The Fail-Safe Illusion: An Ant in the House [pdf]

(zenodo.org)

5 points | by davrosent 10 hours ago ago

3 comments

  • foopod 8 hours ago ago

    I’d frame this as less of an AI problem and more a scaling one. Cheaper development effort brings the same issues we’ve always seen when teams grow too fast: QC and verification gaps, fragmented codebases and culture. Throwing tokens at a problem is a lot like spinning up an offshore team.

    The difficult part imo is the junior pipeline. I don't think we will be able to understand the impact of not investing in juniors for at least a few years, it will be interesting to see where things land.

    • solyin 7 hours ago ago

      Right. Me retiring is great for me in 5/10/20 years, but I'm not getting paid money to fix problems that businesses have because businesses like giving me money because they have problems finding anyone to give it to. If AI keeps growing and business owners can keep giving all their problems to AI, great, but in a reflection of who actually runs society aka anyone who spends money on things that we choose to possibly identify as a collectively demographic eg all bowling ball users, volinists, women, men, car owners, etc; then the price is gonna skyrocket.

      What do you absolutely need to have happen, or that someone else does for free because they can't (aka slavery)?

  • davrosent 10 hours ago ago

    [flagged]