Alright, slight exaggeration on my part. A few
hundred thousand years is closer to the mark. Changes in the local environment (eg: temperature, pressure, etc) of the diamond will affect this.
It's a change that doesn't take place at higher pressures, and will not take place within the rock matrix unless there is a substantial heat gradient. But at atmospheric pressure, free of the rock matrix, diamonds will slowly begin to go back to graphite.
Diamonds that we find within the Earth's crust were formed in the mantle, a layer of liquidised rock at an extremely high pressure and temperature. They grew there just as any crsytal will grow in any liquid, when the liquified rock became saturated with carbon (for whatever reason. It's still not entirely clear how these liquid rocks become saturated with carbon to the extent that it starts to crystallise under high pressure and temperature). For it to be found in the crust, an eruption of high-pressure magma must force it up through the mantle, into the crust.
Having formed under immense pressure and heat conditions, this crystal cools rapidly, and loses too much energy to decompose to graphite as it would prefer to do. It forms a lump, roughly shaped like two three-sided pyramids stuck together.
It's trapped within the rock matrix as
that cools, and there is no effective difference between the pressure at which it cooled, and the ambient pressure to which it is subjected. There's no pressure gradient, and no thermodynamic impetus for the diamond to change form.
This is why you find billion-year-old diamonds embedded in Kimberlite.
Once the diamond is removed from the rock matrix, it is at a different pressure to that at which it cooled, and there's an associated entropy value with that. Crystal defects will now form over time within the diamond structure, dependant upon ambient temperature.
It does take a long time for this to happen. The diamond lattice is very strong, and thanks to the relative speed of the cooling process that it undergoes, it has a very low entropy value. But that value
is there. Whilst the change from diamond to graphite takes a long time, it's still something that we can observe (sort of).
At atmospheric pressure, there's nothing maintaining the force which originally made the diamond lattice the more thermodynamically favourable structure. As the diamond absorbs energy from its surroundings, it gradually gains the energy needed for the molecules to assume the most thermodynamically/entropically favourable arrangement. To observe this process "naturally", we'd need to use a high temperature, low pressure environment (or live for many thousands of years).
You can speed this process up by exposing your diamond to a high-output laser. The energy of the photons is just enough to break bonds within the lattice (at sites where there is a flaw in the crystal or an inclusion of some other atom), and this process will cause a darkening of the diamond at that point. That dark spot is graphite. The photons emitted by the laser will be absorbed by the graphite, raising the energy of the structure as a whole. The diamond darkens, turns black, and crumbles into powder rather quickly once one of these inclusions is formed. If you don't have an enormous laser at home, then you probably won't be doing this to your diamonds (I'm assuming you have a sack of those).
This is why diamond lasers burn out really quickly. The gemstone is turned to graphite by the light energy that it is focusing. Eventually, it is completely "consumed" (and your laser no longer works). But, good news! Diamond lasers made with synthetically manufactured gems are a lot better at handling this phenomenon. They contain fewer impurities, and a lot more energy is required to break any of the bonds within the lattice.
Diamonds made synthetically will overall not last any longer than natural diamonds in other applications though, and other gems are a lot better for producing coherent light emissions. Synthetic forms of the various
aluminium oxide gemstones are popular in this regard.
So, to summarise, if those diamonds were left in their rock matrix, they'd maybe last
another billion or so years. But once they're removed, they will begin to decompose (although at a rate that's so slow we can effectively ignore it). We can speed up this process by increasing the amount of energy in the lattice structure (heating, firing lasers at it, etc) or we can slow it down by increasing the pressure on the lattice, by cooling it down to remove latent energy (at absolute zero, a diamond would be perfectly stable, perfectly flawless, and last forever), or we can simply grind it to sparkly dust and use it to make drill bits, saw blades, and abrasive discs.
Larger diamonds are also found mainly as
lumps rather than tetrahedral crystals. They look somewhat amorphous and often contain dark inclusions where the centre has begun to revert back to graphite before it was cooled.
The "little double tetrahedrons" are smaller crystals and have overall a lower entropy value, so are closer to perfect crystal structures.
Source: My Inorganic Chem notes from last year, plus Google. Having finished this, I skimmed the
Wikipedia article on diamonds to see if there was a convinient summary of any of this I could link you to. Wikipedia seems more interested in talking about other properties of diamonds. It does note that diamonds will decompose into graphite over time, but says nothing about it, other than describing the rate of conversion as "negligable".
This is a condensed version of several dozen pages of text overall, and I might have omitted something or even said something that's complete bollocks by accident. I make no apologies for this. It's not meant to be something you can base a thesis on, just a note of potential interest originally based on a comment that appeared on a messageboard on the internet. I have not researched this in great depth myself, and if you're really interested, I suggest turning to google yourself for further information.Anyhow, I decided to show the rest of the square avatars I sketched out. I scanned the other sheet, and coloured a few in... then got tired of it.
Randi, there
is a borg cube among them. Notice how different it is from the Minecraft diamond block.

I coloured these three before I got bored:
Square:

Borg:

Cobwebs and question mark:
