| Website and occasional blog | https://www.logophile.org |
| Infrequent macro photos | https://www.flickr.com/photos/tikitu/ |
| Software | https://github.com/tikitu |
| Ex-Twitter | https://twitter.com/tTikitu |
| Website and occasional blog | https://www.logophile.org |
| Infrequent macro photos | https://www.flickr.com/photos/tikitu/ |
| Software | https://github.com/tikitu |
| Ex-Twitter | https://twitter.com/tTikitu |
RE: https://neuromatch.social/@jonny/116350870063793276
This stuff is so fascinating as a glimpse of what happens when you remove human oversight completely and just let an agent-loop completely run wild. Also fascinating because doing this significantly better feels totally attainable while still getting a lot of benefit from the tools: this mess is a CHOICE.
RE: https://wandering.shop/@offby1/116341713293151884
the linked piece is excellent: tl;dr, better argued than I can: the risk of agents in "intellectual" fields is giving you freedom to choose which of the steps in work you want to skip doing yourself. Knowing what you can personally skip (given the option) is what you earn from a career of learning and re-learning.
My greatest professional accomplishment of the year: I got my exec & manager teammates saying "point positive," a term from whitewater rafting and kayaking.
Meaning: when facing hazards, point people toward where to go/what to do, rather than drawing attention to everything to avoid.