Also I never said that I "have no equal". I am saying that the death of my career has been predicted for no less than 10 years now and it still has not happened, and I see no signs of it happening; LLMs produce terrible code very often.
This gives me the right to be skeptical from where I am standing. And a bit snarky about it, too.
I asked for a measurable proof, not for your annoyed accusations that I am arrogant.
You are not making an argument. You are doing an ad hominem attack that weakens any other argument you may be making. Still, let's see some of them.
---
RE: choices, my choice has been made long time ago and it's this one: "I will become quite good so as to be mostly safe. If 'AI' displaces me then I'll be happy to work something else until my retirement". Nothing more, nothing less.
RE: "startup universe", that's a very skewed perspective. 99.99999% of all USA startups mean absolutely nothing in the grand scheme of things out there, they are but a tiny bubble in one country in a big planet. Trends change, sometimes drastically and quickly so. What a bunch of guys in comfy positions think about their bubble bears zero relevance to what's actually going on.
> I can't find it now, but there's at least one company that is doing enterprise-scale refactoring with LLM's, AST's, rules etc.
If you find it, let me know. That I would view as an interesting proof and a worthy discussion to have on it after.
"Your analogy with the automakers seems puzzlingly irrelevant"
"Your take is rather romantic."
That's pretty charged language focused on the person not the argument, so if you're surprised why I'm annoyed, start there.
Meta has one: https://arxiv.org/abs/2410.08806
Another, edited in above: https://www.linkedin.com/pulse/multi-repo-ai-assisted-refact...
Another: https://codescene.com/product/ai-coding
However, I still don't recognise the names. The one I saw had no pricing but had worked with some big names.
Edit: WAIT found the one I was thinking of: https://www.youtube.com/watch?v=Ve-akpov78Q - company is https://about.grit.io
In an enterprise setting, I'm the one hitting the brakes on using LLM's. There are huge risks to attaching them to e.g. customer-facing outputs. In a startup setting, full speed ahead. Match the opinion to the needs, keep two opposing thoughts in mind, etc.
Or build an AI agent to transform his speech into something more understandable.
https://docs.aws.amazon.com/amazonq/latest/qdeveloper-ug/tra...
To be honest, I have no idea how well it works, but you can’t get much bigger than AWS in this regard.