[chartered_stack]

I think we're returning to CLIs mostly because typing remains one of the fastest ways we can communicate with our computers. The traditional limitation was that CLIs required users to know exactly what they wanted the computer to do. This meant learning all commands, flags etc.

GUIs emerged to make things easier for users to tell their computers what to do. You could just look at the screen and know that File > Save would save the file instead of remembering :w or :wq. They minimized friction and were polished to no end by companies like MSFT and AAPL.

Now that technology has got to a point where our computers now can bridge the gap between what we said and what we meant reasonably well, we can go back to CLIs. We keep the speed and expressiveness of typing but without the old rigidity. I honestly can't wait for the future where we evolve interfaces to things we previously only dreamt of before.

[ewoodrich]

It’s less rigid than a command line but much less predictable than either a CLI or a GUI, with the slightest variation in phrasing sometimes producing very different results even on the same model.

Particularly when you throw in agentic capabilities where it can feel like a roll of the dice if the LLM decides to use a special purpose tool or just wings it and spits out its probabilistic best guess.

[chartered_stack]

True the unpredictability sucks right now. We're in a transition stage where the models can understand intent but cannot constrain the output within some executable space reliably.

The bridge would come from layering natural languages interfaces on top of deterministic backends that actually do the tool calling. We already have models fine-tuned to generate JSON schemas. MCP is a good example of this kind of stuff. It discovers tools and how to use them.

Of course, the real bottle neck would be running a model capable of this locally. I can't run any of models actually capable of this on a typical machine. Till then, we're effectively digital serfs.