So they’re mostly overblown report slop with no real-world value until an actually skilled and experienced human sifts through it to find the gems and properly polish them? Much like pretty much anything else AI produces? Yeah I thought so.

As a senior software developer, the reports of my job’s demise remain greatly exaggerated. I do worry about the lack of junior devs being afforded the opportunity to be gainfully employed while they earn the same level of experience I have, though. Considering they’re called “learning models” I’m not convinced the current crop of AIs are actually capable of the quality and speed of learning that an actual human is, deeply flawed creatures though we are. AI cannot achieve the same without access to enormous quantities of training data which, while it might be available in sufficient quantity for the right price and energy cost, won’t be of sufficient quality to actually learn valuable insights from, especially so the further we move into replacing all carefully curated human knowledge with an AI slopfest, reducing the signal-to-noise ration and actually get further from solving the “discovery problem” that has plagued us since the dawn of the information age.

Junior devs, don’t give up. We haven’t given up on you. (only the corporate AI bubble leaders have, and the bubble will eventually deflate even if it doesn’t catastrophically pop, which it probably will)