He told me that, when desktop publishing started happening in mid-'80s, the consensus in his business was that computers wouldn't have much of an impact on "real printing" because the kerning you got from digital sources was so bad.
When I hear people saying "there's no way a computer could do that properly", I think of 1985 print shop owners confident in how their handmade kerning provides value to customers.
He used to do a lot of the game manuals for a video game company and they used to give him finished video games every year for us kids to play.
I'd say 1 is the hardest for computers and 2 and 3 are the hardest for humans.
There's a cliff in artificial intelligence that people have been saying is just on the horizon for decades. Eventually they'll be right.
Did it?
My belief has always been that all design work should be measured and have to withstand the "does it actually provide value" test, since otherwise, how do you know one design is better than another?
I'm going to say that yes, 100%, it provides value by creating a more appealing display. Frankly, it's embarrassing to have your brand displayed without the typography handled properly. If you don't see the difference then maybe you should do yourself a favour and not delve into the art of type because there is no turning back once you start noticing the bad stuff.
Our marketing department swore by bad design. Not just simple, accessible. Intentionally bad. They believed it grabbed more attention, and that their customers felt safe/found them more trustworthy.
When we tried out more polished marketing we had measurably fewer responses.
Not the same thing.
I would be interested in a simulation based on big data or whatever, but I am passionate about typography as art.
But eventually digital became good enough to be largely indistinguishable to the unprofessional eye, while also being 10x cheaper.
Point being: the print shops may have misinterpreted "customers want professional-looking results" as "customers want handmade kerning on offset plates". The specific technique wasn't really relevant, even though it seemed that way to the ingrained pros. (There's perhaps a lesson there to today's web developers who spend their time arguing about frameworks, languages and flat design icons rather than customer-oriented results.)