Update 5/28/08: Over in the talk-polywell forum, I actually did a bunch of homework on what the US could do if we tapped into our unproven reserves (mostly off the coasts). At the end, I summarized the Dyson review: Here it is:
Current proven US oil reserves = 21 Bbbl
Mean US estimated undiscovered reserves (as of 1995, which are the most recent data I could find in a moderate Googling--go figure) = 44 Bbbl
Current US daily production = 5.4 Mbbl/day
Current world daily production = 83 Mbbl/day
So US production = 6.5% of world production.
So now we need to do a modest conjuring trick to turn undiscovered reserves into daily production. Since 21 Bbbl proven reserves results in 5.4 Mbbl/day, we can assume a reserve-to-daily-production conversion factor of .03%, which I will arbitrarily de-rate by half due to the fact that new wells are finicky and offshore wells are especially finicky--say .015%. Multiply that by the 44 Bbbl of undiscovered reserves and you get:
Possible additional US daily production = 5.6 Mbbl/day
Now, if you add that additional 5.6 Mbbl/day to the current daily US and world production, you get:
New US daily production = 11 Mbbl/day
New World daily production = 89 Mbbl/day
New US percentage of world production = 12.4%
Incremental world production = 5.6/83 = 6.7%
Which is nothing to sneeze at, especially given how steep the demand curve is for oil. Sliding the supply curve just a little bit to the right makes the price plummet. So, all-in-all, I agree with Simon's statement.
However, a couple of caveats.
First, let's all agree that all this nonsense about "energy independence" is just that--nonsense--as long as you're just talking about petroleum. Oil is a globally traded commodity, so even if all that US-produced oil stayed in the US, that production would affect the world price of oil, just as rest-of-world production would still affect the US price. In other words, oil is truly "fungible". If the US were able to supply its entire domestic demand (which is something like 20.5 Mbbl/day) and somebody were to nuke the Norhern Gulf oilfields, the US price of oil would still skyrocket along with that of the rest of the world. (The exception to that would be if we completely isolated US production and demand from the rest of the world, which is such a bad idea that it doesn't even warrant consideration.)
Second, none of these numbers include unconventional petroleum: shale oil extraction, coal gasification/liquefaction, etc. If you add that stuff in, we could be truly petroleum-independent for a very long time--maybe even long enough to get tokamaks to work! Bottom line: there are still plenty of dead, hundreds-of-millions-of-years-old critters to set fire to.
Which brings me around to one last interesting tidbit. Freeman Dyson has a review of a couple of recent books on the economics of global warming mitigation, here.
(http://www.nybooks.com/articles/21494 in case the link didn't take).
The review of the Nordhaus book is especially interesting for our little group here. Nordhaus did some net-present-value modelling on investment in global warming mitigation over the next 1 to 2 centuries. Nordhaus uses standard IPCC data (FWIW) to compute the cost of doing nothing (i.e. business as usual) at about $23T by 2100 in today's dollars. Relative to that, he computes the NPV of several investment scenarios and gets the following numbers:
Nordhaus's computed optimal program = $3T better than nothing
Kyoto with US participation = $1T better
Kyoto without US participation = $0
Business as usual = $0 (since we're normalizing to this)
The "Stern report" policy = $15T worse
The Gore policy = $21T worse
But here's the really interesting one: Nordhaus also computes the effect of a "low-cost backstop" technology that can either obviate the need for fossil fuels or absorb excess carbon from the atmosphere relatively cheaply. Dyson wasn't thinking about fusion as such a policy (and you should take a look at his proposal--it's interesting) but fusion would definitely count. Based on such a technology, Nordhaus's model yields (drumroll, please):
Low cost backstop = $17T better
Dyson's conclusion, which seems pretty reasonable to me, is that we ought to invest modestly in fossil fuel mitigation (aka carefully targeted carbon taxes) and invest heavily in possible backstop technologies.