We're doing a very dangerous thing here. We're crossing the line between computers computing, and computers advising. When a computer computes, you can be assured the answer is correct. When a computer advises, the average person takes that information at the same value. There's a long laundry list of reasons why this is horrible. I went over a few in my essay about platforms being the enemy. http://tiny-giant-books.com/Entry2.html?EntryId=recEUbufzhAv...
There's a deeper moral/ethical issue at stake here. If you program a computer, you are responsible for that computer never presenting a misleading view of reality to the user. People don't distinguish between their tax program, GPS, and voting on reddit. When you use the interface to guide or subtly mislead people? You're hurting millions of people just a tiny bit at a time.
So many net-level effects involve these tiny changes that are impossible to evaluate. That delays or completely obfuscates the feedback loop. Very bad stuff here.
There's a deeper moral/ethical issue at stake here. If you program a computer, you are responsible for that computer never presenting a misleading view of reality to the user. People don't distinguish between their tax program, GPS, and voting on reddit. When you use the interface to guide or subtly mislead people? You're hurting millions of people just a tiny bit at a time.
So many net-level effects involve these tiny changes that are impossible to evaluate. That delays or completely obfuscates the feedback loop. Very bad stuff here.