Man, tough choice, use an existing library or boil the planet to generate one on demand 🤔🤔🤔

a relative time formatting library that contains no code

The library is two text files (code) that are processed by an LLM (interpreter) to generate code of another type. This is not that new in terms of workflow.

I think what makes this the worst is the fact that the author admits that you can’t be sure the library will work until you generate the code and test it. Even then you cannot guarantee the security of the generated code and as you do not understand the code you also cannot give support or patch it.

When Performance Matters

If performance of a datetime processor is not relevant, what is? The author mentions they would like a browser implementation to be fast, documentable, fixable. However, operative systems, browsers, and other complex systems are made of little utilities like this that have very well documented functionalities and side effects.

But the above isn’t fully baked. Our models will get better, our agents more capable.

The whole assumption is that instead of creating a good stable base that anyone can use we should be just shtting out code until it works.

Eventually the hardware will be good enough to support a shitty bloated browser so we don’t need to optimize it.

Eventually people will harden their PC enough so we shouldn’t care about security.

This software library has no software, only specs!

So… Its a spec.

I wonder how much it really costs to have claude opus implement simething like this, the real cost, not the “get hooked on ai” prices we have now.

A library with no code, no support, no implementation, no guarantees, no bugs are “fixable” without unknown side effects, no fix is deterministic even for your own target language, …

A spec may be language agnostic, but the language model depends on trained on implementations. So, do you end up with standard library implementations being duplicated, just possibly outdated with open bugs and holes and gaps and old constructs? And quality and coverage of spec implementation will vary a lot depending on your target language? And if there’s not enough conforming training it may not even follow the spec correctly? And then you change the spec for one niche language?

If it’s a spect or LLM template, then that’s what it is. Don’t call it library. In the project readme don’t delay until the last third to actually say what it is or does.

I originally shared this after stumbling upon it in one of Martin Fowler’s posts.

The article reminds me of how my mother used to buy dress patterns, blueprints if you will, for making her own clothes. This no code library is much the same, because it offers blueprints if you wanted to build your own implementation.

So the thing that interests me is what has more value - the code or the specifications? You could argue in this age of AI assisted coding that code is cheap but business requirements still involve a lot of effort and research.

To give a non-coding example, I’ve been wanting to get some cupboards built, and every time I contact a carpenter about this, it’s quite expensive to get something bespoke made. However, if I could buy blueprints that I could tweak, then in theory, I could get a handyman to build it for a lower cost.

This is a very roundabout way of saying I do think there are some scenarios where the specifications would be more beneficial than the implementation.

I thought this would be about the Ai generated games, that has no code and runs like a video. Can we speak about software, if there is no code?