalable research]] for a much deeper dive.  

### What makes something not product-focused enough for a startup
‘Product-focused’ is meant as shorthand for ’things startups are incentivized to do.’ So things that are not product-worthy are those that don’t obey [[§Startup Constraints]].  It’s important to explicitly call those implications out because that call-out can shine a light on where to dig for hidden gold. 

1. **There’s no convincing story for how the project’s output produce a massive and competitive ROI on a reasonable timescale for a VC-funded startup** 
*Massive ROI* Venture capitalists need to believe that any investment in their portfolio could potentially return the fund. 
*Competitive ROI* Investing isn’t just about absolute returns - any investment is in leu of another investment. So if the stock market is absolutely crushing it, there’s less incentive to invest in higher risk technology. Even within the VC realm, many atom-based technologies have the same risk or higher than a software startup but expected return is lower (especially when you take timeline into account) or farther in the future. From the cold hard numbers SaaS is often just a better investment.  
*Timescale* [[J.C.R. Licklider]] started work on the ARPA project to build what we would consider “modern computing” in 1962 and perhaps the first real ‘value capture’ event based on it was Apple’s IPO in 1980. [[Venture capitalists need to fund things that increase valuation in a few years and return in 5-7 years]] so this trajectory would have been untenable as a startup. Venture timescales often lead to projects being pushed to focus on a specific application or being acquired before achieving the technical goals they set out to hit. There is nothing wrong with either of these routes per se, but they can lead to technologies failing to live up to their potential. 
2. **It will take a lot of development work to sell anything for a bootstrapped startup**
Startups do not *need* to raise money from VCs. However, that does mean that if the founders are not extremely wealthy ([[SpaceX]], [[Blue Origin]]) the startup needs revenue early on. If the project will take a lot of development work before it can sell anything the startup needs to get revenue somewhere. Some bootstrapped startups use consulting work to create revenue to fund development. Unless it’s wildly profitable, the consulting approach severely limits the amount of time and money that can be spent on development, possibly dragging it out forever. [[Some projects need effort above a certain threshold to move forward]]. [[SBIR]] grants can provide another option, but they are relatively small (order $1m over several years.) 
3. **It’s hard to capture the value the project creates**
Markets are great, but there is no law of economics that people can always capture the value they create if they are clever enough. (See [[Schumpeterian Profits in the American Economy - Theory and Measurement]].)  Research often falls into this trap because [[Innovation requires multiple people]]. Often the person who figures out how to make a lot of money off of a discovery or invention is not the person who created it, nor are they even in the same organization.  “But what about intellectual property ?” There are *many* reasons intellectual property does not ensure everybody involved in an innovation capture the value they create but a short list includes: many people in many organizations contribute to most innovations, patents only work for a narrow set of innovations, and patents have massive downsides - including potentially capping how much value an innovation can create. 

A more subtle issue than the question of value capture itself is that trying to capture the value created by an innovation can severely hamstring the total value it creates. Consider the [[Graphical User Interface]] and other personal computing innovations created at [[Xerox PARC]]. Arguably they’ve created trillions of dollars of value for the world. Would that be the case if Apple had a patent on the GUI and mouse? Probably not. This situation is not unique. 

[[It is hard to capture value from research]]

4. **The project is changing one or more components of a bigger system or process** 
A startup that is trying to improve or change a component of a bigger system needs to either convince whoever runs that system to adopt the change (which is especially hard when no particular entity runs the system) or the startup needs to build the entire system themselves with the change incorporated. 

Obviously, the change needs to happen eventually, but many paradigm shifts (especially in complex systems) initially *lower* performance, at least on traditional metrics. [[Metrics can cause paradigm lock-in]]. Change happens either when the new paradigm gets enough reps to catch up to the old way of doing things or people see the new paradigm in action and accept that the traditional metrics weren’t capturing all the important features of a system. Both of these situations are hard for a traditional startup to achieve because both approaches take a lot of time and resources before showing ‘traction’ and the potential customers (the people running the systems or consuming their outputs) will assert that the idea is dumb up until its not. 

See [[fillerFundamentalManufacturingProcess2020]] for more.

5. **Innovations that can’t be encapsulated into a product**
There are some innovations that it’s just hard to productize. Process innovations often fall into this category - especially those that don’t depend on new technology per se, but on a new way of doing the same thing with basically the same tools. The [[Bessemer Process]], for example, used basically the same crucibles but you put different things in them. Social process changes (like Kaizen[^1]) are even harder to productize. 

In these situations, the vast majority of the value generally accrues to the end-product producer or consumers. The naive advice would be “start a firm that’s doing the end-to-end thing” and outcompete the incumbents. While that should theoretically work, it often runs into the reality of extremely complex products that you would need to reinvent from the ground up (like if you had a better way of making passenger jets), heavily regulated industries where it would take massive lobbying just to use a new process, and other scenarios. A startup doesn’t usually have the time or resources to tackle these scenarios. The alternative approach is to take on a change-management consulting role, which usually doesn’t capture enough value to be VC-fundable so it needs to be profitable early on. Additionally consulting runs head-long into the issues in point #4 above. 

In some ways this point is intertwined with #4. 

6. **Startups shouldn’t tackle market, channel, AND technical risk** 
New technological paradigms usually come with all three of technical, market, and channel risk. You simultaneously need a lot of work to get the technology to a point where it’s competitive with alternatives, to figure out who wants to use it, and to establish how they’re going to buy it. 

It’s common wisdom among VCs that it’s not a good idea to invest in companies that are taking on two, let alone three of these risks. This heuristic isn’t misguided.  The chance of failure goes up *dramatically* when an organization tries to tackle tow, let alone three of those risks. Startups should only tackle one (or *maybe* two) of these.  The reason therapeutics focused startups can exist is that they have almost no market or channel risk - there is a well established pipeline to being acquired by a pharma company and insurance companies are guaranteed to pay for drugs that go through FDA approval and target big conditions.

At the end of the day, the work you need to do to drive adoption of a technology is often very different from getting it to a certain performance level. The work priorities drive the type of organization you create. As much as we like to think that startups are about building technology, they are actually about selling technology and adopting it for specific markets so that they can do that better. 

### Related
- [[§Academia Constraints]]
- [[§Startup Constraints]]
- [[§Corporate R&D Constraints]]
- [[§Philanthropy and Non-Profit Constraints]]
- [[Systems research requires a lot of work that is expensive and uninteresting]]

[^1]: [[Kaizen is Toyotas process for continuous business process improvement]] 



[Web URL for this note](http://notes.benjaminreinhardt.com/Things_that_are_not_paper-worthy_enough_for_academia_and_not_product-focused_enough_for_a_startup)

[Comment on this note](http://via.hypothes.is/http://notes.benjaminreinhardt.com/Things_that_are_not_paper-worthy_enough_for_academia_and_not_product-focused_enough_for_a_startup)




Program design is a coordination tool

There is a tradeoff between an organizational culture that is good at addressing market uncertainty and one that is good at addressing technological uncertainty

Sarah Constantine conversation 8 Jul 2021

How to think about the differences between FROs and private ARPA programs

Jerry Neumann conversation 1 July 2021

Week of 11 April 2022

General-purpose telerobotics needs work that’s not suitable for a for-profit company

Apoorva Mehta conversation 10 Feb 2022

Software startups can get away with doing research

Week of 30 Aug 2021

14 Jul 2021

7 Jul 2021

9 May 2022

Derisking is actually deuncertaintying

# When should an idea that smells like research be a startup?
# When should an idea that smells like research be a startup
*Notes about language.*
1. I’m going to be trying to use the word ‘uncertainty’ in the technical sense of* [[Knightian Uncertainty]] *that carries a lot more meaning than just “we don’t know.”
2. I’m referring to things as “ideas” (in the [[Activity Space]] sense) to avoid “projects” “programs” etc. because whether it looks more like a project or a program ([[The distinction between areas, programs, and projects]]) probably has something to do with whether it makes sense as a startup.  
---

We can say for certain some things about ideas that smell like research:
0. To some extent they are attempting to do something, usually discovery or invention-wise, that nobody has done before. In other words, they do not have a global [[Critical path]]. [^1]
1. They will not be able to generate cash flow directly from their main activities for some amount of time.  
2. As a result of #1, the activity will need some kind of money factory.[[Innovation orgs need a money factory]].
3. That money factory (or multiple money factories) will need to put money into the idea for some amount of time before it produces a ‘working’ result — a good business, a technology that other people can use, an acquisition etc.  
4. There is unpredictability about the timescale to get to a ‘working’ result. More uncertainty about critical paths means more unpredictability on the timeline, or even the timescale. (Keep the fact that [[Mismatched Buxton Indexes lead to shear forces]] in mind.) 
5. In large part because of the timescale unpredictability, there is unpredictability about the amount of money the idea needs to get to that ‘working’ result. Cost is often tightly coupled to the amount of time it takes. Hence, more critical path uncertainty -> more time unpredictability -> more cost unpredictability.
6. The time and money are necessary to address technology risk (a chance that the technology won’t hit a performance metric) and technology uncertainty (not knowing what the metric actually is). 
7. Researchy ideas have some amount of uncertainty about potential outcomes: what the form of the output will be, whether there is a market for that output, whether the output will be of a form that can capture value from that market, whether the program will organizationally be able to be a business that can do that value capture. 

If the idea is housed in a new for-profit organization, ￼#1+#2+#4 means that except in rare cases, the organization needs to be a high-growth startup because in order to provide competitive time-discounted ROI on the money the company needs to do the work over a potentially long timescale, it needs to target massive returns. Being a [[Grahamian startup]] then subjects the organization to [[§Startup Constraints]]. 

One could imagine situations where a for-profit research program could avoid committing to be a high-growth startup. A for-profit could get away with targeting lower growth/eventual value if funding comes from non-dilutive grants or ‘friendly’ investment where the investors don’t expect to make as much return on their money as they could elsewhere. However, both of these options come with their own downsides. These downsides include needing to keep investors happy because ‘friendly’ investment generally means ‘not market competitive’ so investors become much less replaceable. The danger is that the product becomes ‘things that make investors or grant-givers happy’ rather than things that maximize the potential of the idea and push the [[Knowledge frontier]].  (Or even make the most money in the long run.) Friendly-funded for-profits also experience a lot of pressure towards efficiency, and [[Efficiency biases systems towards false negatives]]. (Of course this efficiency bias is a cultural thing and not tied just to profit-seeking.)

Outliers dominate both the results and these conversations so I’m going to call out the elephants in the room:
- [[SpaceX]] (it’s even debatable whether this counts as an idea that smelled like research — at first they weren’t even trying to do something nobody had done before — there was a pretty clear but very hard critical path). 
- [[Genentech]]
- A number of therapeutic companies
- [[OpenAI]]
- [[DeepMind]] 
The fact that the list is so short compared to the list of researchy for-profits that have failed to deliver on their promises seems indicative. The ‘failure rate’ is even higher than the average for startups. This list also points to the fact that the question might be very discipline dependent. Only SpaceX is the non-therapeutic, non AI company on the list.  

At the end of the day, no set of criteria can definitively say “this project shouldn’t be carried forward by a new for-profit organization” except maybe if desired output could never be sold. (Though even then someone could point to profitable but unintended discoveries as a reason for researchy work to be a startup). The nature of [[Knightian Uncertainty]] means that it’s certainly possible to flip all heads in a row on a large-dimensional, non-stationary coin and pull off a successful research startup. However, I would argue that the more the unpredictability in each of the steps smells like uncertainty instead of risk ([[Uncertainty always involves risk but risk does not always involve uncertainty]]), the less it’s a good idea for the project to be carried forward by a new for-profit organization. 

A decision tree may be the correct way to look at the question of whether a program is too researchy to be a startup. These questions are ordered roughly in terms of importance, with “no” answers strongly indicating that a start:
- Can the desired output be sold?
- Is there a global [[Critical path]]?
- Is there an objective ‘measure’ of what ‘working’ means? How smooth is it?
- Is there a bounded amount of market uncertainty?
- Is the technology a modular product (ie. does it slot into existing systems)?
- Would investors consider a low ROI acceptable? 
- Would investors be willing to wait a long and unpredictable amount of time to get that return?
- Are investors ok with an illegible process without clear milestones? 

### Can the desired output be sold or unlock a product?
If you are setting out to create something that could never be sold, the program probably should not be a startup. There are examples of accidental but profitable discoveries that happened in the course of pursuing something out of motivations besides profit — discovering a heart drug while studying tree frog poison in the amazon for example — but you might as well just buy stocks chosen by a random number generator. 

### Is there a global critical path?
No global critical path means that you need to run a “fat” process without clear justification. (Justification requires explaining how it fits into a bigger picture which requires a global critical path). Without a global critical path, “progress” is nothing but a narrative and the fat process makes it look like you’re doing a lot of dicking around. As a result, investors will get fed up at some point and either force the research into a premature product or shut the whole thing down. 

### Is there a smooth objective measure of what ‘working’ means?
One way to look at uncertainty is that there’s no definition of what ‘works’ means: this could be in terms of the technology itself (technology uncertainty) or in terms of how it becomes a good business (market uncertainty). Note that this is different from technology or market *risk* where you know what winning looks like but there are a lot of reasons why you wouldn’t get there. [[Uncertainty always involves risk but risk does not always involve uncertainty]]. Technology risk sounds like “there’s a chance that we can’t get the efficiency of this process above 35% but the back of the envelope says there’s nothing that should prevent that and we have a development plan to get there.” (A metric and a vague sense of a probability distribution over that metric). 

Creating new technology will almost always involve *some* uncertainty. Technology’s modular/combinatorial nature[^2] means that there will be uncertainty at some level of the hierarchy. Sometimes that uncertainty will propagate up to the top-level technology if it is an irreplaceable subsystem. However, my extremely hand wavy intuition is that [[You cannot have massive uncertainty in subsystems without it propagating upwards]]. 

Even if you have a clear measure of what ‘working’ means, the way in which that metric changes matters. A smooth metric means that incremental changes can cumulatively move the needle (which means that you can show nice graphs trending towards success). However, some researchy ideas involve situations where the needle doesn’t move at all until it does a massive jump (ie. They depend on breakthroughs).  

### Is there a bounded amount of market uncertainty?
Situations where you’re guaranteed a massive market if you can hit a clearly defined target lend themselves to researchy startups. Theraputics are a prime example here[^3]: “if you can create a cancer drug that passes FDA approval, you are guaranteed a $B market.” In the past, mainframe computers had a bit of this flavor — if you could hit a clear performance metric, you had a guaranteed market. [[It is possible that healthtech and therapeutics in the early 2020s is to biotech what calculation and mainframes were to  computers in the 1960s]]. 

A big reason that market uncertainty can kill researchy startups is less about investment and more about the fact that [[There is a tradeoff between an organizational culture that is good at addressing market uncertainty and one that is good at addressing technological uncertainty]]. At some point a researchy startup needs to do a dramatic gear shift into growth and product-market fit mode. This transition often either prematurely kills the research potential or the company dies because it’s being run by people with a research mindset. 

### Is the technology a modular product (ie. does it slot into existing systems)?
[[Startups are good at point changes]]. If there is a component in a system that can be improved or swapped out in order to make the system cheaper/more profitable/more efficient, startups are an excellent mechanism for making that happen. The same goes for projects in industrial research labs — if you can, say, save the company a ton of money by swapping out the sheathing on telephone cables, the path to success is straightforward. However, to improve, many systems need many simultaneous changes (or to be swapped out for a new process entirely[^4]). 

[[Systems must take performance hits to get out of local optima]] and people are rarely willing to pay to *decrease* system performance. (This is a major contributor to [[The (idea) valley of death]]).

See also [[Systems research requires a lot of work that is expensive and uninteresting]]

### Would investors consider a low ROI acceptable? 
A lot of research involves dicking around, or “fat” processes. If you need to target a high ROI from day one, you need to be as efficient as possible. ROI is all about efficiency, and [[Efficiency biases systems towards false negatives]]. There will just be less [[Slack - concept]] in the system, cutting off lines of work that can’t be justified up front (especially if they require resources to evaluate). 

### Would investors be willing wait a long and unpredictable amount of time to get that return?
A key characteristic of research is that regardless of how long it takes, that amount of time is *unpredictable*.[^5] Combined with the fact that progress is often not smooth, it takes a constitution of steel for investors not to start getting antsy at some point. People are patient until they’re not. The pressure to show some progress towards a progress makes sense from a business sense but can be at odds with developing a powerful technology. Premature specialization is a common failure mode in researchy startups. ([[Ideally corporate R+D enables work on a general purpose technology before it’s specialized]]) .

### Are investors ok with an illegible process without clear milestones? 
In addition to unpredictable timelines, a lot of technical research involves constantly shifting targets. You can set milestones (and it is often useful to do so), but they need to be incredibly fluid. Unlike business progress, which can be boiled down to a few core metrics like customers or revenue regardless of how much pivoting is happening, research progress is much harder to boil down to a legible metric.  

Some questions that are proxies for the degree of technological uncertainty facing a project:
**How smooth is the technology’s fitness landscape? That is, how much could a single change shift outcomes?** 
[[The smoothness of a technology’s fitness landscape has a big effect on a program’s technological uncertainty]]
**Do you know what it means to improve?**  

**Do you know how you will know that you know how to incrementally improve the system?**

### Related
- A specific subset here is  [[General-purpose telerobotics needs work that’s not suitable for a for-profit company]]
- [[What does ‘working’ mean?]]
- [[Uncertain things basically have infinite risks]]
- [[Knightian Uncertainty]]
- [[There might be many situations with intractable uncertainty or randomness in the highest order bits]]
- [[§Startup Constraints]]
- [[Jerry Neumann conversation 1 July 2021]]
- [[Research is especially susceptible to Goodhart’s Law]]

### Braindump
- One trick is that the non-stationarity of the coin flip means that if the people on the program know that they only get a set number of flips before they produce a working thing
- [[Startups are the upper left corner of the divergent-convergent lean-fat 2x2]]
- Are we allowed to throw out “what about Tesla/SpaceX”
- There needs to be a lot of excess value creating from the research work 
- [[Startups need to conform to investor timescales]]
	- [[Venture capitalists need to fund things that increase valuation in a few years and return in 5-7 years]]
	- The average stock market return is 9.2% 
	- Average over 2010-2020 13.6%
		- $1->$6.77 over 15 years
		- In order to beat that by even 2x a startup would need to turn $100M of investment into $6.8B returned to investors, so valuation would probably need to be $10B
- Startups working on general purpose technology tend to get sidetracked
	- Magic Leap
	- 
- Manufacturing a physical thing that has never been manufactured before has not led to great results outside of electronics 
	-  [[Startups need to produce a scalable product]] 
		- “But what about Tesla and SpaceX” 
- Need an actual breakthrough, as opposed to just grinding on current paradigms 
- TRL of the overall system
- There’s also the point that [[Organizations get good at one thing]] — organizations that are optimized for doing exploratory research and finding critical paths 

[^1]: In other words, they are not in the lower left quadrant of  [[raoGUTSGrandUnified2018]]
[^2]: See [[arthurLogicInvention2005]]
[^3]: See [[Therapeutics have established sales channels]]
[^4]: See [[fillerFundamentalManufacturingProcess2020]]
[^5]: See [[narayanamurtiCyclesInventionDiscovery2016]] 

[Web URL for this note](http://notes.benjaminreinhardt.com/When_should_an_idea_that_smells_like_research_be_a_startup)

[Comment on this note](http://via.hypothes.is/http://notes.benjaminreinhardt.com/When_should_an_idea_that_smells_like_research_be_a_startup)






Benʼs working notes

§Startup Constraints

Links to this note