Conference notes : Nihilism and toothpaste?
This talk was as strange as the title I’ve given to the article. It was an improvised talk given by Romeu Moura due to another speaker being unable to attend, and probably, one of the more thought-provoking talks of the whole conference.
Why do hotels don’t have toothpaste? This was the question everything started with, and with an emphasis on the normal amount of time he had spent researching this, he was able to walk us through the thought process of challenging prescribed beliefs.
Descriptivism vs. Prescriptivism
Prescriptivism: Dictates how something should be used, enforcing rules of correctness and standard forms.
Descriptivism: Describes how something is used, observing and analysing actual usage patterns without judgment.
Why don’t we have toothpaste in hotels then ?, it is due to how “ratings” work, a list was made based on the best hotels (descriptive), but it has since been used as a single source of truth (prescriptive).
It’s easy to relate this to software development, how commonly as an industry do we treat things that come from observation and analysis, for example, the usage of micro-services for distributed software put in place by Netflix, google, etc as “dogmas”, “you must use micro-services to be scalable”
The same can be said about TDD, Agile, monoliths (bad), they are success stories that become the “standard”, it becomes true by consensus, but not by us continuously verifying that it stills stands in our current reality.
Sensation of Efficiency vs. Real Efficiency
He didn’t really go into much detail about this, but it’s something that struck to me due to its relation to AI.
As shown in https://arxiv.org/pdf/2507.09089, we tend to highly overestimate how big of a gain we get from using this type of tooling in our day-to-day.
I’m not sure if there is any new studies that have come out with a bigger sample size and newer tooling, but it still stays true that is complicated to measure developer productivity (and something that every project manager would love to be able to do), and that jumping to conclusions based on a simple study is not good.
But based on personal experience, I tend to overestimate the gain when I’ve not given it enough thought to wether the task is a suitable candidate for AI-driven development, or if it would be easier/better to simply do it by myself.
I feel like some nuances of AI usage in our workflow are very nicely explained in this article by Unmesh Joshi : https://martinfowler.com/articles/llm-learning-loop.html
What if the feeling of believing that we are being productive outweighs our rational analysis of the actual gains we receive from these tools ?
Nihilism
What does this have to do with software?
Nihilism can be defined as the avoidance of oneself, a way of avoiding suffering (Algophobia).
There are 5 behavioural patterns that, when used to avoid oneself, are considered signs of Nihilism :
- Self hypnosis
- Orgy of feelings
- Petty pleasures
- Mechanical activity
- Herd formation
And if I try to relate it to software (from what I understood) :
The hypnosis The technique or the tools we use to produce software (TDD, DDD, etc) become the goal, instead of the fact that we’re building a product that might either not be working or that we are not proud of.
The mechanisim We think the processes we follow (Agile, Domain Discovery, etc), absolve us from the outcome of our decisions “I followed the rituals; therefore I’m not responsible for the consequences”
The herd Safety in the “common” way, if everyone is doing micro-services, if we all fail while doing it, no one is individually to blame.
At the end, each is a way to avoid ourselves and the fact that we might fail or not be happy with the output of our work/life.
Conclusions
The hotels will still not give you toothpaste, but next time you find yourself in a conference don’t hesitate to go to that “weird” talk that doesn’t sound as productive as the rest. It might actually be the one that ignites something inside of you.
Open questions
- When does descriptive wisdom become valuable description?
- On algophobia, are we afraid of suffering, or afraid of judgement?
- Is the solution for this behavioural patterns individual (mindfulness) or structural?