Scary article
NY Times
Nuff said
We had “coding without coders” in late 90s (maybe 2000s) with VB and Access databases. Some of my coworkers maintained such “software” previously written by not-a-dev.
And then there was “low code” fad about ten years ago? There was “coding” with diagrams and such, like Scratch but for serious people.
And what will regular developers do? Probably the same old shit, digging in decades-old, hastily-written, and now LLM-generated code, making it all work, and adding functionality. While “architects” and management will draw diagrams (with AI now!), and try to abstract everything into the cloud (and now into AI probably, somehow)
While I personally don’t like AI, I do think it is changing things. I don’t think it’s ever safe to run code without oversight from an actual programmer, but AI will likely affect the number of programmers being hired in a non negligible way.
If it’s run in a good sandbox, it’ll be safer than most of the code you run.
Then you add in controlled interfaces/gateways to give it “just enough” power to do something interesting… and you audit the hell out of those.
Risk is something that has to be managed, because it usually can’t be eliminated.
If you sandbox anything it’ll be safer than otherwise. Not really sure what you’re suggesting. I would still want the code reviewed regardless of the safety measures in place.
I wrote a program that basically auto organizes my files for me. Even if an AI was sandboxed and only had access to the relevant files and had no delete privileges, I would still want the code reviewed. Otherwise it could move a file into a nonsensical location and I would have to go through all possible folders to find it. Someone would have to make the interfaces/gateways and also review the code. There’s no way to know how it’s working, so there’s no way to know IF it’s working, until the code is reviewed. Regardless of how detailed you prompt, AI will generate something that possibly (currently very likely) needs to be adjusted. I’m not going to take an AIs raw output and run it assuming the AI did it properly, regardless of the safety measures.
Yeah, if you want it to write code that will act on important data or outside a sandbox, then a code review is still advised, even if only a sniff-test.
There was “coding” with diagrams and such, like Scratch but for serious people.
Yep. Genesys IRD is kinda good (well-tested), Genesys Composer I hate, and recent cloud Genesys workflows are not usable for anything but demos without pain.
