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

The ‘best people’ usually know their market value so by default they can capture more monetary value by spinning out their own thing or going to academia where they can capture more glory. The exceptions to these may be when doing corporate research relaxes constraints.
This assumes People are professionally motivated by money, glory, autonomy, or impact.
Money
If an individual researcher is motivated by money and create something valuable, they can capture more of the value they create if they spin off a startup.^1 Therefore, researchers who stay in R&D labs long-term aren’t motivated by capturing value.
Glory
Researchers can capture more glory because companies have the incentive to keep many more things secret. In academia, the only thing that keeps people from publishing is fear of being scooped. In a company you want to keep the research secret until you are already in a position where you can execute on the research better or you have no interest in executing on the research. Therefore, you would expect researchers who are doing things relevant to the company’s core business to have restrictions on what they can do and therefore less publishing and glory-capture than if they were in academia.
One exception to the ‘glory-restrictions’ on business-adjacent research may be if the company feels like they’re the only ones who can take advantage of the research. This suggests that one of the reasons Bell Labs didn’t have a unique model may have been able to publish so much was because their monopoly position assured them they would be the only ones able to execute on the research. Google Brain may fall into this same category - they have an effective monopoly on search and video hosting.
Autonomy
The “something else” could be doing interesting or novel work that would otherwise require months of grant writing. <Months of grant writing is one of the things that makes people think twice about academia> However novel, interesting work is often not aligned with the parent org because the value of innovations is not always clear andBig companies cannot start small things. If research doesn’t fit into an existing business unit, only outliers are incentivized to keep it going which means that researchers in corporate R&D can only be autonomous for a finite amount of time.
Impact
Corporate R&D can potentially have large impact if you’re working on something that’s aligned with the main company. For example if you’re working on research that can become part of an existing Google product, it will affect billions of lives. However, to sound like a broken record because Expensive research needs to address an existential threat eventually at an organizational level to maintain support, anything research not connected to a core product will have a finite lifetime and probably won’t have a large impact.
One way that Corporate R&D can get the best people is if they provide resources that would be hard to come by otherwise - the massive compute at haoMessySecretiveReality2020 and Google Brain for example, or access to world-class labs at pharma companies.
The incentives to get the best people to stay may also be part of the reason why Corporate R+D orgs are expensive - you need to provide them with hard-to-get resources, which usually translates to $$. The other way to keep the best people is to make them feel exceedingly comfortable - call this a ‘local optima’ - which is also expensive.
(Of course Bell Labs is one of the few examples of an innovation org that was aligned with its money factory for very specific reasons)
^1: This is especially true in the world of software but may hold less true in areas with higher capital costs. Also depends on the company’s willingness to acquire. Are there ideas that create a lot of value for a big company but would make terrible startups?

Related

Innovation orgs need to be aligned with their money factory.

References

Web URL for this note

Comment on this note