Tesla Enhanced Autopilot
In June 2013, Tesla CEO Elon Musk announced he would drive a Tesla Model S electric car across the U.S. to show off the company's new Supercharger fast-charging network.
Indeed, two Tesla Model S sedans successfully completed the run from Los Angeles to New York City in February 2014.
Last year, Musk decided to repeat the stunt, but with a twist: This time, a self-driving Tesla would take him across the country.
DON'T MISS: Tesla upgrades self-driving sensors, hardware; full autonomy test next year? (Oct 2016)
Musk and the rest of the company, however, have been quite preoccupied with a tougher-than-expected task: getting its less-expensive Model 3 compact electric sedan into volume production.
While 30 hand-built examples were delivered to the first buyers last July, fewer than 2,000 Model 3s were delivered during 2017 as a whole.
Musk has made frequent references to the "production hell" in which the Model 3 and Tesla were stuck, though his comments earlier this month suggested prospects were looking up for higher volumes this year.
Elon Musk signs new 2013 Tesla Model S at Tesla Store opening, Austin, Texas [photo: John Griswell]
And what of the self-driving Tesla trip?
As Musk acknowledged in a conference call for financial analysts on February 7, he had missed the deadline—the end of 2017—he had set for the trip:
I've been meaning to address this because obviously I missed the mark on that front. Our focus is very much on Model 3 production, so everything else kind of took a second place to that.
CHECK OUT: Why did Tesla take capabilities out of latest Model S + Model X? (Nov 2016)
But we could have done the coast-to-coast drive, but it would have required too much specialized code to effectively game it or make it somewhat brittle and that it would work for one particular route, but not the general solution.
So I think we would be able to repeat it, but if it's just not any other route, which is not really a true solution.
There followed a lengthy discussion of Musk's excitement about the applications of neural networks to autonomous-vehicle technology, and a comparison to Google's AlphaGo self-driving software.
Tesla Model S at Supercharger site in Ventura, CA, with just one slot open [photo: David Noland]
"Timing-wise," he concluded, "I think we could probably do a coast-to-coast drive in three months, six months at the outside."
And, he confirmed, that would be using software that would be immediately available to Tesla owners whose cars have the appropriate suite of sensor hardware.
Musk reiterated his belief that competitors who plan to use lidar in their self-driving systems would end up with inferior technology.
Using lidar would be a "crutch," Musk suggested, leaving those competitors with cars that include "a whole bunch of expensive equipment, most of which makes the car expensive, ugly and unnecessary."
Tesla plans instead to use what he called a "sophisticated neural net ... capable of advanced image recognition" that he envisions, combined with advanced radar detection.
"Perhaps I am wrong," Musk concluded. "In which case, I'll look like a fool. But I am quite certain that I am not."