http://www.youtube.com/watch?v=7ImvlS8PLIo

This is a great talk, with some very complex ideas clearly explained.  Particularly good is the discussion of why the universe looks flat.

There are a few of points that Krauss discusses that I want to add to.

Firstly, Krauss says that the flatness of the universe means that it could have come from nothing.  That is a bit of a problem, as what he actually means is that it could have come from a quantum fluctuation.  But that fluctuation would have required some framework of physics in which quantum mechanics operates.  The term "fluctuation" is also a problem, as what is it that is fluctuating when there is no space and time.  I don't disagree with his conclusions, but his language does not really deal with the question of what the universe could have come from.  However, he makes a good point that the true language of physics is mathematics, so trying to put things into words is a real problem. 

Secondly, Krauss makes an interesting point about the problems of infinity and eternity.  In an infinite and eternal system (which can involve more than just our universe), quantum mechanical fluctuations have time to make anything at all.  This has led to some philosophical problems.  Science assumes that we are typical observers of reality.  But in an eternal universe which reaches heat death, there is endless time for quantum fluctuations to create - out of nothing - intelligent beings.  Those should outnumber us infinitely.  (These beings are called 'Bolzmann Brains', after the great mathematic physicist Ludwig Bolzmann, who came up with the idea).    There are various possible explanations as to why this could be wrong.  Perhaps universes can't be eternal (they would have to last a very, very long time indeed for a Bolzmann Brain to appear).  But also, it is most probably very, very much more likely that a random fluctuation will create a new universe than a brain.  So, young universes with evolved intelligences will vastly outnumber those with spontaneously randomly arisen intelligences.   It is a weird problem, but this is the kind of thing one ends up with when dealing with infinity.

Finally, it is stretching things a bit to make any kind of prediction about what the universe will be like in 100 trillion years (I think that is what he said), especially in a talk that describes how our understanding of cosmology changed so dramatically just over a decade ago with the discovery that the expansion of the universe is accelerating.  Who knows what we will find in another decade or two?  Predictions aren't wise when the system you are using to predict is changing so much.  There is even a possibility that the acceleration of the universe is itself going to speed up, leading to a rather dramatic destruction on all matter, perhaps on a timescale of only tens of billions of years.   Unlikely as this probably is, it is still not excluded based on observations.  As we have no idea what dark energy is, there is the possibility that the universe could enter a phase of deceleration at some time.

Regarding the environment of future astronomers in his scenario, I also feel Krauss is possibly mistaken to suggest that they could find no evidence of the Big Bang.  This may just be possible, as event horizons give off radiation, and due to expansion we would effectively be in a galaxy surrounded by an event horizon.  It is just barely possible that an extragalactic probe could detect this, and realise that space is expanding.  OK, this is very unlikely, but if we are talking of technologies trillions of years hence.

But anyway, a really fantastic talk; taught me a thing or two.