I used to work as a scientist in an engineering company. To understand the difference, you could ask both a scientist and an engineer what π² is. The scientist will tell you it is a number with an infinite number of digits close to 9.8696. The engineer will tell you it is about 10. After working for long years with engineers I understand that finding a good enough solution fast enough is often more useful than having the most exact solution. Engineering is one of the most useful things a person can learn, and many of the big problems humanity is facing, like for example climate change, will need at least in part engineering solutions.

Having said that, engineers frequently think that everything is an engineering problem. I was thinking of that when reading the latest news about Elon Musk telling Twitter employees that he wants them to work "hardcore" on coding to get the current Twitter to some ideal Twitter 2.0 state. He pretty much told the same thing to his employees as Tesla; but creating an AI for an electrical car is an engineering problem, and creating a non-toxic "digital town square" for everybody is not an engineering problem.

Selling the blue verified checkmark on Twitter for $7.99 was clearly as mistake, although there was nothing wrong with the coding. My first thought when I heard it was that it would have been better to do the opposite, to ask that every Twitter user either lets his identity be verified, or pays $7.99 for the luxury of remaining anonymous. The fake Pepsi account stating that "Coke is better", and the fake Coke account offering to put cocaine back into Coca Cola sure were funny; Eli Lilly losing billions on the stock market because of a fake "insulin is now free" tweet probably less so. It just shows that if everybody has access to a verified blue checkmark on Twitter, then "verification" doesn't mean anything anymore, and we need to treat all tweets as fake.

There is probably an engineering solution to a non-toxic Twitter that involves biometric facial recognition as verification with an absence of anonymity and legal consequences for criminal abuses of "free speech" equal to the existing consequences in the real world. However, that would not be a mass market product, and would be very hard to monetize. The current version of Twitter probably could be improved, but I think that would involve mostly psychologists, behavioral economists, and maybe even anthropologists, not mainly engineers. Elon Musk might be the wrong man for the job.

- posted by Tobold @ 4:37 PM Permanent Link