CO2 emissions is a problem now.   Because the feedback effects could kill us.  By the time we take action, the biosphere could already have lost its capacity to restore the damage.

And because it will take so long to deal with the CO2 emissions of very long lived equipment: power plants last 50 years, cars last 15 years, planes last 30+ years, etc.  China has something like 100 coal plants under construction, and once built, carbon sequestration is that much harder to implement (to the point where it will probably never be implemented on those plants).

Peak oil?  It's a big unknown.  My own view is that we will hit PO (by the definition of an exhaustible resource) but more likely in 2020-30 than on Thanksgiving Day 2005 (Deffyes' date, I believe).

There are a *lot of alternatives to conventionally produced oil out there: tar sands in Canada, heavy oil in Venezuala, the Fischer-Tropf coal-to-oil process, NGLs.  All of these can, and will be scaled heavily (Canada alone will be 5 m b/d).  And there is still Arctic oil, deep water oil, plus further exploration and production in Africa.

But note the CO2 consequences of widespread coal-to-oil projects.

The state-owned oil companies that control most of the world's oil resources are underinvested and undermanaged.  Saudi Aramco is perhaps the best, but in Mexico, Russia, Iran, Venezuala etc, there are serious issues.  So there is likely to be more oil squeezed out there.

* peak gas, or at least accessible peak gas, worries me more, oddly. We can substitute gas for oil in a lot of applications, but we don't have ready substitutes for gas (except more coal!).  And from Simmons gas graphs at least, 'old gas fields don't fade away, they just die'.  We don't readily have the transport infrastructure to get the gas from where it is, to where it is needed, even if we do have that gas.