When I was at Amazon last year, the bragging (from the AI poo-bah in my section of Amazon, note) about AI included "look at the total line count of commits from the heaviest AI users!"
So if AI screws something up and re-writes it and then screws it up again, needing another re-write, that counted as more positive than if it was done correctly, and simply, the first time.
I don't know where you're working but LLM enhanced development has skyrocketed our rate of feature development. As an example, a project roadmapped to take 7 months was delivered in only 4.5 because of CC/Codex.
I'm confused how anyone could believe it isn't an enhancer, unless they have refused to use any of the technologies.
You're measuring success with time to delivery, that's a reasonable metric. Same with volume of features shipped. Also good. LoC or tokens burned... not so much.
I can confirm it's an enhancement in writing code specifically. We've been actively using CC in our company for more than 6 months already.
Notably, the product itself isn't really better for users. And almost everything else apart coding now takes the bigger percentage of time. So as devs we could either just fuck around and refactor endlessly, or chill out and "complete the sprint in 30% of time". It was known for a long time that churning out code is not the bottleneck.
Yeah I’ve experienced much the same as you. Like it’s overwhelmingly clear from everything it’s enabled for us that we’re going far, far faster than we ever have, and the guardrails we have in play have helped guard the architecture and make it even harder to commit a bad PR. Sometimes in reading these comments I’m left wondering what sorts of experiences people are having elsewhere that’s left them this soured on its usage in business.
Moving faster doesn’t necessarily mean delivering business value faster. You may be moving faster in the wrong direction.
More code doesn’t necessarily mean delivering more business value. You’re piling on debt and if that debt is growing faster than the value of the code, you’re actually losing.
And even if you somehow are delivering more, better code, faster, and without building technical debt: writing code is somewhere around 1-5% of the actual work and time that it takes to deliver a software product. At least in all the places I’ve ever worked. You are optimizing the wrong thing.
You might want to read my earlier comment in the overall thread re: delivering business value and moving the needle features while still building with solid architectural principles. As I mentioned above, the AI stuff - with the rails I’ve put in to guide it - has actually lessened tech debt introduced, not made it worse.
I’m well aware of what you’re saying, but we are definitively moving faster in the correct direction. If this hasn’t been the case for where you’ve worked, my sympathies!
I didn't send it very high up the chain (and was looking for a job at the time anyhow) but mostly got back snickers from peers and an "I know, but this is a directive from above"
I tried to make exactly this a year ago. Built on Cloudflare using all of their primitives: https://crawlspace.dev -- It didn't work too well (so don't bother trying it).
I have no plans in downloading Atlas either, but I think your browsing isn't used for training unless you opt in.
> By default, we don’t use the content you browse to train our models. If you choose to opt-in this content, you can enable “include web browsing” in your data controls settings. Note, even if you opt into training, webpages that opt out of GPTBot, will not be trained on.
