> What is alarming, however, is the degree of autonomy and depth of exploration the AI displayed.
Claude Code made a change on March 26th to skip asking for most permissions. See this quote "Claude Code users approve 93% of permission prompts. We built classifiers to automate these decisions":
It's helpful to rewrite software every 2 years as your team turns over. The code you understand best is code you wrote yourself, and no one likes maintaining other people's legacy code.
This is the risk of being a consumer in the AI world - companies are running extremely lean on real humans and are deferring support to AI chatbots with no real reasoning abilities...
Also an issue with scale - for example, Google having similar issues of not handling small, isolated cases.
As someone involved in hiring, there are a few factors going on:
1. There's no available headcount. No one is leaving since there are no opportunities elsewhere. Especially those on visas.
2. When we do have headcount, we get 1000 resumes within an hour or two. We also have a stack of referrals to get through first. A lot of these resumes are experienced folks with backgrounds in FAANG.
3. Supply remains high (100k CS undergrads every year including global supply) whereas innovation is low.