There’s a subtle distinction between freedom to suggest ideas and freedom to pursue ideas

# §Bell Labs Model

### Questions
- [[Why was Bell Labs Different?]]
- [[Why hasn’t Bell Labs been replicated?]]
- [[Was Bell Labs an anomaly?]]

### Conclusions
- [[Bell Labs and AT&T had timescale alignment]]
- [[Bell Labs is one of the few examples of an innovation org that was aligned with its money factory]]
- [[Bell labs had a budget of 2.8 billion 2020 dollars]]
- [[Bell Labs was only profitable in conjunction with AT+T]]
- [[AT+T benefitted significantly from creating better technology]]
- [[Regulation drove many of Bell Labs’ actions]]
- [[All the work at Bell Labs needed to tie back to the system]]
- [[Bell Labs empowered young people]]
- [[Bell Labs hired people before they had a role or expertise]]
- [[The work at Bell Labs that was most valuable to AT+T and the work that was most valuable to the world was fairly disjoint]]
- [[The famed bell labs research was a fraction of their activity]]
	- [[Bell labs was primarily not researchers]]
- [[Bell labs brought many disciplines together under the same roof and broke down barriers between them]]
- [[Bell Labs did not give researchers complete freedom]]
- [[Bell labs asked people to do useful things but left them room to say huh that’s funny]]
-  [[Bell Labs management was extremely light]]
- [[Bell Labs didn’t have a unique model]]
- [[Bell Labs selected for people with only academic pedigrees and with a sense of adventure]]
- [[Bell Labs was not a business]]::NW::
- [[Bell labs existed before continuous improvement was expected]] ::NW - stub::
- [[Bell Labs was a dance of individuals and institution, prototyping and manufacturing, invention and discovery]]::NW::
- - [[Bell Labs enabled free radicals]]



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

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




Bell Labs

gertnerIdeaFactoryBell2012

Bell Labs hired people before they had a role or expertise

The good idea fairy is the failure mode of a single decision maker

§Bell Labs Model

Labs Document Outline

# Bell Labs empowered young people
“Empower young people” is another one of those admonitions like “have light management” ([[Bell Labs management was extremely light]]) that sounds uncontroversially good, but is both hard to do in practice and easy to screw up. 

Bell labs’ tendency to empower young people is illustrated by [[Jim Fisk]] being in charge of radar/magnetron work at age 30 in the middle of WWII. 
![](../BearImages/662F1380-98EC-4427-A641-F5F686CE2269-4422-0007B95693CDD5A5/IMG_0144.jpg)

Empowering lightly-or-unproven people can clearly backfire, so doing it depended on a substrate of trust - both upwards, that higher management wouldn’t second-guess the decision and downwards, that the empowered young people could handle it. Where did that underlying [[Trust]] and conviction come from? It’s not clear to me, but getting to the root of it seems important for riffing on Bell Labs.  [[Research requires more trust than other disciplines]].

Keep in mind that empowering young people was probably not a uniquely Bell Labs thing ([[Bell Labs didn’t have a unique model]]) but a broader cultural trend in the first half of the 20th century. However, [[Mervin Kelly]] was even willing to promote young men *at the expense of* veterans, which is doubly risky and vaguely violates other cultural norms like the idea of loyalty being rewarded. 
![](../BearImages/55AED6C9-A8DD-4602-9607-405B596C2992-4422-0007B966D9720770/IMG_0148.jpg)
![](../BearImages/12DF4EE9-39E5-4C58-BBFE-56D91A675344-4422-0007B96859A81B8A/8BCC0E41-9CD4-41D2-9B8D-CD10BAEB8D73.png)

Also note that it was empowering technically-trained young people to do things that were in or adjacent to their areas of training. This stands in contrast to what feels like a more 21st century Silicon Valley take on “empower young people” which is to say “why shouldn’t we bet on a 22 year old with a philosophy degree to start a drug company. Empower Young people!” [[People who have done a thing should be in charge of a thing]]. 

### Related
- [[The more trust an institution has, the less it needs formal process]]
- [[Bell Labs hired people before they had a role or expertise]]
- [[Research requires more trust than other disciplines]]



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

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




When is push or pull appropriate for innovation orgs?

Bell Labs empowered young people

Comparing and contrasting Bell labs, DARPA, and Venture Research

# Bell Labs hired people before they had a role or expertise
Throughout its history, there are examples of people being hired because they were smart and had experience vaguely in the area of electrons, etc. On the other hand, they seem to consistently hire people with good *PhD* academic pedigrees. [[PhD academic pedigrees are higher signal than most undergrad pedigrees]].
![](../BearImages/738E9E8A-6A50-4FDC-9A8C-E30CB3A8DFA6-4422-0008049E8C06D111/IMG_0136.jpg)
This strong pull recruiting was complemented by, in some situations, enabling people to do a rotation and figure out which group they wanted to work with.  

This style of recruiting is very much in line with the “pull” approach. [[When is push or pull appropriate for innovation orgs?]]. In terms of ideas going in, Bell Labs required almost none, DARPA requires some, and Venture Research requires extremely well thought-out ideas. [[Comparing and contrasting Bell labs, DARPA, and Venture Research]]. It requires a lot more evidence, but this pattern suggests that the longer the ‘gestation’ someone has in an organization before they need to be productive, the more you can get away with getting just “good people.”  This conclusion might seem banal, but it’s important for getting organization design and expectations right, and people often screw it up. If you’re going to hire people before they have a role or expertise, you need to have appropriate internal structures in place. Similarly, if you *don’t* have those structures in place it’s important to have a good external-idea vetting process.

### Related
- [[Bell Labs empowered young people]]



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

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




Research requires more trust than other disciplines

For a certain class of research, deliverables can be detrimental

Bell labs brought many disciplines together under the same roof and broke down barriers between them

Bell Labs enabled free radicals

Bell labs asked people to do useful things but left them room to say huh that’s funny

All the work at Bell Labs needed to tie back to the system

# Bell Labs management was extremely light
Bell Labs management gave researchers autonomy, both to work on what they thought was best and with whom they wanted. While on paper this is what most management claims to do, in practice that’s rarely the case. So it’s worth digging into what light management at Bell Labs tactically entailed and perhaps how they were able to walk the walk instead of talk the talk. 

![](../BearImages/FBB01D7F-FB72-4B5D-A54A-87F6D9F98C8A-88049-0007B3CAFC68AFC1/IMG_0206.jpg)
Bell Labs had explicit policies that prevented managers from controlling who could collaborate with whom. My hunch is that the policies were more a codification of culture than anything else and someone citing the policy and telling their boss ‘you can’t stop me!’ was rare. However, it’s telling that unrestricted collaboration was part of the culture worth encoding. This managerial getting out of the way was one of the ways that [[Bell Labs enabled free radicals]]. 

![](../BearImages/358508FB-EFA0-44E8-997D-0CE7B250A8CA-88049-0007B3C5E0753C7D/IMG_0158.jpg)
`A believer in granting a degree of autonomy to researchers, he had not asked about, and had not been kept appraised of Bardeen and Brattain’s work. What’s more, there was a tendency at Bell Labs to confine important developments to middle management for a purgatorial period.`  
Bell Labs management did not put pressure on researchers to produce demos and there was an explicit time delay built in as the demos moved up the management chain. This approach — basically the inverse of the [[Media Lab]]’s “demo or die” — is worth calling out because expectations around demos have a strong effect on organizations that isn’t discussed enough. In some ways demos are valuable because they put pressure on people to get their work into a form that’s legible to people outside of their thought bubble. On the other hand they can often lead to creating “movie set products” where they are basically shiny shells that can do the demo and nothing else. Bell Labs’ management seems to have walked that line well. In order for the purgatorial period to work, moreover, the middle managers need to have in turn not felt pressure to surface demos in order to advance their careers. 

[[John Pierce gave four reasons why Bell Labs worked]] and `A technically competent management all the way to the top` is the first item on that list. I suspect that the technical competence helped the management stay light because it enabled trust between the managers and researchers. [[Research requires more trust than other disciplines]]. This trust meant that there was less dependence on metrics and milestones. The trust, and its resulting light management and efficiency is why [[People who have done a thing should be in charge of a thing]]. 

Interestingly, the management style at Bell Labs actually seems similar to [[JCR Licklider]]’s management at ARPA - find awesome people, get on the same page about the precise problems they’ll go after, and then let them rip. [[DARPA does multiple levels of top-down problem generation and bottom-up solution generation]]. 

At the same time, we shouldn’t hero-worship Bell Management.  Pierce may also have just been a bad manager and light management just kept him from doing damage….
`The brunt of his management work [at Bell Labs] had consisted of dropping in, unannounced, on colleagues in their labs to ask how work was progressing.`
![](../BearImages/A9B10EF2-FDC0-4C59-BA09-8A12263E8E2E-88049-0007B5D6B1243317/IMG_0253.jpg)

They still committed classic management false negative (preventing someone from doing something useful.) [[Types of management error]]. [[Gordon Teal]] still had to do the classic “work overnight on the side project” move to show that he actually could create high-purity chunks of germanium. ![](../BearImages/D5380E7F-E486-4A6B-991D-ECC088E27D58-88049-0007B4AFC620D39D/IMG_0169.jpg)

And of course, one has to hold the simultaneous views that while Bell Labs’ management was light, [[Bell Labs did not give researchers complete freedom]].  [[All the work at Bell Labs needed to tie back to the system]]. My current mental model for integrating these seemingly contradictory is similar to a coconut - high-resistance exterior and low resistance interior. [[Bell labs asked people to do useful things but left them room to say huh that’s funny]]. 

### Related
- [[Venkatesh Narayanamurti Podcast 20 May 2020]]
- [[odlyzkoDeclineUnfetteredResearch1995]]
- [[Andrew Odlyzko Podcast]]



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

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




Life Style Guide

What traits do high-potential PMs have?

Organizations run by practitioners have less bureaucratic bs

Pieces of the Action Forward Intro Notes

Successful professors are not necessarily the best researchers

Bell Labs management was extremely light

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

PARPA needs a constitution with checks and balances

People do fundable work

fuEnablingMissionImpact2021

Explicit rules are a substitution for learned pattern matching

Where is the best physical place for PARPA?

Bottling Lightning

Fingerspitzengefühl — Finger Knowledge

The people setting research priorities at AFRL are often not technical

The class of unskilled abstraction weavers

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

Tool accessibility is underrated for heuretic development

ICT Distillation

Tacit knowledge about the tradeoffs inherent in drivers of change is important for well run governments

ICT Outline

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

Job titles should reflect different levels of responsibility

Short term feedback loops are hard when you need to be doing two things at the same time

# People who have done a thing should be in charge of a thing
*I’m going to focus on research because that’s the primary concern that’s driving this note and I think that’s a better cruising altitude on the ladder of abstraction. However “people who have done a thing should be in charge of a thing” extends to almost everything.*[^1]

[[Understanding a thing has to do with intuiting the affordances of a thing]]. It’s extremely hard to build up intuition for a thing without having done it. So at the end of the day ‘having done a thing’ is actually a heuristic for ‘has intuition about this thing.’ 

This heuristic is important to call out because it is not always true! People who have done a thing do not always have intuition for a thing and people who have not done a thing can sometimes build intuition for a thing! As far as I can tell, the only way to tell whether someone has an intuition for a thing is a track record of things that could only be accomplished through having an intuition for the thing or if you have intuition for the thing yourself. It’s [[The Map is not the Territory]] problems all the way down. So while ideally you could know someone’s intuition directly, the probability of intuition in someone who has done the thing is high and the probability of intuition in someone who has not done the thing is low. 

In research especially, there are very few metrics you can use on their own distinguish whether a project is a good idea or a bad idea or how it is progressing once it’s started. [[Many things cannot be measured well]]. Even in the case of “everything is working great!” it is easy for researchers to fool themselves, optimize for the wrong thing, or overlook a core problem.  

“Being in charge of a thing” is a rather fuzzy concept. I think what I’m grasping at is along the lines of “controlling the fate of thing” - this includes deciding whether a project/person gets approval funding, how well it’s doing, coming up with ideas for what should be done in the first place. Abstracting: the levels are binary go/no-go, progress measurement, and creation. You need different levels of intuition for each of these - they’re ordered from least intuition to most. 

At some point someone needs to be in charge and they can’t have done everything, so how do you both have a [[§Trusted Hierarchy]] and stick to the assertion that people who have done a thing should be in charge of a thing? Delegation! 

The “gotcha!” comes in determining what counts as ‘having done the thing.’ Does someone who worked in a biology lab doing experiments on rat brains count as someone who has ‘done’ biology? Should they be in charge of a molecular genetics project? This is where you need to defer back to the fact that the actual [[Highest order bit]] is 

Of course, the contrapositive also holds:
`People who have not done a thing should not be in charge of a thing`  Specifically people who have not done research should not be in charge of research

A corollary to this is that a PM’s (project, program, or product) power over a thing should be attenuated according to how much they’ve done the thing they’re PMing over. 

In terms of investing at least, a counterpoint brought up by [[Sarah Cone]] is that sometimes someone will overspecialize and end up being *less* qualified to be in charge of a thing in an area that’s normally quite close. For example, someone with a PhD in molecular biology but doesn’t understand CRISPR applies an old mindset to a new technique and isn’t willing to update their priors. While on the other hand, an excellent, applied, and diligent person may be able to be better at investing at least.  

### Related
- This tweet/video is an amazing example of what happens when someone who has no technical experience with a thing being in charge of a thing https://twitter.com/chittimarco/status/1338810527565361152?s=12
- [[Venkatesh Narayanamurti Podcast 20 May 2020]]
-  [[Organizations run by practitioners have less bureaucratic bs]] 
- [[Legible games enable people outside of an institution to control it]]
- [[Institutions need to differentiate between individuals inside and outside the institution]]
- [[Book Review - Seeing Like a State]]

[^1]: Of course there are exceptions. I suspect the strength of the assertion varies with the dimensionality of the thing. 



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

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




Uncertainty always involves risk but risk does not always involve uncertainty

Heilmeier — Licklider — Braben — Elon quadrants of Research Management

As a platonic ideal, research is an unscheduled quest for new knowledge and inventions with hard-to-predict outcomes

A system of credit and funding for non-paper artifacts could be a trust network

Bureaucracies remove the need for trust

Opacity is important because it allows people to be open about failure

Tensions are the core theme of the PARPA writeup

Research orgs don’t scale

Introducing Speculative Technologies Outline

Research requires intrinsic motivation

Frontier-pushing technical people will be willing to work with a new organization

Quantized Heilmeier Catechism

How much does the reproducibility crisis have to do with low trust and tacit knowledge?

riconFundPeopleNot2020

Decreased trust is crippling research

Purpose behind ICT Writeup

Venture research built trust by iterated meetings to refine precision around the idea

 two decades of public engagement around Nanotechnology

Leo Christodoulou meeting 23 Nov 2020

Research projects have a halting problem

Science should be less open unless you want to dig deep and engage

# Research requires more trust than other disciplines
The uncertainty inherent in research means that there are only a few people (if any) who understand the underlying outcome distribution of doing the work. [[Uncertainty always involves risk but risk does not always involve uncertainty]]. Unless you are one of those people *and* you have the resources to undertake the work ([[We should strive to unbundle Elon Musks Brain]])  the person with the resources needs to trust the person with the underlying distribution. 

Abstractly, research requires doing weird things by its very nature. In order to create things that nobody has created you need to do things that nobody has ever done. Research often involves long periods without legible artifacts and sudden direction changes when you realize that one line of inquiry is no good or another opens up.  It requires trust to be comfortable with someone doing all those things while you’re paying them or your reputation rests on them. This pattern is present in other disciplines, but in startups the output is more legible ($$$) and in art, the stakes are usually lower.

Research also requires trust around the quality of the work and results. While theoretically a paper should clearly communicate everything a researcher did and enable you to replicate their results, anybody who has done research knows this is basically BS. There are so many ways that research can be misleading. This isn’t about fabricated data and outright lies (though you do need to trust that too.) Results can be exaggerated, conclusions can subtly not follow, experiments can be cherry-picked, a million regressions can be tried before one works, constraints can be ignored, etc. Many of the subtleties that make the difference between bad and good research are not even conscious on the part of the researcher. There is no way to enumerate and check for all possible problems so it comes down to trusting that the researcher did good work.  

Empirically, you see over and over again that high-output research organizations have higher-than-normal levels of trust. DARPA is full of stories where someone walks out of a ~20 minute meeting with millions of dollars. That can only happen because of trust. [[Bell Labs management was extremely light]] and the specific stories make it clear that it was light because the managers trusted the researchers to do good work. Highly-productive “scenes” like the physics community in the early 20th century were small and everybody knew each other.  

Legibility is one way that people avoid the need for interpersonal trust. [[Legibility enables action]] but when you’re trying to create or discover something new, it’s hard to know what to even measure and prematurely imposing measurement often leads to negative effects. [[Many things cannot be measured well]].

Arguably, the terrible effects of peer review ([[Peer review has failed as an institution]]) came out of a desire to replace trust with legibility in research, or at least make that trust more legible. [[Trust is hard to scale]], and as the number of scientists and the amount of funding increased, funders and other scientists needed a legible way to judge the quality of research. 

[[Andy Matuschak conversation 9 May 2021]] pointed out that research also requires much more *self* trust. By its nature, research is a very “lumpy” process that requires logical leaps and conversely will involve large stretches of time where it feels like one is getting nowhere. *And you might be right.* The inherent newness of good research often means that other people’s opinions and experience aren’t great guides. So you end up needing to trust your logic and intuition that you are heading in the right direction and working on something worthwhile. 

### Related
- [[Trust-building timescale and budgets tend to be similar in different domains]]
- [[Bell Labs management was extremely light]]
- [[As a platonic ideal, research is an unscheduled quest for new knowledge and inventions with hard-to-predict outcomes]]
- [[How do you have skin in the game around ideas?]]
- [[Opacity is important because it allows people to be open about failure]] - it requires a lot of trust to actually be open about failures 



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

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




Coauthors are committees

Alexey Guzey conversation 17 Aug 2022

# The good idea fairy is the failure mode of a single decision maker
“The good idea fairy” is what causes the commander to get it into their head that something is a good idea and nobody will tell them no.  There are many historical examples of this: Napoleon or Hitler wanting to invade Russia, the Kodak founder wanting to build an instant video thing. 

The good idea fairy is what decision by committee is designed to mitigate. [[Decisions made by committees lead to median results]], which *can be a good thing* in several situations. I suspect that the situations where you want a committee boils down to a bunch of [[Nassim Taleb]]-ian concepts: whether or not you’re in extremistan and whether there is infinite or capped downside/upside. In situations with capped downside and infinite upside committees are probably bad. In normalstan committees are probably good. In situations with infinite downside (COVID?) and capped upside the situation is more unclear. On the one hand a committee prevents Napoleon from invading Russia. On the other hand it also keeps the COVID vaccine from getting out for months?

I originally heard of this from [episode 13: eigenrobot vs Steven godofsky — robot friends — Overcast](https://overcast.fm/+pYCxXEXYw/46:44)

### Related
- [[Individuals vs Collectives]]
- [[Bell Labs was a dance of individuals and institution, prototyping and manufacturing, invention and discovery]]
- [[Ambitious individuals are not empowered to do anything besides start companies]]
- [[Bell Labs empowered young people]]
- [[People who have done a thing should be in charge of a thing]]



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

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




Legibility enables action

Legitimacy cascade

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

Trust is hard to scale

The more trust an institution has, the less it needs formal process

Trust is a core driver of the funding mismatch

What are the alternatives to grants?

Small grants are different animals from big grants

Faith based on desire works in human systems

There are many trust dynamics around introductions — Introduction readiness levels

Institution-builders need to build relationships directly with funders

The mechanism by which we trust institutions is the same mechanism by which we trust people

Social capital

People are hesitant to try things when they don’t know where the line is and the consequences are large

Academia judges via novelty because there are no metrics but there are no metrics when the thing is still research

One way of measuring the trust level of a relationship is how recently you talked to someone and what you talked about

Trust cannot be quantified but it can be taxonimized

There is a tradeoff between trad creditability and the ability to do something weird

Reputation is a tensor while status is a vector

Dating is an underactuated control system

# Trust
### Related
- [[§Misalignment between funders and nonprofits in research]]
- The Trusted advisor framework breaks trust down into (Creditability + Reliability + Intimacy)/Self-orientation (See [The Trust Equation: A Primer | Trusted Advisor Associates - Training, Workshops, Trust Education](https://trustedadvisor.com/articles/the-trust-equation-a-primer)) 

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

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






Benʼs working notes

Bell Labs empowered young people

Related

Links to this note