zlacker

here is the thing, even those editors are relict of the pasts, the code is still in the center in these editors. thats something we need now, but not in the near (2 years, 5 years, 10 years?) future.

then the prompt is the coding, the reasoning is the execution, the code just an abstract layer that we do not care to much about i.e.: like assembly, machine instructions.

we know it exists, bit even here on hackernews i would guess only a small fraction know how it really works on a detailed level.

there will still be coding, instructions (prompt) -> execution (reasoning and AI code and code execution -> feedback (debugging to AI then and one point to the user)

bur actual looking at the code, well, thats only when this cycle annoyingly fails.

so current IDEs are still built from an code first mindset. this will not be the IDE of the future.

so basically OpenAI bought a Dinosaur