> Apart from some limited, low-resolution medical trials for people who've lost their eyesight already, this is firmly in the arena of science fiction and not going to happen any time soon.

Define any time soon.

The year I was born the ZX80 was released, it cost about two weeks wages (revolutionary at the time for it's price point) and had a 3.25Mhz processor with 1KB of memory.

Against that backdrop imagine walking up to someone on the high street and handing them a Nexus 5 and pointing out that it's floating point performance was about 50% faster than the Cray-1 which was then the state of the art (at about $8 million then dollars) oh and it has 256 times the RAM and a two orders of magnitude more storage and costs less than a weeks wages and run for a full day off a single charge consuming less power than the Cray 1's circulation pump.

It's a hard call to say "not going to happen any time soon" when we have proven the basic principle.

> oh and it has 256 times the RAM and a two orders of magnitude more storage and costs less than a weeks wages and run for a full day off a single charge consuming less power than the Cray 1's circulation pump.

Yeah but mobile processing power is not increasing at the same pace now as it was 2 years ago. The pace is clearly slowing down, just like computational power on PCs too. You don't get 2x faster processors every year anymore. They are adding more cores, but that's not nearly the same thing as single thread performance.

The basic principle proven so far is that, with an invasive surgical procedure, retinal function can be permanently replaced with technology that electrically stimulates the retinal ganglion cells (i.e. the side of the optic nerve that terminates in the eye) - and that it works in some blind people.

This predicts nothing about augmenting the optical input in people with healthy eyesight. I mean, who is going to volunteer to have their normal eyesight irreparably damaged to play a VR game?