- > And then there’s the code that we rely on for bank transactions, package deliveries, medical results, satellite launches, airline flight paths, self-driving cars, mortgage payments, and nuclear power plants. This is durable code, and it’s going to stay that way.
I can tell you from first hand experience that, since long before LLMs were invented, critical software supporting these industries is held together with duct tape and baling wire (and Excel). "Durable" does not mean "good". In my experience production code is often ugly, poorly abstracted, full of special cases and hacks, but most importantly it works.
- I think the big assumption that this article hinges on is that AI will never create durable code. That cannot be true. Now that we’ve finally got a taste of the singularity, humans will never stop trying to bring about AGI, which by definition must write code that is exponentially more durable than anything a human could write. The question is ‘when’.
- I don't know why people are still eager to spike the ball so quickly after every wave of progress. Things are still moving SO FAST.
----------------
Ok it can complete a line of python, it will never write a full, correct function.
Ok it can write full, correct functions but it will never write full working programs.
Ok it can write full working (disposable) programs but it will never write real, mission-critical (durable) code.
- I would say durable systems are what run the world. The code making up the components are not necessarily durable! Just like living cells die off but maintain the consistent identity of the overall organism, software systems often deal with an endless churn of dependencies and components without changing their overall functionality.
If you have a system of a dozen components with well defined interfaces, you could imagine an LLM rewriting and redeploying each component one at a time, without changing the observable properties of the system.
- AI and vibe coding does not prevent the creation of good, robust and durable code. All it takes is for the coder to think carefully about the functions and not fall for the temptation to make the LLM add a bunch of fluff and features "just because they can".
- I write deplorable code , personally
Hacker News