Enabling technologies must be developed while doing serious work

In How can we develop transformative tools for thought? Andy Matuschak and Michael Nielsen argue that Tools for thought must be developed while doing serious work. This is true not just narrowly for tools for thought but broadly for any enabling technology. Processes and social technology is absolutely (and perhaps particularly!) included in ‘enabling technology.’^1 In the same way that it’s ineffective to develop a tool for writers without actually using it to write something I see many serious proposals for how to manage research that are divorced from actually managing research.

Developing a technology while doing serious work is more than just giving a tool a context of use <Tools need a context of use>. A context of use is necessary but not sufficient. You can deeply understand a context of use - perhaps you’re even a practitioner who once held the role you’re building the technology for - and still fail to develop it while doing serious work. This looks like creating a technology outside the main flow of the work being done and then running workshops, externally consulting, writing papers, or selling tools with tenuous connection to the outcomes of the work itself. Many people develop technology while working on ‘toy problems’ instead of serious work. Toy problems are great! They are an excellent way to develop intuition. However, technologies developed on toy problems alone are like elaborate battle plans drawn up in a war room - they rarely hold up to first contact with the gritty complexities of real work.

Developing a technology while doing serious work has two criteria. First, it means that the work is worthwhile regardless of the enabling technology. Second, ‘while’ doing serious work (instead of ‘for’ serious work) means that technology development is done inside the loop of the overall project. Handing someone technology — a new process or tool — for them to use in a serious project with the intention of getting feedback afterwards misses hordes of illegible gaps; this situation often ends up with technologies gathering dust in the corner with generic reasons like “it was too slow.”

Admittedly, developing an enabling technology while doing serious work is extremely hard. You need to run an additional feedback loop on top of the work itself - “is this helping? How could I improve it?” It is inevitably a distraction from the work itself! Either you need to inject an additional person with new skills into the project and increase coordination costs or, if the same person is both doing the work and building the technology, they must split their time instead of devoting all of it to the serious work. The potential payoff needs to be worth the productivity hit and realistically, it rarely is.^2 <It is hard to capture value from research>. This is why, historically, people have often created new enabling technologies under duress - especially when the existing ones absolutely would not cut it. Many modern research funding practices emerged from Vannevar Bush’s reorganization of the entire US research system that was only possible because of WWII. Modern systems engineering and project management was created in large part because Nike missiles kept blowing up and then NASA projects kept going over budget. Modern equity-financed corporations were the only way to raise the capital to build the new railroads. Arguably, coronavirus-spurred remote work is also an example. ^3

Of course, a crisis isn’t always needed to develop new enabling technologies. However, some of the ingredients of a crisis are indeed necessary - the slack to try unproven things, people willing to tolerate short-term inefficiency, and most importantly: a good reason to deviate from best practice.

Canonical media


^1:<Sidenote>I primarily have process and social technology in mind here but I believe that it does generalize.
^2:This is actually a very rational reason many people are resistant to changes in their tools and workflows. It’s not because they don’t like new things on principle. Instead, there is inevitably a learning curve for any new technology and people become very adept at working around even imperfect tools and processes so the potential benefit over the old system in practice is smaller than you would expect from the outside.
^3: This was written in early 2021, so we’ll see how well these words stand the test of time.