The Future of Home Computing Radical Changes Ahead

Computer Science
96 followers
#Future #home #computing
My take on the possible future of home computing, based on increased component integration, the rise of the agentic command interface (ACI), and increasingly hardware transparency.
My recent, related video on “Agentic AI: The Good, the Bad & the Ugly” is here: • Explaining Agentic AI: The Good, the Bad &...
AI Generated·May contain errors
Welcome to another video from explainingcomputers.com.
This time I'm going to share my thoughts on potentially dramatic changes in home computing.
These are being driven by three developments that are happening right now, and which,
in combination, recently got me thinking about where computation in the home may well be headed.
And so, I thought I'd make this video.
Developments in home computing have often been driven by increasing component integration,
or, in other words, incorporating more and more parts into a single package.
The first electronic computers were based on these, vacuum tubes,
and we then moved on to computers built from individual transistors.
But, in the 1960s, that changed because we have a development of the integrated circuit,
which have multiple transistors on a single device, on a silicon chip.
And then, in the 1970s, we had what was then called very large-scale integration.
Loads and loads of transistors being put onto one chip,
and that allowed the development of the microprocessor, and in turn, microcomputers.
And it was development of microprocessors and microcomputers
that first really started home computing.
However, increasing component integration is about far more
than making increasingly powerful central processing units.
For example, if we look across to this early 1990s PC,
we see that it contains lots of expansion cards,
with almost no connectivity integrated into the motherboard.
And so, early desktop PCs not only had separate graphics cards,
but also network cards, sound cards, and cards for serial and parallel ports.
In stark contrast, if we look inside a modern tower PC,
here, an AMD 5600G system, no expansion cards are included.
This system also relies on a GPU integrated into the CPU.
And so, here, no separate graphics card is required.
And the inside of this tower PC is rather empty.
Looking to the future, already the next phase of component integration is occurring.
Not least, in November 2020, Apple launched its M1 processor.
This has since been followed by the M2, M3, M4, and M5,
all of which are technically a SIP, or a system in a package,
which integrate a CPU, GPU, and RAM into a single component.
To be clear what's going on here, in Apple M chips,
the RAM is not included on the processor die.
But the processor die and RAM chips are all packaged into a single SIP module.
Having the RAM integrated with the CPU and GPU provides performance improvements.
And I personally think that, in time, it will become very common.
In fact, in 2024, Intel did release some core ultra laptop processors with integrated memory.
Admittedly, it then rejected the idea of doing so again,
citing the impact on its margins of having to purchase the RAM
as a relatively small buyer in the memory market.
But, given that Intel also made clear that on-package memory improved performance
and reduced power consumption,
I still think that RAM integrated into a SIP
will be the future for most home desktop and laptop computers.
It's just that the x86 platform needs time to adjust.
Integrating the CPU, GPU, RAM, and other components
clearly removes the opportunity to make individual component upgrades,
and, in turn, will impact PC form factors.
Indeed, if we go back to the Apple website and click on Mac,
notable by its absence is the Mac Pro.
Just to remind us, if we go back in time and take a look,
the Mac Pro looked like this, and was a very expensive tower desktop PC.
But, in March 2026, the Mac Pro was discontinued.
So, having transitioned to SIPs,
Apple no longer sells a computer with a tower form factor.
And, the end of the Mac Pro is one of the events
that recently got me thinking about the future of home computing.
For, if Apple believe, as they clearly do,
that even developers, professional video editors,
and other power users no longer need a tower PC,
then something has changed.
Today, whilst many of us still prefer a traditional tower desktop PC,
due to component integration, I have no doubt that, in the future,
the dominant desktop form factor will become the mini PC.
Apple is already indicating the direction of travel.
Now, without doubt, today, for the best desktop gaming experience,
you want to have a traditional tower PC with a separate graphics card.
However, decent gaming performance has already been demonstrated
on systems based on a SIP, on M-chip based Macs,
and on some of the Intel laptops that use their SIPs
with their integrated CPU, GPU, and RAM.
And, I'd be staggered if we don't get to the point
where, in 10 years' time, the best computing performance,
the best gaming experience, isn't delivered on a SIP-based system.
And, if that's the case, that's where gamers will go.
I know that many gamers like to have a system
where they can go, oh, look at this lovely tower desktop PC.
I like to say, oh, look at this lovely tower desktop PC, too.
But, if the best performance is delivered elsewhere,
that's where the market, that's where gamers will gravitate to.
And, sadly, this does mean, I think, we will see the slow demise,
maybe the rapid demise, of the tower PC form factor.
Not just due to performance issues with a SIP,
but due to the practicalities.
Because, if you've got a SIP-based system,
there's going to be very little you can plug and unplug,
probably very little at all.
And, therefore, there's no reason to have a computer
with that upgrading and changing around flexibility.
And we will, therefore, see the dominance of the mini PC.
Early home computer users were presented with a command line interface
and often had to type in their own programs.
Things then evolved to a point where computers booted
to a graphical user interface, or GUI,
with users then selecting and launching applications
Today, this is, of course, still the way that most home computing works.
However, Microsoft and others are advocating an evolution
to a new human-computer interface model.
As I discussed in a recent video,
this is intended to be based on agentic AI,
with software agents, rather than people,
increasingly being the entities to launch and use applications.
Or, in other words, if the computing industry gets its way,
the future interface model will be users asking software agents
to do things for them on computers,
rather than users directly operating most applications themselves.
So, will this really happen?
Well, there is, of course, a great deal of scepticism.
And personally, I don't use AI.
I don't use it to write anything, to come up with video ideas.
I don't use any of those things that YouTube gives us now to say,
oh, generate this idea for your video.
I don't write scripts for that.
I don't edit documents with it.
I don't allow it to scan and summarise my email.
I don't allow it to write email messages.
I don't use AI to produce images or to alter images.
I don't use AI to edit my audio,
although some people in the comments seem to think I do.
I don't. It's also edited in Adobe Audition.
I don't use AI anywhere.
Maybe I see a few summaries on the top of a Google page
and try and glance at them and move past them.
But basically, I don't use AI.
And I know lots of people don't like all this AI stuff.
And clearly, Microsoft has pulled back from some of the AI in Windows,
but it hasn't pulled back from the underlying agentic AI agenda.
And whilst lots of people say they don't like AI, they don't use AI,
I'm not quite sure that's true,
because someone is using some of this.
And whenever you have people say they don't use AI,
then go, oh, of course, I use it to summarise my emails,
or I use it to do this or to do that.
It is creeping in.
And changing computing, particularly fundamental interface change,
takes a lot of time.
And I think that the big tech companies are playing the very long game here.
And Microsoft has just pushed it too much,
and then they're pulling back again.
But on the surface, this move towards,
we will ask computers to do things for us
rather than directly using computers ourselves,
that is going to become a very, very significant
human-computer interface model.
After all, many people never really wanted a desktop PC.
They didn't want to dedicate part of their home
to having a desk or a table or something with a PC on it
and all that stuff.
They just wanted to do things the PC allowed them to do.
And if the agentic interface system takes off
and people go, oh, I can do things that way,
that's what they'll do.
You know, smart speakers, people were deeply cynical about,
not that many years ago.
They're now pretty common.
People are used to giving voice commands.
They'll do things for them and sort things out.
I think the same thing will happen with agentic computing.
And of course, this is happening in the context of the smartphone.
People have got smartphones to do lots of things around
in terms of home computing already.
And so, you know, the days of a desktop PC
operated by the graphical user interface,
for all things, I think are going to be numbered to some extent.
Yes, we will still do certain things, of course,
sitting on the computer.
But the writing, I think, is on the wall for the GUI
being the dominant interface
as we move towards increasing use of agentic AI.
Three things prompted me to make this video.
The first was Apple discontinuing the Mac Pro.
The second was doing research for my video on agentic AI
and discovering how advanced it had become.
And the third was the realisation
that I now do a lot of my own home computing on my television,
on this screen that I'm looking at in my living room right now.
Putting all of this together led me to conclude
that home computing is going to get less visible.
And in part, that's because hardware will get smaller
due to component integration.
But it's also because the move towards
agentic command interfaces
will lessen the reliance of home computers
on keyboards and rodents and touch.
And there'll therefore be less requirement for desktop PCs
and even laptops to consume prime living space.
Rather, we'll be able to put home computing devices
away somewhere, maybe in a cupboard, something like that.
And they will communicate remotely
with the devices we physically interact with.
We will command them using probably voice interfaces
a lot of the time.
And they will send content to smartphones, tablets,
maybe smart speakers or to televisions as needed.
And in some ways, we're already seeing the start of this.
We've already got smart speakers
that can send content to a television.
And indeed, I suspect that in time,
at least some desktop home PCs will converge
with smart speakers.
Now, as I just mentioned,
I'm doing an increasing proportion
of my own home computing on my TV.
And I do this using this device here.
This is a combined keyboard trackpad wireless device.
And this controls a Chrome OS Flex PC
that is hidden behind the TV.
And I first set this kind of thing up
about seven years ago now
so I could watch YouTube on the TV.
And then as they came along other streaming services
and also so I could watch my own YouTube content
before upload, I like to do a lean back test.
But I've realised in the past year or so
that most days I'm now doing work processing
using this on the television.
And I'm accessing spreadsheets
and I'm doing image editing on the TV.
And I realised very recently
that the field of view I have of the computer screen
when I'm using it on the TV to lean back like this,
is actually slightly bigger
than the field of view I have on my laptop
when it's sitting on my lap.
That's a 13.3 inch laptop.
So it's hardly surprising it works.
I'm getting a slightly bigger experience.
Things are slightly larger to look at
when I'm working on the TV screen.
And this is before any agentic AI interface in place.
I'm using traditional import devices.
And so when we have the opportunity
to have an agentic command interface
where you can say to an agent,
do this thing for me,
chuck the content to something I can see,
I can see that occurring.
Imagine, for example, in the future,
you might go to some remote agent somewhere
on a hidden desktop PC.
Could you take all of the content I shot
on my smartphone and my camera yesterday?
Could you edit it into a video
of this length in my normal style
and show it to me on the TV screen?
And it'll edit the video, put it up there.
If you'd like it, fine.
If you don't, you'll go,
could you make this change in this change?
You'll iterate like that with the agent.
Once you're happy, you'll say,
upload it somewhere.
It does it for you.
That is a perfectly reasonable,
fairly near future scenario.
And I'm not suggesting here
that all dedicated screens for desktop PCs
and even laptops in the home will disappear.
There'll be a need to have multiple screens
in homes with multiple people, not least.
But dedicated home computing space,
I think, will become a little less common.
Computing is going to be around us more.
It's going to be invisible,
built into things,
commanded in different ways
as agents take control of it
and as the hardware gets smaller.
Now, inevitably, in this video,
I am being a bit controversial
to promote debate.
And in that context,
I thought I'd also ask a broader question,
which is this.
Where will most future home computer power
be located?
And it could be in desktop
or laptop computers.
For some people,
it could also, as today,
be in smartphones or tablets.
For others, it may reside
in a hidden mini PC
accessed via a remote keyboard,
games controller, smart device,
or agentic voice interface.
Potentially, component integration
may also allow a lot of home
computational power
to migrate into TVs.
Or, as another possibility,
10 years from now,
most of the computational power
in many homes may be located
in a domestic robot or driverless car
that streams ancillary computing services
as part of its software subscription.
Or, in the future,
the bulk of home computing power
may be located in the cloud.
This is, after all,
where the uptake of generative AI
is already taking
some home computing users.
Hopefully, this video
has got you thinking.
Right now, for many people,
the future of home computing,
their own home computing,
is on hold
due to today's very high RAM
and storage prices.
And, when this situation resolves,
there is no guarantee
we're going back to a state of the world
quite like it was before.
Home computing has never stood still,
and it certainly has always been based
around tower desktop PCs.
And so, I for one,
I'm pretty certain
that home computing
could change drastically
in the future.
But now, that's it for another video.
If you've enjoyed what you've seen here,
please press that like button.
If you haven't subscribed,
please subscribe.
And I hope to talk to you again
very soon.
