How can you be sure that you understand quantum mechanics, if you have never studied it?

When laptops plummet in price, so do margins on laptops.

I feel there's still money to be made in the high-end market, and I do not feel laptops will ever match desktops in performance/$. If you have two components of the same architecture, same performance, the one that is larger in size is probably cheaper to produce.

Taking your car analogy further, trucks and larger vehicles are still being sold today; they are still useful. There are still pianos that need to be moved.

So I am puzzled when you say:

Some use cases/applications that I can think of that would benefit from the higher performance of a desktop:

1. PC gaming market is still going to be around
2. Growth of 3D media, I certainly see a need for higher performance computing here
3. Media creation/editing (3D modeling, video editing, etc.)

Given that people will have remotely-accessible storage residing in the home for their portable devices, I could also imagine people running a more powerful personal server machine in their homes.

Perhaps your mobile applications can offload expensive calculations to your home server, who knows? Bringing the data center to the home, in a way.

And so I feel that's another area where there's a need for more powerful stationary computing.

The article references a Forrester Research report, and the article really missed an important point in the original report:

http://forrester.com/rb/Research/us_consumer_pc_market_in_2015/q/id/57210/t/2

http://blogs.forrester.com/sarah_rotman_epps/10-06-17-steve_ballmer_right_pc_market_getting_bigger

Well, when I search for restaurant reviews on Google, I often see Yelp near or at the top of the results.

With online reviews, I would agree with you if the establishment in question has only a handful of reviews.

On the other hand, if I see a place with 150+ positive online reviews, it's likely that place has good food.

Yes, there are shills, unreasonable customers, people with poor taste, not every good restaurant will be reviewed, some duds will get good reviews, etc., but it's better than having no information at all.

Technically, you could implement such intricate server-side dependencies, but I'm not convinced that makes business sense.

Developing this code will be more costly because you've now made your single player, locally run game a networked game. You will have increased support costs as customers complain about their game not working because of some network failure. Your game is now vulnerable to denial of service attacks. You now have to maintain servers for your game, which incurs maintenance. hardware, bandwidth, electricity costs. You will get bad press, and drive away some % of your paying customers.

You will potentially get some % of the pirates to buy your game. Is this worth it?

I think the author's idea that this is an experiment has merit, I do remember when Ubisoft released a Prince of Persia game without DRM to see how it sell.:

"But this will make everyone hate them.

Perhaps. Make no mistake. Ubisoft will lose customers and earn much nerdrage over this. But they are engaged in a grand experiment. They are seeing if an adequately pirate-proof game can make money. Will keeping cracked copies off the Torrents for a month make extra sales? And enough extra sales to make writing PC games worthwhile? Because the current system, where 90% of the copies out there are pirated and only megahits that could turn a profit on that 10%, doesn't seem to be working."

That said, I think he overestimates the difficulty of cracking this specific implementation.

I think the intent is closer to:

"Cheating kills your game because it makes legit players not want to play it anymore, whereas pirates don't affect your legit users"

If you really wanted to do hardware-specific optimizations today, you'd probably be better off finding parts of the code where you can execute things in parallel. You could get good performance benefits from multicore processors.

1. that would be an absolute nightmare in terms of code maintainability, debugging, etc. Windows source code today is far larger than Windows 3.0 code

2. compilers have improved since the Windows 3.0 days, they're far better than humans at optimizing in many cases

3. Itanium architecture is explicitly parallel, writing assembly for that is completely different from writing x86, it's far from a simple conversion

4. Game performance is limited mostly by your GPU, and CPU usage by your OS is minimal compared to CPU usage by the game itself. Gamers won't see much benefit from this.

The effort required to write and maintain the code in assembly is so great that it's infeasible in practice, and you wouldn't really get that much benefit from doing it.

whoops, accidentally modded you down, posting to remove mod

then you just don't like the Beatles, no music has universal appeal