Post

Any experts on #LLM who can shed some light on this? As we now can use ChatGPT and such to generate code for an entire application/system, why do we have to go through a high level language and not just generate machine code directly?

Erik Hellman

@hellman.io

The obvious exception is a web client, but for other targets this should be possible? Are we not doing this because the LLMs haven't been trained on application binaries?

2/12/24, 8:01 am0 reposts0 likes

No replies yet.