Why do we listen to people with no skin in the game?
I’m talking in the context of industrial programming. A few statements I believe to be true:
- If you aren’t doing it hands-on, your opinion about it is very likely to be completely wrong.
- Being influential doesn’t automatically make you right.
- Talking about coding and delivering software under real constraints are two very different things.
- Before clicking on a programming video, it doesn’t hurt to check the creator’s credentials.
Yet we keep listening to influential people without much skin in the game. Say your extremely AI-pilled director plays around with LLMs for an afternoon, builds a few throwaway applications with some prompts, and comes away convinced that software engineering is solved. Now they demand 10x productivity and won’t hear about organizational friction or unavoidable delivery delays. All the features were needed yesterday, and all of you peasants are falling behind. Just use AI to steamroll everything.
It’s easy for someone who doesn’t build software day to day to become enamored with LLMs. I use them a ton, but I also run into their limitations more often than someone who has one-shotted a todo app or automated a few recurring tasks.
I don’t pretend to know how to run a huge organization. So I’m not a fan of people who don’t build anything substantial day in and day out telling me how software should be written or claiming that it has already been solved. If you write software for a living, you still have to listen when your manager or director tells you how to do your work, unless you wanna lose the job. That doesn’t make them right, and believing it does can be harmful to your psyche.
To be right, you need to walk the walk and pick up the pager when the vibe-slopped software crashes at 3 am.
Another group is software influencers. Some of them love making blanket statements like:
- Everything should be rewritten in Rust
- I will never use a GC’d language like Go
- TypeScript should just die at this point
- Only boomers write C#
- Kotlin is Java from Temu
- Why you need XYZ tool
- Stop using X and use Y instead
- You are falling behind if you don’t do XYZ
Well, of course they do. How else would they keep up with the algorithm?
These people were in the right place at the right time. There is a lot of money in yapping about software. I can name a few channels run by folks who worked in big tech once and then turned that into a career in podcasting or talking about software.
I have learned a lot from some of the guests on these shows. Unlike the hosts, their guests are typically folks who built successful careers writing large-scale software. I have no problem listening to and learning from them. But take advice from the hosts with a grain of salt, especially if all they do is talk about software.
The same goes for people who tell you XYZ AI skills are all you need. Keep in mind that these are the same people who used to sell courses on hot JavaScript frameworks, then jumped on the blockchain and NFT bandwagon, and are now riding the AI wave. Imagine if your run-of-the-mill civil engineer got obsessed with tools and churned out sandcastles while yammering nonstop about how fast they could build them?
Of course they’d think rewriting everything in Rust is a good idea: they run a Rust consultancy, ffs! Swap Rust for any language or tool and you’ll see similar behavior.
These blanket statements rarely consider the fact that an organization picks a language based on the available talent pool in the local market, how quickly they can build a product with it, and, for startups, which language the founders knew when they started the company. Changing that later is quite difficult, and spending time on it is often the wrong thing to do. LLMs don’t change that.
Advice is cheap, so keep an eye on the incentives of people who hand it out wholesale. None of this is new, and if you have been in the industry for more than a few years, you probably know it already.
This makes me wonder: why do we listen to people who have no skin in the game?