Nice to see your critical take on the whole AI taking over the world thing (looking forward to audio on this). I believe that AI's and super AI's could emerge, and it's worth considering seriously. I just get a little put-off with the sense of inevitability about it (either we're doomed, or it'll be great). The story often feels to pat and neat in the Singularity crowd.

One big problem with the picture is the whole idea that more intelligence = more capable / effective. Perhaps. But perhaps most possible super intelligences are also super insane and not very good at impacting the world (or people). In other words, is there a diminishing returns scenario for investing in "intelligence", where more and more intelligence gets you less and less marginal improvement in capability?

The talk looks really interesting. Any luck with the video yet?