Hydrogen is a clean tailpipe emission in search of every relevant technology. Chu is right, and the National Academies of Science agrees with him (or vice-versa).

Science Musical, theme like Harper Valley PTA song. A scientist sings and points out other crazy ideas of the ridiculing scientist who also dance and sing.
Use MIT’s Technology Review web articles and blog entries for reference material.
??? “You say hydrogen fuel cells are impossible to fly? Well, I say a TIAX analysis of system cost presented at the DOE 2009 Merit Reviews earlier this year estimated a cost of $23/kWh. And you with your solar cells are way down that line my friend”??? la la la.
I would pay to see it.

simply by redefining 'hydrogen' storage as water in a tank.  The vehicle would then produce hydrogen as needed with a very small reserve tank for sudden increases in need such as climbing a hill pulling a trailer.

This is a well known process - water plus magnesium combines under heat and other conditions to form hydrogen plus magnesium oxide.  The magnesium oxide would them be recycled at a fueling station, likely run by the same people running gas stations today.

New pellets of magnesium would be piped into your tank when the spent oxide had been removed.

you'd top off your tank with water also.

The same reaction also occurs with aluminum.  Aluminum forms a protective skin preventing 'rust' or oxidation.  However in the presence of a small amount of gallium the protective skin won't form and aluminum will react almost explosively with water at room temperature, evolving hydrogen gas.

'gas' stations (now fueling stations) would truck the oxide back to a plant that would turn it back into metal, which requires alot of electricity, which can come from any source including renewable sources.  This is where the energey is 'put into' the fuel.  Alternatively the fueling stations could have a patch of solar panels, wind turbines or wire up to other electric source and turn the metallic oxides back into metal. 

These are well known processes but currently they are concentrated near where the ores are originally mined.  Aluminum is one of the most abundant materials in the earth's crust.  Likely this process would need to be spread out, closer to the energy use, with market patterns and power availability determining placement.

The last miracle, making fuel cells cheaper depends on 1) research finding the optimum material and configuration,  2) having a humongous chemical plant on the mississippi churn out the plastic membrane by the mile, lowering the price.  So the last miracle can happen too.

An additional miracle will happen when we switch over:  the efficiency of fuel->fuelcell->motion is twice that of ICE - internal combustion engines, so we will be saving energy.

This is such a one sided comparison. I mean, really, an FCX compared to a Cube? A fairer comparison would be an FCX and, oh say, the Tesla Roadster. But wait, the Tesla would be cheaper, it would be able to travel farther, it would be faster off the line. Oh, and it's available to the general public for about $875,000 less. Did I mention it costs less to fill up? Yes, Hydrogen is a little faster to refuel (by not much), but otherwise there's no difference. DId I mention it's available right now? Ok, I'm done.
Used travel trailers
in FL

After years working on E85 fuel systems, PEM and SOFC at GM and Visteon/Ford and now batteries - I can safely say it is time to mothball attempts to put fuel cells on vehicles.  The componentry, reliability and infrastructure gaps keep this as a science project.  We will never see hydrogen fuel cells in the next 25 years.  Focus should be on electrification of transportation, cleaning up the grid, and high speed rail.

If fuel cell cost drops to $78/kW or even $400/kW, (down from $4000/kW) then it makes sense to use them for water and building heat (from the 40-60% waste heat), and put the electricity into the grid. Stationary applications (also with no size or weight limits) are cost effective well before that of cars. Because fuel cells are suitable for micropower, they can be used for cogeneration wherever heat is required, so a relatively low efficiency in fuel kWh/fuel-BTU doesn't matter because the waste heat can be reused. Likewise, FCs would be convienient for small-scale biogas sources at a farm, sewage treatment plant, etc. At this price point, fuel cells for building heat become cheaper than gas turbines now used for peak power (also inefficient).

The DOE is deemphasizing fuel cells for autos, emphasizing fuel cells for stationary applications. It may be that stationary applications will yield the market and manufacturing scale needed as a first step to bring the price down another order of magnitude. This makes a lot more sense for the next 10-20 years-- we should be spending a lot more for stationary fuel cell research (now) than for auto fuel cells.

There is much debate these days comparing battery-electric vehicles vs fuel cell, but in most cases, the comparison is between current battery cost and projected future fuel cell cost. At current prices, batteries are cheaper than fuel cells, and we don't know what future battery/ultracapacitor costs might be. The EEStor hype seems no less credible than the $30/kW fuel cell hype.

If you start with renewable electricity batteries have much higher efficiency from electricity to motion but if you start with natural gas the efficiencies are not very different.  Hydrogen compression is about 85% eff and auto thermal reforming is about 85%. Compound these losses with the claimed 60% eff of honda's fcx and you get 43% eff.  If you take the same natural gas and generate electricity with a combined cycle gas turbine with an efficiency of 60%, run the electricity through a 93% eff grid and a 85% eff electric car the composite efficiency is 47%. This is the best case fossil fuel to electricity generation efficiency.  The grid average is not 60%. A 47% vs 43% efficiency difference is not that compelling. 
Natural gas distribution and storage exists.  I think it is worth considering natural gas stored on the car with an on board reformer and natural gas reformed at the gas station with hydrogen storage on the car.  The carbon foot print of transportation might shrink faster if dependant on more than one supply chain.   
Pursuing multiple battery chemistries makes since. I think transportation fuel cells are worth still considering under the scenarios where the well to wheels efficiency isn't terrible. There are many other factors to consider besides efficiency but the well to wheels efficiency argument is often used to make a case that fuel cells make no since whatsoever.
The point made above about fuel cells having many applications such as distributed power is also important to avoid branding of all types of fuel cells for all applications under a single term with negative connotations.