I am in general agreement that various changes are ongoing and consider it likely that Google will be or is already indefinite-beta-iteration on the factors he mentions. But I also believe (1) they haven't got an actual implimentation handle on several, (2) that his reasons for why Google is pursueing some is incorrect, and (3) there are serious usability problems left unmentioned.
Note: Being on dial-up I have not seen/heard the video. My information is derived wholely from the accompanying text.
...everybody who does a search will end up with different search results based on their prior search history. Personalization is going to be "web-wide"...
"We can't search for any term and look at rankings because I can get different results than you for exactly the same query
This started years ago with geo-targeting differences between cctld search domains, i.e. google.com or google.ca. Similarly with IP geo-targeting - being in Canada I have the google.ca choice of searching 'the web' or 'pages from Canada'. However, when I request 'the web' there are commonly differences from my same search on google.com. And when I use a proxy the results from both shift again reflecting the geo-location of the proxy server.
Now, as mentioned, individual personalisation is a growing SE interest. But there are (so far) insurmountable hurdles between all the real and persona conflicts. Multiple family members using one computer, work and home, research and personal, being in one country needing Local results from another, ad nauseum.
...a search for "java". Some are looking for programming, some are looking for the beverage, and some are looking for travel information. Theoretically, personal search will know that and provide results accordingly.
Note the word 'theoretically'.
In the real world it is never always one or the other. And the SEs are far far far from auto-magically divining which what when. Context remains elusive.
The biggest problem is that while SEOs have a fairly good idea where the SEs want to go they frequently fail to see the abyss between vision and implimentation. In many cases it still remains 'you can't get there from here'.
Bruce says this [universal search] will be the most significant factor of 2009. Google looks at soundtracks and convert them to text. They can jump into the middle of videos, scan sites to determine if you have video/images or not. There are a lot of variables in universal search, and some of these variables tell how engaging your site is.
He is much more generous than I to Google. I don't think they will give up many of those video spots to sites other than YouTube. WikiPedia aside, universal search is NOT about highlighting anything but other Google properties. And it was that (to Google) distressing dominence of Wikipedia that drove them to create Knol. Google really really wants those top five SERP (75% of clicks) as often as possible, really really wants to keep that traffic in-house.
Every change Google makes is to put more ads in front of each person. And personalisation is designed to better target ad delivery not better target results.
According to Bruce, this is all going to really help Google fight spammers. He says it will affect the way scrapers and spammers work, because if they are just scraping text, they wouldn't be able to rank.
It just means that as went text, then images, so video. Scrape and add AdSense and buy AdWords to drive quick traffic, same old now improved with someone else's vid. Google fights spammers because they pollute their algo, however, Google loves scrapers (but not publicly) because they are a major revenue stream.
I agree that SEOs need to adapt. They have always needed that ability. I totally disagree with that 'user benefit' kool-aid. Every single change is a Google revenue benefit. With a a user spin for painless consumption.
When looking at new SE (especially Google) offerings/changes/benefits look at what it does for the SE, then you will understand whether to, and how to, game the change.