By this point I assume we'll have moved on from ChatGPT to more agentic systems that basically everyone agrees at least in principle are useful even if they might disagree with how those systems are made or their resource use or their perceived (or real) impact on art etc. bsky.app/profile/jdp....

While these agent systems generate mountains of grounded long text and are therefore quite good at reasoning and programming tasks, they still have limited access to the outside world since robots are expensive and people are uncomfortable just letting them walk around doing stuff.

🚨 “everyone agrees” 🚨 I think the last four years have taught us that SF tends to be way too optimistic about this part. I’m skeptical about the implicit underlying hypothesis that as these systems become more useful, more people will accept them. I think agentic systems will freak ppl out.