Pretty good article, but I think he misses the point on a few things.
For many, many issues, doing field repairs just isn't practical. If the problem is a high-density IC, replacing that takes very specialized, very expensive equipment, and very specially trained operators. Those parts are soldered en-mass on the production line, and removing and replacing just one of them is not simple.
So you can't really expect manufacturers to have repair shops set up, just to handle the repairs that are do-able. Couple that with people actually do want the newer wiz-bang features, and repairs are less practical. Your repair is competing against the new product. And with all that specialized equipment and training, expect an analysis/repair to cost $100/hour.
And I just can't see the manufacturers providing repair guides for the average Joe. Those cost money to create and keep up to date, and support the inevitable questions. And product cycles are very short these days, so they'd need to constantly be doing this.
But things certainly could be better than they are. More modular designs so that a bad module can be swapped out, etc. Before I bought the last round of Apple MacBook Pros for the family, I verified that t was pretty easy to get to the hard drive and memory and fan - things you might be likely to replace or upgrade if you keep your computers a while. But the more recent ones look pretty bad in this regard - soldered on memory chips, no swap/expand, etc.
Sure, the average person will just buy new, but I will keep trying to find laptops with easy to replace modules (and load Linux on them).
An example: I bought Sonicare toothbrushes. The instruction say that when the battery no longer holds a charge, wrap it in a towel, and (I am not making this up!) hit it with a hammer to crack open the case. Then disconnect the battery, and recycle it.
That sounds pretty bad. But OTOH, if the battery has an expected 10 year life, and the motor/actuator/thingee has a similar life expectancy, maybe it's not so crazy?