I feel like I still have yet to see any decent answers to this question; Are professional SDE paying for Claude on their own dime, and then logging into their personal account and somehow integrating Claude Code (or other LLMs) into their work repos that way?
The startup I work for has chosen their flavor of AI subscription and its frankly not developer focused. Instead they chose Google because of the productivity tools in the Google App suite.
I want to try Claude Code but the reality is that I don't want to be the martyr that tells my team lead that I want to see if AI can do (parts of) my job for me. It feels pretty sketchy or maybe even completely wrong to use something like this on a company repo I don't own without permissions, so I haven't done it. I suppose I will just have to continue to wonder if the agentic coding services are smoke and mirrors because I somehow don't know anyone who has used them extensively either and I have no clue when I will be able to use one with the strings attached of it being on a free-tier...
Yes I pay for the most expensive Claude sub with my own money and use it at work.
I also have to use it via a proxy server I set up to get around the corporate firewall which explicitly blocks it. The company like the results but wouldn't like how I get them..
More corporate ridiculousness
Our company set up some kind of Wise debit card thing where we each get our own number, and they told us "try out any AI tool you want."
So I subscribe to a new one every month to try out while still shoveling like 150$/mo at Claude cause it's consistently been the best and the one I use the most. Cursor as well has been good for their completion model which surpasses anything else I've tried for inline/multiline/jump completions.
But I've also tried supermaven, codeium/windsurf, copilot, zed. I guess from the company's perspective, a couple hundred bucks a month is well worth the time of keeping us all up to date with ai tooling.