Are Apple Silicon Macs ‘fast enough’?

[Bodhitree]

In years to come all this will be irrelevant because computers won't exist as we currently know them where the heavy lifting is carried out locally and determined by the specification of your computer. Computers will just be windows (pardon the pun) to a virtual machine that you access through terabit internet, OS and all. Processing power, storage capacity and RAM will be things you subscribe to on monthly or yearly plans; your machine itself will only need the absolute bare minimum processing power required to shuttle huge amounts of data to and from the internet. Playing AAA games will require no greater local processing power than your TV currently needs to stream and display a 4K film. Sounds far-fetched? Not really. Try telling someone in 1993 that by 2023 features in your new car will be enabled or disabled remotely over the internet by the manufacturer depending on whether you subscribe to optional services.

That is one possible future. It depends on how small and how cheap computing devices get. An alternative future might be that in another twenty generations computers will be 20x as fast, fit into a lapel pin and cost $10, so that compute power is everywhere rather than being piped across the net.

In that kind of a future, low-power architectures will have a key advantage, and interfaces will likely have been reinvented. It all depends on what happens first, pervasive, publicly available glassfiber data cabling for fast internet, or cheap and small compute power.

Certainly bigger jobs will get put in the cloud more and more. If you had asked a desktop publishing professional whether he’d ever trust Adobe to store his data on the internet thirty years ago you’d probably have gotten a ‘zero chance’ back, but look what happened. Software makers are not stupid, they know that in extracting maximum value from their users they have to control the process.

 

[Joe Dohn]

In years to come all this will be irrelevant because computers won't exist as we currently know them where the heavy lifting is carried out locally and determined by the specification of your computer. Computers will just be windows (pardon the pun) to a virtual machine that you access through terabit internet, OS and all. Processing power, storage capacity and RAM will be things you subscribe to on monthly or yearly plans; your machine itself will only need whatever is the absolute bare minimum processing power required to shuttle huge amounts of data to and from the internet. Playing AAA games will require no greater graphics-ability or number-crunching ability on your local machine than your TV currently needs to stream and display a 4K movie.

Sounds far-fetched? Not really. Try telling someone in 1993 that by 2023 features in your new car will be enabled or disabled remotely over the internet by the manufacturer depending on whether you subscribe to optional services. The future of computing (probably of everything) will be wholly subscription-based, to keep you in the honey-trap of perpetual payment.

I don't think so. If we had that vision, it would make sense to make computers dumber over time, not more powerful – after all, why bother giving the user power at all if we can process everything in a server?

The flaw with processing everything in a server, naturally, is that you can easily overload it when everyone accesses it. And if the server is overloaded, then no one can access it.

Instead, the path we are taking is that we process everything we can locally, and then we offload the tasks that the user can't do locally in the server (e.g, AI processing). Even so, there are many nuances to that; e.g, Google Docs is rather trivial, but is processed on the cloud; some applications are on the cloud, but use the browser to make the user process a few request and then synchronize it, and so on.

 

[yellowhelicopter]

The question is, will there be another such an ah-ha moment, when computers double in speed again compared to the M1?

If you mean a quality leap like the (successful) transition to ARM was for Apple and performance boost it brought, there will be, but definitely not anytime soon. Now it will develop steadily and relatively slow in terms of performance increasing.

 

[Joe Dohn]

If you mean a quality leap like the (successful) transition to ARM was for Apple and performance boost it brought, there will be, but definitely not anytime soon. Now it will develop steadily and relatively slow in terms of performance increasing.

There's quantum computing. A quantum chip is so much more powerful than a regular chip that it renders our cryptography algorithms useless in seconds.

However, there are two problems with this.

1. If they became popular before regulation, they would could chaos in the current banking system. Of course, we could address this by making those firms adopt quantum computers first and designing more powerful cryptography systems, but banks are conservative by default, so it would take a while.

2. Quantum computers need to be cooled at sub zero temperatures to remain stable, and we haven't found a way to stabilize them commercially.

The day point 2. is addressed, I'm sure we will have another massive breakthrough. But that could take a while.

 

[MajorFubar]

I don't think so. If we had that vision, it would make sense to make computers dumber over time, not more powerful – after all, why bother giving the user power at all if we can process everything in a server?

Because were not there yet. Not even nearly. In terms of having internet connections and servers so fast that VMs can run everything, we're currently in 1983 compared to 2023. Currently the money is in selling faster and faster computers. But that won't always be the case, and IMO will change when the infrastructure exists to shuttle gigabytes and gigabytes of data to and from immensely powerful cloud-based computers (could be quantum computers as per the post before this) each serving thousands of users, which subscribers will access as a service, like streaming a 4K movies from Netflix but the experience will also be realtime interactive like using a 'real' computer.

 

[Joe Dohn]

Because were not there yet. Not even nearly. In terms of having internet connections and servers so fast that VMs can run everything, we're currently in 1983 compared to 2023. Currently the money is in selling faster and faster computers. But that won't always be the case, and IMO will change when the infrastructure exists to shuttle gigabytes and gigabytes of data to and from immensely powerful cloud-based computers (could be quantum computers as per the post before this) each serving thousands of users, which subscribers will access as a service, like streaming a 4K movies from Netflix but the experience will also be realtime interactive like using a 'real' computer.

We're not in 1983 at all. While some places do lack an Internet connection, many urban centers have at least a rudimentary form of it.

The problem is that if you rely on an Internet connection, you're expecting it to remain stable 100% of the time. It doesn't make sense to overload servers with "simple" things, such as checking your photos or typing a text document. So, even in a world with rock-solid Internet services (which is not currently feasible in real life), you would be overloading servers in peak times for very trivial reasons. Which is why I simply don't see the cloud being used for ALL one's computing needs.

 

[thebart]

Not according to my mom. If she clicks something and there's a spinner for 3 seconds the Wi-Fi is bad. If she taps the screen and it takes the keyboard half a second to come up she needs a new phone. If she moves the mouse and the screen doesn't turn on immediately why don't I get myself decent computer?

 

[MajorFubar]

We're not in 1983 at all. While some places do lack an Internet connection, many urban centers have at least a rudimentary form of it.

The problem is that if you rely on an Internet connection, you're expecting it to remain stable 100% of the time. It doesn't make sense to overload servers with "simple" things, such as checking your photos or typing a text document. So, even in a world with rock-solid Internet services (which is not currently feasible in real life), you would be overloading servers in peak times for very trivial reasons. Which is why I simply don't see the cloud being used for ALL one's computing needs.

I'm not quite sure why you're not getting what I mean. My reference to 1983 vs now was an analogy, and has nothing to do with who or what currently does/does not have internet.

What I meant was, compared to the technology I'm talking about, we are at '1983' levels of technology and infrastructure, where someone connecting to the infant internet with the handset of their landline-phone plugged into a 1,500baud acoustic coupler thought they were living the dream. Now look at what we've got. And the same will happen again.

And when it does (be it another 40 years, or less, or more), I'm convinced the transition to subscription-funded cloud-based computing for most domestic users is inevitable, and for those people it will entirely replace the need for super powerful home computers, because they will be able to rent whatever power and specification they need, and it will be delivered to them as a service, in the same way we currently subscribe to TV services, with virtually zero detectable latency. Maybe not in my lifetime because I'm >50. But gen z's will see it. Maybe millennials too. The timescale is difficult to predict. But I think the inevitability is 100%.

(edited for further clarity.)

 

[Joe Dohn]

And when it does (be it another 40 years, or less, or more), I'm convinced the transition to subscription-funded cloud-based computing for domestic users is inevitable, and it will entirely replace the need for super powerful home computers, because you will be able to rent whatever power and specification you will need, and it will be delivered to you as a service, in the same way we currently subscribe to TV services, with virtually zero detectable latency. Maybe not in my lifetime because I'm >50. But gen z's will see it. Maybe millennials too. The timescale is difficult to predict. But I think the inevitability is 100%.

Even if we argue on infrastructure alone, there's still privacy issues. Whenever your files are on the cloud, not only they are subject to leaks and you are at the mercy of the people hosting them, but you have no privacy. Local computers, on the other hand, will always give the privacy you need.

I don't see this as something that can easily be solved.

 

[MajorFubar]

Even if we argue on infrastructure alone, there's still privacy issues. Whenever your files are on the cloud, not only they are subject to leaks and you are at the mercy of the people hosting them, but you have no privacy. Local computers, on the other hand, will always give the privacy you need.

I don't see this as something that can easily be solved.

That is a very valid concern of course...but it seems to be one which we are already mostly comfortable to live with, given the millions and millions of Apple iCloud / Google Drive / Microsoft One Drive subscribers. There are thousands if not millions of servers already out there holding and shuttling your personal data.

If I didn't know better I would say we are already being gently groomed into the inevitable widespread acceptance of cloud-based computing.

 

[Joe Dohn]

That is a very valid concern of course...but it seems to be one which we are already mostly comfortable to live with, given the millions and millions of Apple iCloud / Google Drive / Microsoft One Drive subscribers. There are thousands if not millions of servers already out there holding and shuttling your personal data.

If I didn't know better I would say we are already being gently groomed into the inevitable widespread acceptance of cloud-based computing.

Speak for yourself. I don't share all my content in the cloud, and I'm definitely not comfortable giving ALL I have to Google or Microsoft.

 

[MajorFubar]

Speak for yourself. I don't share all my content in the cloud, and I'm definitely not comfortable giving ALL I have to Google or Microsoft.

I wasn't really speaking for myself. I was speaking in general. It's really not possible to exist anymore without your personal data being stored electronically somewhere. So fully-online computing is just an extension of that.

The biggest security concern rn (all be it the least likely) is ID fraud caused by e.g. hackers, phishing, fake websites, or the professional mishandling of personally-identifying information which is already stored on countless servers worldwide about you. Unless you choose to exist on the fringe of society without even a bank account (and definitely not with an online presence, which you obviously already have or else you wouldn't be here), you have very little to worry about what Google and Microsoft might do with something innocuous like pictures of your pet cat. But feel free to be paranoid about it. That's 100% your right.

 

[Joe Dohn]

I wasn't really speaking for myself. I was speaking in general. It's really not possible to exist anymore without your personal data being stored electronically somewhere. So fully-online computing is just an extension of that.

Of course that SOME of my personal data will exist somewhere. But trusting all of it to a random company? Not at all. E.g, I find what Windows does of attempting to set up One Drive automatically very intrusive and annoying. And I definitely do NOT trust all my files to my iPhone either.

 

[MajorFubar]

That's fine. But all I'm saying is, the concerns you have about storing what you consider to be private information on some large enterprise's warehouse servers are dwarfed into insignificance by the personal information which is already stored about you. That is all.