You might argue that since those LLMs build on all of humanity's knowledge, they should belong to everyone. But where do you draw the lines? Or make a practical case.
They kind of do belong to everyone. The technology and majority of training corpus is available to essentially everyone.
The training and inference infra are owned and operated, but I can't see an argument that any machine that processes public domain (or stolen, if you prefer) info should be available to everyone for free.
But those training and inference costs will go to zero. Think about your cell phone today versus $1m+ supercomputers in the 1980's.
We're living in a transitory blip where capitalists and gold rushers are getting rich arbitraging the cost of processing against the non-cost of corpus. We can argue about morals (it doesn't bother me much) but it is a narrow window and it will be remembered the way Compuserve is: a precursor to the actual revolution, worth a footnote.
Sure, that's what I mean by where to draw the line. An artist taking inspiration from others is not obliged to give up the work to the public.
Sidenote: It may be tricky/impossible in the future to uphold intellectual property laws. If anyone is able (for instance) to prompt-create all their software, a software patent is worthless.
I'm also baffled. AI that is substantially smarter than us is a very potential threat to us - and we won't even be able to comprehend what most of those threats may be.
It's head-in-sand denial because actually facing the threat (and our total inability to stop it, imo) is terrifying, and people don't like feeling that. So they're just angry and cynical instead. Plus it makes them feel smarter for whatever reason.
I am very product focused with my projects. I have Claude write 100% of the code but I challenge its suggestions and make damn sure there are ways for it to verify functionality and correctness. I don't read much of the generated code but am adamant that there are tests: unit, integration and (if possible) against other implementations. And I care a LOT about performance.
So far (last 6-7 months) I've built tons of stuff in my spare time: A pdf generator lib, a scheme implementation (R5RS and R7RS) with AOT compilation, a screenplay editor, a code editor, a ripgrep-like lib/engine/cli tool for fast search in a workspace, a markdown parser, a Fountain (screenplay) parser, a lib for dealing with updating apps, tooling for finding duplicate code and generating codemaps - and much more.
Cool. I'm also working on a Scheme compiler for embedding. Bytecode VM as well as AOT compilation to Zig. 100% written by Claude Opus under my supervision and guidance. I've given it an extensive set of tests and benchmarks (r5rs and r7rs) which helps A LOT. I currently use it embedded in a modal prose editor, mostly running integration tests for now.
reply