Jeff Hawkins gets shit on a lot in ML because his theories haven’t produced models with results as good as the mainstream approach, but I’m glad they keep working at it and keep coming up with interesting ideas.
I just ran across a paper [1] from Numenta on Time Series Anomaly Detection using HTM last night which provide a benchmark[2] with some existing approachs. (But it seems to me there is no NN based approach in them. )
Finally: "Those who have done really great things generally report, privately, that it is better than wine, the opposite sex, and song put together."
This is strongly contradicted by Mihaly Csikszentmihalyi's evidence. Nearly everyone he talked to (many Nobel Laureates among them) said their greatest achievement was raising a family or something in personal life.
Strictly speaking, these two claims don't contradict each other. Talking to a book writer/researcher in an interview doesn't count as talking privately.
But what Google does is as far as any sane donator/sponsor would go. We can't really expect them to keep sponsoring more doughnut shops after they find out their money is not being put into good use.
Alternative explanation: Star scientists are just highly productive scientists. After they die, their not-as-productive collaborators can't keep up with the amount of publication as high as before. Naturally collaborators publication percentage falls while Non-collaborators' rise.
So isn't it perfectly normal that it also takes several years for the total value of IPOs and acquisitions, i.e. money that flows out of startups, per year to catch up with the rising trend of total startup investment, i.e. money that flows into startups, per year?