Interesting discussion, just a few thoughts to add (not too detailed, but perhaps some time in the future more)
My personal opinion is that too much is being made of the lactate, though I recognize that this is the "measurable" molecule that is easiest for coaches to use to pinpoint the physiological stress of the exercise. I don't however believe it to be as crucial as seems to be suggested here, because lactate is really just a symptom of what is happening to the various process that produce and remove it. What it ultimately reveals, I would argue, is the degree of sympathetic response to exercise, and that's where I think there is likely a difference.
Just on lactate, remember that lactate PRODUCTION is the result of the flux through the glycolytic pathway. Lactate REMOVAL and OXIDATION are a function of lactate transporters (called MCTs in the muscle) and the enzyme LDH. And OVERALL lactate LEVELS are a function of the balance between production and removal/oxidation.
In terms of production, the capacity of an enzyme called pyruvate dehydrogenase to catalyse the conversion of pyruvate into Acetyl CoA in the TCA cycle is one of the key limiting steps in metabolism. Therefore, when the rate of pyruvate production is high (which is primarily determined by the ADP/ATP ratio, the NAD/NADH ratio and the calcium levels), the rate of production of pyruvate exceeds what can be taken up and converted into Acetyl CoA.
So when you exercise hard, and recruit more muscle, and use more ATP, everything is geared towards driving MORE ATP production and this is achieved by increasing the rate of glycolysis. End result? More pyruvate is formed. However, there's now a "bottle-neck", since the pyruvate cannot all be converted to Acetyl CoA in the absence of PDH at high enough levels. So another enzyme, Lactate dehydrogenase takes over and forms lactate.
To cut the biochemistry lesson short, what this means is that if the rate of the pyruvate production is excessively high, lactate forms simply because there's nowhere for the pyruvate to go. The key difference then, would be in the level of the PDH enzyme, and this is an enzyme that exists in the mitochondria.
As you know, the mitochondria are responsible for the chemical reactions that oxidize fat, as well as the TCA cycle and oxidative phosphorylation, and one of the key adaptations to endurance training is the production of more mitochondria. This enhances the ability of the body to produce ATP via oxygen-dependent means, and I believe this is the crucial reason why the lactate may be lower.
However, and this is very important, there is another explanation for why lactate levels may be lower, that is that lactate is used as a source of energy. For more on this, read Brooks' theories on the lactate shuttle. Now, there's been some discussion in this thread about "buffering" - that is a distinct metabolic process from lactate oxidation. I know what Renato and JT are talking about, but buffering is different from oxidation of lactate. Buffering does not produce energy per se, it simply buffers. In order to produce energy, lactate has to re-enter the muscle cell (if it is in the circulation) via MCTs, and then be converted back to pyruvate and then oxidized. Brooks' theory is that this happens in some muscle tissues (ST fibers) to reduce overall reliance on glycogen, and this is what JT mentions in his post.
To return to the question, it would be possible that this "new breed" of runner (a term I disagree with, but anyway, more on that below) simply produces less lactate. Studies are needed to look at this, for the simple reason that what you measure in the blood is the result of a balance between production and re-uptake. Therefore, those lower levels could just as easily be explained by a greater capacity to use lactate as they can the possible reduced production.
I suspect both to be true. There is evidence that good black runners have lower lactate levels than good white runners (this is for good athletes, not elite), and also evidence that they have higher levels of MCTs, the transporter that helps oxidize lactate. Do they produce less than average? Yes, of course, because they have more mitochondria and thus can take the pyruvate into the TCA cycle and use it more effectively.
Before one can convincingly argue that the lactate levels are lower because of "buffering" or oxidation, you'd have to measure production and removal (radioactive tracer studies, never been done on this level athlete), so it's a bit of guess to say this is what happens.
So you have a combination of less production and more removal. Also remember that the fat oxidation is probably exceptionally high. We measured this in elite Ironman triathletes, a sport where glycogen depletion is almost impossible to avoid. And the winners of international Ironmans have an exceptionally high capacity to use fat. The same will be true of Kenyan and other elite distance runners.
The approximate energy cost of a marathon is around 2500 KCal. If 100% of the energy for this comes from carbohydrates, that's 630 g of carbs. That's close to limiting. HOwever, even 20% of the energy from fat means 500g of carbs and 60 grams of fat, and so clearly the picture changes. 40% of the energy from fat (probably an overestimate given the intensity of a 2:04 marathon runner) means 380 g carbs and 112 g of Fat.
And note that this doesn't factor in the re-oxidation of lactate, which as I've mentioned, probably makes a significant contribution.
Anyway, is this enough to explain a jump, and a so-called "new breed of runner"? Not in my opinion, I think we're seeing marginal improvements - the last 6 world records in the marathon average 28 seconds of improvement. That is a 0.4% improvement. Even if the WR was to drop to 2:02:30 within the next year, it's still only a 1% improvement - that kind of tiny gain is not measurable through science, so physiologically, biochemically, even mechanically, I'd argue against anyone who says we are seeing "CHANGE".
What we are seeing is PROGRESSION, and it's probably driven in part by tiny metabolic differences as a result of training differences, the increase in high intensity training is likely increasing lactate removal and oxidation capacity, buffering capacity (again, two different things), and possibly, fat oxidation. There are some interesting studies in cycling that show that high intensity interval training has a profound effect on fat oxidation rates. So we're edging faster.
Still think it will be a while to a sub-2:03 though.