Hello World
Page 27
fn5 A step up from partial automation, level 3 vehicles like the Audi with the traffic-jam pilot can take control in certain scenarios, if the conditions are right. The driver still needs to be prepared to intervene when the car encounters a scenario it doesn’t understand, but no longer needs to continuously monitor the road and car. This level is a bit more like persuading a teenager to do the washing up.
fn6 At the time of writing, in February 2018, the ‘full self-driving hardware’ is an optional extra that can be paid for at purchase, although the car is not currently running the software to complete full self-driving trips. The Tesla website says: ‘It is not possible to know exactly when each element of the functionality described above will be available.’ See https://www.tesla.com/en_GB/blog/all-tesla-cars-being-produced-now-have-full-self-driving-hardware.
fn7 This is a real product called ‘The Autopilot Buddy’ which you can buy for the bargain price of $179. It’s worth noting that the small print on the website reads: ‘At no time should ‘Autopilot Buddy™’ be used on public streets.’ https://www.autopilotbuddy.com/.
Crime
fn1 On this occasion, when I say ‘we’ I actually mean it. This particular study was a paper I worked on with my marvellous PhD student Michael Frith.
fn2 This number relates to how many true matching faces the algorithm missed when it was tuned to avoid misidentification. There’s more on the kinds of errors an algorithm can make in the ‘Justice’ chapter, and the different ways to measure accuracy in the ‘Medicine’ chapter.
Art
fn1 Then again, my 14-year-old-self was also a big fan of PJ and Duncan, so what does she know?
fn2 It’s unfortunate that this book’s format doesn’t lend itself to include musical excerpts, because I really want you to hear how hilariously bad this song is. Google it, will you?
fn3 I’d also thoroughly recommend looking up some of Cope’s music online. I think the orchestra piece in the style of Vivaldi is my favourite: https://www.youtube.com/watch?v=2kuY3BrmTfQ.
fn4 You can also use predictive text to ‘compose’ some text in your own style. Just open a new note and seed the algorithm with a few words to kick you off, like ‘I was born’. Then just repeatedly hit the words that pop up on the screen. Here’s mine (genuinely). Starts off fine, gets a bit more stressed towards the end: ‘(I was born) to be a good person and I would be happy to be with you a lot of people I know that you are not getting my emails and I don’t have any time for that.’
fn5 Cope also found it was necessary to keep track of a number of other measures as the algorithm went along. For instance, the length of a phrase and the length of a piece proved to be integral to a Bach-like output.
Conclusion
fn1 I actually met the CEO of this particular company for an interview. I asked him if he’d ever validated his algorithms to see if they lived up to the claims, and he launched into a long anecdote about how the neural network’s analysis had led to a major Hollywood star being dropped from a movie franchise. When I pointed out that was evidence that people bought into his algorithm, not that the algorithm worked, he said: ‘Well, we’re not running an academic exercise.’
fn2 There’s a trick you can use to spot the junk algorithms. I like to call it the Magic Test. Whenever you see a story about an algorithm, see if you can swap out any of the buzzwords, like ‘machine learning’, ‘artificial intelligence’ and ‘neural network’, and swap in the word ‘magic’. Does everything still make grammatical sense? Is any of the meaning lost? If not, I’d be worried that something smells quite a lot like bullshit. Because I’m afraid – long into the foreseeable future – we’re not going to ‘solve world hunger with magic’ or ‘use magic to write the perfect screenplay’ any more than we are with AI.