I'm not understanding something - are you are saying these investments will result in more complex products from the same input of crude, or more output (i.e. are they increasing the efficiency of refining)?  Also, how is diesel so much more complex that heating oil (I run heating oil in my tractor!)?

From my admittedly ignorant perspective, I find it hard to swallow that refineries were not already pretty good at making diesel and gasoline.  What kinds of throughput improvements are possible for these outputs?  And if the investments are to make other more profitable but lower volume distillates, then does it really matter?

Twilight:

Diesel and heating oil are basically the same until the very end of the refining proces where they are treated differently. Fuel oil is a far lower quality product and is the only significant output that is worth less than crude (see Bangchak slide #12).

A complex refinery does produce a lot more gasoline and diesel from the same barrel of oil. Slide 25 of the Bangchak presentation shows that the refinery currently produces 37% fuel oil, 37% diesel and 15% gasoline. The addition of a $250 -350 million dollar hydrocracker and other equipment changes this significantly. Afterwards it would produce 9% fuel oil, 52% diesel and 25% gasoline from roughly the same crude input.

Refineries have always been good at making gasoline and diesel, they just made less of it from the same input

According to the American Petroleum Association, the typical US refiner produced the following products from a typical barrel of oil.

Distillate: 9.7 gallons
Kerosene-type jet fuel: 4.3
Fuel oil: 2.0
Liquefied gases: 1.9
Still gas: 1.9
Coke: 1.9
Asphalt: 1.4
Petrochemical feedstock: 1.1
Lubricants: 0.5
Kerosene: 0.2
Other: 0.4

Sorry,

"...the refinery currently produces 37% fuel oil, 37% diesel and 15% gasoline." should read 31% fuel oil, not 37%.

OK, so then presumably the investment can result in producing more of the outputs people want from the same crude input, so it's probably an attractive one.  BUT, that's different than investing in refineries or modifications that result in more usage of crude.  So it may be that your point actually reinforces what Stuart is saying, in that the investment that is occurring in refineries is not directed towards projects that would increase the use of crude, but toward maximizing the profitability of the existing crude inputs.  And this would make sense if one did not expect larger quantities of crude to be available.
Two points:
1 It is true that no US refineries have been built in 25 years, but this is very misleading - US refinery output, measured in barrels input, has increased around 30% in the period, albeit almost nil since 2000.
2 Refinery profit is dependent on cost of crude less price of product. If crude is rising faster than product prices, as has been true for most of the period from 1998, refinery profits are declining, reducing interest in new investments. (European supplies of product post K/R has naturally cut product prices and refinery profits.)  The only company significantly expanding capacity in the US is Valero, which buys up old, antiquated refineries, adds substantial new equipment designed to handle cheap sour/heavy crudes, and ends up with higher output. So, most integrated oils are not interested in expanding because their profits are anemic, but Valero is happily exploiting its niche. Eventually, if refinery capacity is a true bottleneck, crude stocks will continue to rise (as they have been for the last year) while product prices rise, increasing refinery profits and renewing interest in expansion. OTOH, OPEC may defend $55 NYMEX, restricting crude to match refinery capacity.