12 Comments

“a simple way to do an analysis is to ask MacBook Air users how many PowerBook 100s they would need to give up their one Air.”

I really like that line of reasoning. I think too many people think that our consumer products have not gotten that much better over the past few decades. Your question forces them to acknowledge the obvious.

Expand full comment

When I saw the release of the new iMac last week, I thought it was crazy the base model today is the SAME price ($1,299.00) as the base model in 1999. Superabundance continues!

Expand full comment

I remember my father buying a new Mac for $3000 in the 1980s. I just bought another for a bit less. And that is not even taking into account inflation.

Expand full comment

The comparison is even better if you don't buy Apple.

My personal historical comparisons: In 1981, I bought a Sinclair ZX-81 which came with no hard drive storage and an entire KB of RAM. That's kilo, not mega. I don't remember the cost but it wasn't useful for much unless you wanted to learn BASIC programmed and paid to upgrade to an amazing 12K of RAM. In 1987/8, on starting graduate work in the USA, I bought an IBM PC with a stunning 20 Mb of storage, and a slow dial-up modem.

What will our devices be like in 2050?

Expand full comment

If we stay on Moore's law and double every 18 months, they will be 262,144 time more powerful. 2^18. AI should make Moore's law obsolete. Unless we allow tyranny to prevail.

Expand full comment

Most of the projections of AGI -- most famously those by Ray Kurzweil -- rely on Moore's Law for AI to arise. They don't say that AI will make Moore's Law obsolete. If I recall correctly, not even Hans Moravec expected that. But you could be right. It's hard to say when we are truly approaching a physical limit. So far, we've always gone deeper or worked around physical limits. I just want to still be alive to see how it works out.

Expand full comment

Remember that economics in not physics. Physics is about atoms. Economics is about knowledge. Physics is about the weight of things. Economics is about the value of things. Atoms without knowledge have no economic value. It’s only when we add knowledge to atoms that they can become resources. Since the amount of undiscovered knowledge is infinite, our resources can be infinite. Economics must move beyond the old thinking of atoms and money and scarcity to the new economics of knowledge, time, and abundance. Thanks for reading.

Expand full comment

Yes, I understand that. I've understood that at least since reading Julian Simon's Ultimate Resource in 1984. He goes into the idea in more detail in Ultimate Resource 2. That does not mean that Moore's Law must continue. That depends both on physical laws and on economics. Even if resources as a whole can be infinite (a question I will write about in a future blog entry), that doesn't mean that microprocessors can be miniaturized indefinitely.

Of course, we cannot say for sure that miniaturization must reach a limit. Our physics clearly is not complete and final. New discoveries in physics could lead to a pushing back of technological limits. Hans Moravec (years before my friend Ray Kurzweil) looked at exotic possibilities for pushing the limits of computing (reversible computing, neutronium computing, etc.). But there MAY be a physical limit. Human ingenuity means we can work around such limits and continue to increase value. My point was about physical limits, not economic limits.

Expand full comment

Yes we owe Julian Simon much for explaining how resources are created and then offering evidence for his model. Some have suggested that Superabundance is Ultimate Resource 3. We tried to make Professor Simon proud.

Expand full comment

“New discoveries in physics could lead to a pushing back of technological limits.”

I would say AI shows us that new advances in coding allow us to use much more of the hardware computing power. AI leverages our ability to discover valuable new knowledge faster. It’s not the raw power of chips that count. It’s our ability to use this power to discover knowledge that counts. We’ve moved from the speed of horses to the speed of electrons on a telegraph wire. AI will move us from Morse code to photons on fiber optic cables.

Expand full comment

$100. Unskilled workers were earning $4.40 an hour. That would put the time price at 22.73 hours. Still a great deal for learning how to code.

Expand full comment

I would have paid in pounds sterling. I earned just a little in my teens, delivering newspapers, working in restaurant kitchens, etc. It must have seemed like a lot to me then but I can't recall how much I paid. (Or if it was a gift.)

According to a UK inflation calculator, 100 pounds in 1982 is equivalent to 472.70 today. (That's just inflation-adjusted, not time-price adjusted.) Given that, at 17 years old and in school full time, I can't imagine how I could have afforded it. The number I found was 50 pounds for the kit and 70 pounds assembled, which makes it a bit more conceivable.

Expand full comment