Where will PC tech go ?

[ging84]

Mr Whippy said:

AI and quantum computing aren't really anywhere near?

You mean they're just buzz words... again?


I remember talking about AI in 1998/99 and how it was going to revolutionise things. And that was pattern matching AI. 20 years later it's *just* getting going on the pattern matching front.


Once AI and quantum computing really get going we'll all properly know about it because we won't have jobs any more. AI will do it all.

I never said anything about AI. You could call that a buzz word but that is more because the term has been so poorly defined that yesterday's AI is today's computing or automation.
AI is a field that is subject to some hype cycling, but has also delivered some fairly fundamental innovations in technology.
Quantum on the other hand has consistently delivered absolutely nothing outside its own field.

[eein]

ging84 said:

Mr Whippy said:

AI and quantum computing aren't really anywhere near?

You mean they're just buzz words... again?


I remember talking about AI in 1998/99 and how it was going to revolutionise things. And that was pattern matching AI. 20 years later it's *just* getting going on the pattern matching front.


Once AI and quantum computing really get going we'll all properly know about it because we won't have jobs any more. AI will do it all.

I never said anything about AI. You could call that a buzz word but that is more because the term has been so poorly defined that yesterday's AI is today's computing or automation.
AI is a field that is subject to some hype cycling, but has also delivered some fairly fundamental innovations in technology.
Quantum on the other hand has consistently delivered absolutely nothing outside its own field.

AI is not a factor of innovation. It is simply a useful technique (when used in the right way in the right place...) to code a lot of complex computing quite quickly, run quickly, update more easily and be create by far lower skilled people. AI, or more specifically Machine Learning Neural Networks, are just transfer functions that could other have been created in a fixed algorithm or other adaptive technique. So they only enable innovation in as much as any technique does that allows more stuff to be done quicker and easier.

[red_slr]

What is incredible to me is the vast majority of our customers order using tablet or phone. Only around 20% use a PC.

I think other than for gaming which is becoming ever more niche then the actual physical desktop PC will slowly die.

I think the old "all in one" PC might make a come back once you can get a CPU/GPU that will run off a small PSU.
Maybe even see TVs becoming PCs.

Tablets are basically oversized mobile phones so there needs to be a big step in thermal management before we see really powerful mobile devices. Its slowly coming now and IIRC Dell are the first to have laptops with new cooling for the CPU (top spec XPS IIRC).

However, I can see a "base station" type device with GPU/CPU being a thing and then display being pushed over WiGig style connections. Several users can hang off the same "base" with a tablet but that can run serious power for games etc.

[theboss]

ash73 said:

Desktop streaming - move it all into the cloud, subscription based, rent gaming pc performance when you need it, 60fps to any simple device with a browser.

The tech around graphics remoting is becoming seriously impressive. I’m working on a project right now to deliver 3D graphics workstations from Azure and it’s fascinating. Plenty of people demoing FS2020 from cloud hosted GPU accelerated desktops in the last week or so. Not sure how well it works for low latency / fast paced gaming but it’s certainly moving in the right direction.

[ging84]

red_slr said:

What is incredible to me is the vast majority of our customers order using tablet or phone. Only around 20% use a PC.

Very much depends on what you are selling.
There is a huge amount of effort that goes on into tracking user journeys across devices. Often a purchase can start with someone looking up on a phone out and about, then completing the purchase at home on a laptop or tablet where they feel more comfortable about making a decision. It can also happen the other way round, particularly if you are sending out email reminders or offers based on what people browsed. There is also a lot of browsing that goes on, or atleast used to go on at work that turns into a purchase on another device where all their payment details are saved.

[J4CKO]

red_slr said:

What is incredible to me is the vast majority of our customers order using tablet or phone. Only around 20% use a PC.

I think other than for gaming which is becoming ever more niche then the actual physical desktop PC will slowly die.

I think the old "all in one" PC might make a come back once you can get a CPU/GPU that will run off a small PSU.
Maybe even see TVs becoming PCs.

Tablets are basically oversized mobile phones so there needs to be a big step in thermal management before we see really powerful mobile devices. Its slowly coming now and IIRC Dell are the first to have laptops with new cooling for the CPU (top spec XPS IIRC).

However, I can see a "base station" type device with GPU/CPU being a thing and then display being pushed over WiGig style connections. Several users can hang off the same "base" with a tablet but that can run serious power for games etc.

Sounds possible but I find with tablets, the games are already pretty impressive even on my five year old Ipad, considering it is passively cooled and not that high performance in the scheme of things.

The limitation is the human sat playing it, the methods to interact, the screen size, tablet gaming is generally pretty limited, FS2020 or any flight sim is going to be crap on a tablet, same with racing games, I like Forza Horizon on my big telly and a joypad does the trick but would be werd propping a tablet up and using a pad.

[CubanPete]

My home PC is ten years old, and runs everything I use it for completely fine, including the odd game.

My work PC (next to my home PC) is a few years old, but high spec, core i7, decent video card, 32gb RAM. It still takes time to calculate FEA models, and they are simplified too and large CAD models can be stuttery on screen. So significantly faster processing / bus speeds would allow me amend models and to look at FEA dynamically. But it would need some clever software to do this.

[eein]

ash73 said:

I can't see fat-client Windows 10 still existing in 20 years. Why bother with a complex ecosystem of OS and apps for different devices? Just run it all in the cloud and deliver a 60fps display to whatever device you are using.

It will still exist - cloud will not take over everything... indeed it's seem a bit of a claw back to non-cloud recently. Security concerns and the lack of ubiquitious low latency free high bandwidth internet will prevent everyone going 'thin client'. I agree this will become common, but it wont kill off fat client OS in a 20 year timeframe.

I can of course see a fat client with seamless offload / backoff to the cloud - microsoft's shift in strategy to the cloud will provide this integration to the masses.

Another possible path over the next 20 years for personal (ie domestic) computing is where some companies are going with some form of separation of internet connected activity from local activity. This is currently becomming common in companies to take their local assets 'offline' to avoid cyber attack and then enable employees to access the wild internet via some hypervisor 'down browse' (there's other ways to do it instead of hypervisors). If the internet gets more dangerous, which the introduction of more encryption will do (DoH, ESNI, etc) then it will become such a big threat to consumers and all their assets (photos, files, etc) that solutions to truely separate these off the internet could appear. If this happens it will only do so if implemented in a near seamless way to the users... some of the corporate solutions for this today are getting fairly seamless.

[anonymous-user]

Landfill.

[eein]

gottans said:

Landfill.

An important angle of course.

I suspect within 20 years not only will recycling of electronics get better and therefore the norm, that other factors will result in an indusctry that digs up landfills to find and extract old electronics for recycling.

There's gold in them (landf)hills! (and other valuable stuff).

[anonymous-user]

eein said:

gottans said:

Landfill.

An important angle of course.

I suspect within 20 years not only will recycling of electronics get better and therefore the norm, that other factors will result in an indusctry that digs up landfills to find and extract old electronics for recycling.

There's gold in them (landf)hills! (and other valuable stuff).

That is true but it is really down to whether it is more economic to recover these materials or dig up more. In truth I can't see the human race doing that much different in the future.

[eein]

gottans said:

eein said:

gottans said:

Landfill.

An important angle of course.

I suspect within 20 years not only will recycling of electronics get better and therefore the norm, that other factors will result in an indusctry that digs up landfills to find and extract old electronics for recycling.

There's gold in them (landf)hills! (and other valuable stuff).

That is true but it is really down to whether it is more economic to recover these materials or dig up more. In truth I can't see the human race doing that much different in the future.

I guess my prediction is that in the next 20 years the cost of digging up more materials will become higher than digging through landfill to recover. I suspect the cost will not just be direct cost, probably some eco tax and reputational cost will contribute to the shift.

If you look at the advances in recycling technology they are phenomenal. I was reading recently about lithium battery reuse and recycling and it's on the cusp of becomming very viable.

[eein]

ash73 said:

eein said:

cloud will not take over everything ... [snip] ... it wont kill off fat client OS in a 20 year timeframe.

I'm pretty sure it will. Think back 20 years, we were still on dial-up... broadband wasn't even a thing.

They can already desktop stream today. Imagine where it will be in another 20 years.

Why even have a fat-client OS on a phone, let alone a PC?

The cloud utopia relies on truely ubiquitious internet - quality, bandwith and latency. If your life is mostly in a city or major town environment today then you'll think we're already there, if you step outside these areas you'll see we're not and the efforts to make it better are just not happening. This has not changed in the last 20 years - I remember standing in the stand at the British F1GP in 2000 excited to use my phone's data conenction to keep an eye on the lap times (remember WAP?!). It worked well, but it turned out this was only becuase the mobile networks had put up event trasmitters for the weekend in rural Silverstone.

Indistinguishable cloud will definetly be an option in 20 years, but I doubt it'll be the only option as there will still be too many places that still need a 'local' service. Of course under the hood that 'local' service may just be a local instance of whatever VMs and containerisation has evolved in to in 20 years so the service running will be identical to that in the cloud, just happens to be running locally on your PC/tablet/phone/other future thing.