While Im still personally skeptical on the ability of these tools to produce a GOOD software engineer, it’s something I should probably consider testing in a limited capacity.
I’ve noticed Deepseek has a few integrations, both official and hobbyist, with text editors like Claude Code. Plus, I’d rather not pay £20/mo for any of this stuff, let alone to any AI company NOT linked to the CPC.
I might consider a locally hosted model but the upfront cost for anything that can run it decently fast/high params are quite prohibitive. My home server isn’t really set up for good cooling!
LLMs are terrible for coding. Keep a personal journal instead. Note down every problem and obstacle you’ve overcome and index them.
I find it works really well for text transformation (ie. arbitrary text into JSON, Yaml, or language-specific class schema) which is something that I’d be down for having in my IDE where I wouldn’t have to build out the context manually every time I made a request.
However, I agree that it’s “terrible for coding”, in that I have tried and failed many times at prodding it to do a complex task that includes some form of problem solving or creative thinking (ie. laying out widgets but not giving any guidance, generating tests based on a natural-language specification) and those cases are where I’m already wary of using it at all.
Doesnt stop my coworkers, though. One of them has gotten claude code access at work and now spits out this shitass code with no context to the wider project - importing classes that already exist via arguments or class properties, weird state management that breaks convention, inconsistent naming and poor documentation… It really reproduces the same kind of quality that a dev would stoop to if they arent being held to account.