What happens when AI gets cheaper and more capable every quarter, but the bottleneck quietly shifts from model quality to memory and evaluation -- and who actually captures value when that happens?
This article comes at the perfect time, really hitting on the core issues. What if durable memory solutions become so seamless and cheap that the 'human bottleneck' for evaluation becomes the only significant barier we have left to overcome?
Hey, great question! I think human bottleneck won't be the only core issue. I think another problem that people do not yet pay attention to is the imporance of implicit knowledge. Underlying my post is that AIs are generally very good at tasks with abundant explicit knowledge with clear success criteria (= programming), but they are really bad at implicit tasks (= customer communication) if they are not given 'harnesses' to control their behavior.
While these harnesses are also rapidly developing, I think implicit knowledge is going to be so fragmented and hard to generalize that people who know how to collecting these implicit knowledge and making it explicit will be another bottleneck in the near future. What do you think?
This article comes at the perfect time, really hitting on the core issues. What if durable memory solutions become so seamless and cheap that the 'human bottleneck' for evaluation becomes the only significant barier we have left to overcome?
Hey, great question! I think human bottleneck won't be the only core issue. I think another problem that people do not yet pay attention to is the imporance of implicit knowledge. Underlying my post is that AIs are generally very good at tasks with abundant explicit knowledge with clear success criteria (= programming), but they are really bad at implicit tasks (= customer communication) if they are not given 'harnesses' to control their behavior.
While these harnesses are also rapidly developing, I think implicit knowledge is going to be so fragmented and hard to generalize that people who know how to collecting these implicit knowledge and making it explicit will be another bottleneck in the near future. What do you think?