4 ms·

I disagree with this definition of "fundamentally different", because nothing in physics stops you from building a computer out of a billion vacuum tubes. That

I disagree with this definition of "fundamentally different", because nothing in physics stops you from building a computer out of a billion vacuum tubes. That's very different than requiring negative energy or something never observed.

somenameforme 2y ago
You're shifting the goal posts there. We're talking about a device that fits in your pocket, powered by a tiny battery. And, by the knowledge of computing at the time. The amount of energy required to power such a device, let alone the space, would have made the concepts plainly impossible, by the limits of what was known at the time.
- Workaccount2 2y ago
  That's all true, but fundamentally compute has worked the same way for decades, maybe even 100 years at this point. The scales are orders of magnitude different, but the fundamentals haven't changed much. I believe that's all they were pointing out.
throwaway63467 2y ago
Interesting question, actually I think building a computer out of vacuum tubes that would rival say an iPhone 15 seems impossible, the waste energy alone would be prohibitive. The Eniac used around 100 W per vacuum tube, for 10^13 tubes that’s 10^15 W, and that’s two orders larger than the global electricity output, and keep in mind you need to deliver that on a small cubic area of your computer and you need to dissipate that heat somehow (a 1 megaton nuke dissipates around 1 MWh of energy, so this computer would have to dissipate the energy equivalent of one billion megaton nukes per hour). Setting off a billion megaton nukes every hour in a small area seems almost as ridiculous to me as converting the entire mass of Saturn into energy. Also signal runtimes and crosstalk would likely make this infeasible as well. So no, you would never be able to scale vacuum tube technology to the current level of semiconductor physics, not even remotely. If you told the Eniac engineers you know how to put 20 billion vacuum tubes in the area the size of a coin and make each of them switch a few billion times per second they would have been quite skeptical I think and relegated you to the realm of science fiction, because the theory of semiconductors and the “exoticly” engineered electron bandgap materials patterned on the scale of a few angstrom you need for them would’ve been largely incomprehensible to them.

somenameforme 2y ago

You're shifting the goal posts there. We're talking about a device that fits in your pocket, powered by a tiny battery. And, by the knowledge of computing at the time. The amount of energy required to power such a device, let alone the space, would have made the concepts plainly impossible, by the limits of what was known at the time.

Workaccount2 2y ago
That's all true, but fundamentally compute has worked the same way for decades, maybe even 100 years at this point. The scales are orders of magnitude different, but the fundamentals haven't changed much. I believe that's all they were pointing out.

throwaway63467 2y ago

Interesting question, actually I think building a computer out of vacuum tubes that would rival say an iPhone 15 seems impossible, the waste energy alone would be prohibitive. The Eniac used around 100 W per vacuum tube, for 10^13 tubes that’s 10^15 W, and that’s two orders larger than the global electricity output, and keep in mind you need to deliver that on a small cubic area of your computer and you need to dissipate that heat somehow (a 1 megaton nuke dissipates around 1 MWh of energy, so this computer would have to dissipate the energy equivalent of one billion megaton nukes per hour). Setting off a billion megaton nukes every hour in a small area seems almost as ridiculous to me as converting the entire mass of Saturn into energy. Also signal runtimes and crosstalk would likely make this infeasible as well. So no, you would never be able to scale vacuum tube technology to the current level of semiconductor physics, not even remotely. If you told the Eniac engineers you know how to put 20 billion vacuum tubes in the area the size of a coin and make each of them switch a few billion times per second they would have been quite skeptical I think and relegated you to the realm of science fiction, because the theory of semiconductors and the “exoticly” engineered electron bandgap materials patterned on the scale of a few angstrom you need for them would’ve been largely incomprehensible to them.