The launch of Google Instant has caused some to predict the end of SEO. This seems hugely overblown to me. Google Instant won't kill SEO. Not even close. It may shift some nuances in SEO, but by and large the art of driving organic traffic from search engines will remain unchanged.
What the introduction of Google Instant does do is change the way searchers interact with their searches. People will probably look at more search results during a search session. The streaming results will create a feedback loop that makes people iterate quickly on their search queries. Searchers will also be guided, inspired and influenced by suggested queries.
But what does not change is that people come to Google to search for specific information, and leave Google when they've found what they were looking for. This is of course Google's bread and butter, and business model. 90% of searchers leave Google by clicking on an organic link, and the remaining 10% click on a sponsored link. This is a highly efficient (and profitable) practice. If Google does a good job, searchers end up at the best possible page to answer their question or solve their problem. It is unlikely that significantly fewer or more people will succeed or fail with their search tasks using Google Instant than using classic Google search.
Since the number of visitors to Google and the number of indexed pages on the web remain essentially unchanged post-Google Instant launch, the average number of visitors to individual web pages through organic search will remain unchanged too. This is a classic example of a zero sum game. Some sites may loose some traffic, some may win some, but on average sites will get the same level of organic search traffic.
If a site ranked well in the classic Google world, it will rank well in the Instant Google world and its inbound organic traffic should remain largely unchanged.
I do however predict some relatively minor shifts (which will benefit some sites at the expense of others):
* The "Fat Tail" will gain some traffic at the expense of both the long tail and the head. The logic goes as follows: When people type a generic search query (for example "San Francisco") they will see suggestions to refine their query (for example "San Francisco hotels"). If Google's suggestions are any good, the suggestions will often prompt people to opt for the more specific query. Likewise when people set out to do a very specific query (for example: "San Francisco hotels that are family friendly") they may be tempted to pause at a suggested query (for example: "San Francisco hotels").
* The difference in clicks between top ranked results (positions 1 - 3) and lower ranked results (4 and lower) will widen. This prediction is based on my own (relatively new) experience trying out various queries today. The interactivity of the page make me less likely to scroll down since there is already so much going near the top of the page.
* Highly ranked generic pages could slightly increase their traffic gap with lower ranked, more specialized pages. This prediction is due to the fact that highly ranked sites with pages relevant to a search intent will start showing up while the search intent is not yet fully clear to Google. For example, when a user intents to search for "Yosemite hiking trails" a highly ranked site for this topic may show up when Google figures out that "Yosemite" is being typed (along with other relevant pages for Yosemite). Once the user sees a page about trails, she may click on it, never seeing the more specialized site that would come up when the query is more complete.
Obviously the above predictions are based on initial thoughts and observations. The real effects of Google Instant will slowly emerge over time.