> If we asked this question in similar terms just a few years ago we wouldn’t have things like consumer VR and AR.
Things won't stagnate just because the sales pitch for a new product might be "no new use cases, it 'only' makes things smoother and a quicker and higher detail".
Getting that out of the way, not every increase in power is equally useful. It's a fair question to ask. VR is good and newly enabled but that doesn't mean we're going to get a bunch of tech with similar potential in the next handful of years.
> Phones in your pocket are performing on-device ML and AR in ways previously thought impossible. They’re being used to shoot actual movies that are shown in actual theaters.
The fancy algorithms are somewhat important to picture quality, but for photos slower would be okay and for video you could run it in post without much disruption.
> Low power tech like smartwatches wouldn’t be possible without these breakthroughs because ultimately faster processors also imply low power devices that can actually do decent amounts of computation.
On the other hand, we've had pretty complex devices that run on button cells available for decades. And a lot of that power is taken up by the screen and radios rather than processing. And... wait a second, the original Apple Watch came out so long ago, it was around 8x slower than the current model.
Basically what you’re saying is that we can get by with worse tech.
Well, yes. You can use a Casio calculator watch from the 80’s to make calculations. We got by just fine before computers existed at all.
But that’s not really the point. The point is that there is value in simple raw horsepower. If there wasn’t, all these chip manufacturers would pack up their bags and shut down their R&D departments, because businesses don’t pay people to make things that have no market value.
It is an inevitability, in my opinion, that extra horsepower will enable some kind of new experience, however minor. Your example of the original Apple Watch is perfect here. The original Apple Watch barely functioned, for staters, and had a reputation for being slow. It didn’t have GPS, cellular, always-on screen, or a number of other features that need a very powerful and power efficient chip, display, and wireless modem/chip. All of these features are things that have market value.
I’m honestly not sure what you’re trying to say we should do here, not try to make computing faster? Why?
> Basically what you’re saying is that we can get by with worse tech.
> I’m honestly not sure what you’re trying to say we should do here, not try to make computing faster?
No. I'm saying that if you look at tech, generation by generation, faster is sometimes just faster. It doesn't always enable new use cases.
There's no extra subtext. I'm not saying we should stop making things faster. I just don't want the benefits of increased speed/efficiency to be overstated.
I don't think "we sorely need faster matrix multiplication" is true. It's just a nice-to-have. Which is still enough reason to make it! But it's a different level of benefit.
> It is an inevitability, in my opinion, that extra horsepower will enable some kind of new experience, however minor.
Sufficiently minor things aren't "genuine new use cases" that we "sorely need".
Eventually you'll hit a new use case, probably, but you might get zero new use cases for a long time.
> Your example of the original Apple Watch is perfect here. The original Apple Watch barely functioned, for staters, and had a reputation for being slow. It didn’t have GPS, cellular, always-on screen, or a number of other features that need a very powerful and power efficient chip, display, and wireless modem/chip. All of these features are things that have market value.
Not very much of that depends on computation getting faster.
I replied to your other comments but now I see that you are arguing in bad faith and I regret the time investment.
> It doesn't always enable new use cases.
It doesn't have to always enable new use cases, you are just moving the goalpost.
> Sufficiently minor things aren't "genuine new use cases" that we "sorely need".
A use case is a use case, you don't get to redefine what you consider a "major" and a "minor" use case to justify your bad arguments. That's another example of moving the goalpost.
> Not very much of that depends on computation getting faster.
That statement is so objectively wrong, power efficiency is a trade off with processing power, if we can make a faster chip we can make a similarly capable chip that consumes less power.
Things won't stagnate just because the sales pitch for a new product might be "no new use cases, it 'only' makes things smoother and a quicker and higher detail".
Getting that out of the way, not every increase in power is equally useful. It's a fair question to ask. VR is good and newly enabled but that doesn't mean we're going to get a bunch of tech with similar potential in the next handful of years.
> Phones in your pocket are performing on-device ML and AR in ways previously thought impossible. They’re being used to shoot actual movies that are shown in actual theaters.
The fancy algorithms are somewhat important to picture quality, but for photos slower would be okay and for video you could run it in post without much disruption.
> Low power tech like smartwatches wouldn’t be possible without these breakthroughs because ultimately faster processors also imply low power devices that can actually do decent amounts of computation.
On the other hand, we've had pretty complex devices that run on button cells available for decades. And a lot of that power is taken up by the screen and radios rather than processing. And... wait a second, the original Apple Watch came out so long ago, it was around 8x slower than the current model.