M&A: xAI, A SpaceX Company

[meekGee]

Your brain and mine require something like 20 to 25W of power. I don't think that the current models of AI capture what's really going on (by many orders-of-magnitude, even correcting for the magnitude of analog-to-digital conversion). I don't know the correct algorithm, but I do know that this isn't it. The current algorithms *might* be equivalent to the retina, but that's about it.

Nobody is claiming it is...

But does it have to work the same way?

[xvel]

Your brain and mine require something like 20 to 25W of power. I don't think that the current models of AI capture what's really going on (by many orders-of-magnitude, even correcting for the magnitude of analog-to-digital conversion). I don't know the correct algorithm, but I do know that this isn't it. The current algorithms *might* be equivalent to the retina, but that's about it.

A small LLM speaks dozens of languages fluently and remembers the entire Wikipedia. It has superhuman abilities in some areas and is lacking in others. It is not a brain, it is something else, but it is not a simple system, as you are trying to imly.
Your digression on analog-to-digital conversion also suggests that you have not much idea about either of the two things as deep learning models are fundamentally continuous analog systems that are approximated by digital computers and spiking neurons of biological brains are actually very digital in nature.

[Vultur]

I see no reason to expect exponential growth in energy use over the next century or so. That's been driven so far by exponential growth in world population + more and more of the world population being brought into one connected economy. The first is slowing drastically and the second is self-limiting (there are no more markets as big as China and India).

Growth may effectively end. It isn't a fundamental law that it will/must continue.

[meekGee]

I see no reason to expect exponential growth in energy use over the next century or so. That's been driven so far by exponential growth in world population + more and more of the world population being brought into one connected economy. The first is slowing drastically and the second is self-limiting (there are no more markets as big as China and India).

Growth may effectively end. It isn't a fundamental law that it will/must continue.

If you're talking about the conversation between TM and myself, I don't think either of us took it literally or seriously.  Just juggling nomenclature and OOMs.

Nothing remains exponential across that many OOMs, some other limit always kicks in.

[InterestedEngineer]

I see no reason to expect exponential growth in energy use over the next century or so. That's been driven so far by exponential growth in world population + more and more of the world population being brought into one connected economy. The first is slowing drastically and the second is self-limiting (there are no more markets as big as China and India).

Growth may effectively end. It isn't a fundamental law that it will/must continue.

If you're talking about the conversation between TM and myself, I don't think either of us took it literally or seriously.  Just juggling nomenclature and OOMs.

Nothing remains exponential across that many OOMs, some other limit always kicks in.

exponential curves are an interesting topic.  Yes, they all eventually turn into logistics (S) curves, due to some limitations.  See for example DRAM prices, basically no change for 10 years.  We have hit a physical cost/bit limit.

OTOH, you CAN turn something that has matured (top of the S curve), into a new exponential growth.  And there's no better example than "new space" (SpaceX), regarding launch costs per kilogram.

Did physics change?  No, the same chemical limitations in general apply, we haven't gone with fission or fusion.

So what changed?

1.  Material Science (high volume 3D printing, new alloys)
2.  CFD being able to simulate an entire rocket engine - which was due to exponential reduction in compute price/performance in the last 30 years.  The Methalox engine, it can be argued, isn't truly possible to make reliable without the ability to simulate it.  This is happening in many fields.
3. The ability to reuse rocket hardware, which is a combination of compute power, better engines, sensors, and maturing engineering, and the willingness to try and fail a lot (itself a societal product of  the 1990s software boom)
4.  40+ years of reusable heat shield improvements (slow but steady), reaching critical mass.
5.  Probably stuff I'm missing.

Falcon-9 is about 3x cost reduction, and Starship will be another 10-40x on top of that.  That's definitely in the exponential growth area of the S-curve

So, getting around to the topic of energy, do we have such an exponential growth?  Kinda.  Not in amount of watts produced for sure - we are nearing the top of that s-curve.   But in the ability to produce those watts cheaply in space from solar power?  Absolutely.  It's now less than $1/watt of capacity to install solar power now on Earth, it was a lot more 2 decades ago.  Mix that with the  exponential drop in launch costs, and plenty of sunshine in space, and you have a new opportunity to grow the power produced by the human race.

And rather than beaming it down (which Elon says will never be efficient enough), we can just beam down the small few hundred megabytes of results from compute (AI, whatever), which is micropennies of energy.

So I think if one counts the net power produced by the human race, we will escape our current "top of the S curve is readily apparent" limits and go on another exponential binge.   Doing this in space for compute is good, we aren't taking away electricity from starving babies.

[meekGee]

I see no reason to expect exponential growth in energy use over the next century or so. That's been driven so far by exponential growth in world population + more and more of the world population being brought into one connected economy. The first is slowing drastically and the second is self-limiting (there are no more markets as big as China and India).

Growth may effectively end. It isn't a fundamental law that it will/must continue.

If you're talking about the conversation between TM and myself, I don't think either of us took it literally or seriously.  Just juggling nomenclature and OOMs.

Nothing remains exponential across that many OOMs, some other limit always kicks in.

exponential curves are an interesting topic.  Yes, they all eventually turn into logistics (S) curves, due to some limitations.  See for example DRAM prices, basically no change for 10 years.  We have hit a physical cost/bit limit.

OTOH, you CAN turn something that has matured (top of the S curve), into a new exponential growth.  And there's no better example than "new space" (SpaceX), regarding launch costs per kilogram.

Did physics change?  No, the same chemical limitations in general apply, we haven't gone with fission or fusion.

So what changed?

1.  Material Science (high volume 3D printing, new alloys)
2.  CFD being able to simulate an entire rocket engine - which was due to exponential reduction in compute price/performance in the last 30 years.  The Methalox engine, it can be argued, isn't truly possible to make reliable without the ability to simulate it.  This is happening in many fields.
3. The ability to reuse rocket hardware, which is a combination of compute power, better engines, sensors, and maturing engineering, and the willingness to try and fail a lot (itself a societal product of  the 1990s software boom)
4.  40+ years of reusable heat shield improvements (slow but steady), reaching critical mass.
5.  Probably stuff I'm missing.

Falcon-9 is about 3x cost reduction, and Starship will be another 10-40x on top of that.  That's definitely in the exponential growth area of the S-curve

So, getting around to the topic of energy, do we have such an exponential growth?  Kinda.  Not in amount of watts produced for sure - we are nearing the top of that s-curve.   But in the ability to produce those watts cheaply in space from solar power?  Absolutely.  It's now less than $1/watt of capacity to install solar power now on Earth, it was a lot more 2 decades ago.  Mix that with the  exponential drop in launch costs, and plenty of sunshine in space, and you have a new opportunity to grow the power produced by the human race.

And rather than beaming it down (which Elon says will never be efficient enough), we can just beam down the small few hundred megabytes of results from compute (AI, whatever), which is micropennies of energy.

So I think if one counts the net power produced by the human race, we will escape our current "top of the S curve is readily apparent" limits and go on another exponential binge.   Doing this in space for compute is good, we aren't taking away electricity from starving babies.

That's all well understood and agreed.

The question was whether the use of solar in space could grow to be comparable with the total solar insolation the Earth receives.

Because exponents will get you there, what's 100,000,000x between friends.

My point was, no.  It'll stop long before that, simply because Earth itself as a consumer is finite.

Maybe when we're all living in space it'll become relevant, but then the whole comparison to Earth's energy balance is irrelevant.

[Twark_Main]

The question was whether the use of solar in space could grow to be comparable with the total solar insolation the Earth receives.

Because exponents will get you there, what's 100,000,000x between friends.

My point was, no.  It'll stop long before that, simply because Earth itself as a consumer is finite.

Coal owners in the 1800s comforted themselves with similar hand-waving stories why the effects of CO2 could never change the total energy balance of Earth. Earth was equally finite then too.

When we first invented plastics, we never imagined that microplastic pollution could ever get so massive that it becomes an issue. Yet here we are.

I, for one, welcome our strange new "driving by looking out the front windshield" overlords.   


Also, if you think humans can't effect the total exergy (heat engine) of Earth, do a quick check on the current impact of a little thing called "agriculture" on the total primary energy (*exergy) available for the biosphere.

This isn't some theoretical / unimaginable scope and scale. We're applying major stress to this huge system already, and this would be on top of that.


Note that impact can greatly exceed the watts of power you generate as a human. Cutting down a rain forest and replacing it with solar panels comes with no guarantee that the solar panels use energy as effectively as the previous rain forest.   

[Vultur]

The question was whether the use of solar in space could grow to be comparable with the total solar insolation the Earth receives.

Because exponents will get you there, what's 100,000,000x between friends.

My point was, no.  It'll stop long before that, simply because Earth itself as a consumer is finite.

Coal owners in the 1800s comforted themselves with similar hand-waving stories why the effects of CO2 could never change the total energy balance of Earth. Earth was equally finite then too.

Not really. I don't think the idea was considered at all (even to be dismissed) in the 1800s. Svante Arrhenius write something brief about it on the early 1900s though. (But not as a problem. Not only was the amount of fuel burned far smaller then, the *second order effects* that cause the real problems weren't understood. A 1 or 2 degree rise would have sounded more beneficial than harmful in northern/western Europe.)

But that aside, there's a huge difference. Not only was world population way lower then, only a tiny proportion of it (pretty much Western Europe and the Northeast/Midwest US) was industrialized at all. So the accessible demand, and thus energy use, grew enormously over the course of the 20th century, as world population grew AND a much larger proportion of the world population was brought into the world economy.

That's not really happening any more. Sure not all the world population is participating in the global economy ... But with rural electrification in China and India, there aren't any *equivalently large* untapped markets.

This situation hasn't existed in at least 200 years; there's no real history to go by as to how that will interact with a technological/industrial economy, as the entire industrial age has been one of very rapid demand growth.

[Oersted]

Relevant:

https://twitter.com/elonmusk/status/2000331928109809757

[Twark_Main]

The question was whether the use of solar in space could grow to be comparable with the total solar insolation the Earth receives.

Because exponents will get you there, what's 100,000,000x between friends.

My point was, no.  It'll stop long before that, simply because Earth itself as a consumer is finite.

Coal owners in the 1800s comforted themselves with similar hand-waving stories why the effects of CO2 could never change the total energy balance of Earth. Earth was equally finite then too.

Not really. I don't think the idea was considered at all (even to be dismissed) in the 1800s. Svante Arrhenius write something brief about it on the early 1900s though.

(But not as a problem. Not only was the amount of fuel burned far smaller then, the *second order effects* that cause the real problems weren't understood. A 1 or 2 degree rise would have sounded more beneficial than harmful in northern/western Europe.)

Arrhenius's paper was in 1896.  I chose my dates carefully. 

Not that it would make a lick of difference, of course. "No no, it's totally different today. The difference is that we can see the problem coming, which is why we shouldn't do anything...."   

But that aside, there's a huge difference. Not only was world population way lower then, only a tiny proportion of it (pretty much Western Europe and the Northeast/Midwest US) was industrialized at all. So the accessible demand, and thus energy use, grew enormously over the course of the 20th century, as world population grew AND a much larger proportion of the world population was brought into the world economy.

That's not really happening any more. Sure not all the world population is participating in the global economy ... But with rural electrification in China and India, there aren't any *equivalently large* untapped markets.

I imagine that in 1896 you would've argued that everyone only "needs" an electric clothes iron and a toaster, so there's no way this industrial revolution thing has much steam left in it anyway.   

There doesn't seem to be a ceiling on quality-of-life. Those rural Indian and Chinese people aren't "done." They want all the same creature comforts, including those from AI, and they should have access to it. People usually get (rightfully) upset if you suggest otherwise.

This situation hasn't existed in at least 200 years; there's no real history to go by as to how that will interact with a technological/industrial economy, as the entire industrial age has been one of very rapid demand growth.

"And that's why I'm so confident in my prediction." 

But seriously, I agree this is an uncertain moment.  What we set into motion now will shape things for decades and even centuries to come. That's why I'm not unhappy to see a little extra "tread carefully" mindset among the people shaping that future.



 I hope we can eventually circle back to Reason #1. I drew the distinction because all the "real" economics is in there.

[InterestedEngineer]

I"m confused.  How does using googlewatts of power in space negatively affect the Earth at all?

The fact that we don't have to generate the power on Earth is a positive effect.  What's the negative?  (leaving aside AI armageddon predictions, let's stick with energy)

[Vultur]

Arrhenius's paper was in 1896.  I chose my dates carefully. 

Hmm, ok. Last time I looked i thought that was just the greenhouse effect in general, the connection to coal burning wasn't until a later publication. But I could well be wrong.

I imagine that in 1896 you would've argued that everyone only "needs" an electric clothes iron and a toaster, so there's no way this industrial revolution thing has much steam left in it anyway.   

There doesn't seem to be a ceiling on quality-of-life. Those rural Indian and Chinese people aren't "done." They want all the same creature comforts, including those from AI, and they should have access to it. People usually get (rightfully) upset if you suggest otherwise.

But the connection between energy use and quality of life isn't linear. In many cases, new technologies are more energy efficient - an electric car's motor might be 90%+ efficient at converting stored energy to motion whereas a gas car's engine is more like 30%. LED bulbs are more energy efficient than fluorescents which are more energy efficient than incandescents. Etc.

And I don't think current AI improves quality of life. IMO it is a net *detriment*. Google, and the Internet in general, has distinctly become less useful and reliable in the last four years.

And that's why demand growth or the lack thereof matters. So far, jobs displaced by technological advancement have been replaced by more jobs elsewhere in the economy. But in a low demand growth world, that may no longer be true - and Ai/automation may no longer increase quality of life, rather decrease it.

"And that's why I'm so confident in my prediction." 

I'm *not*. What I'm questioning is the apparent assumption that rapid growth *will* continue and the question is who wins out/makes money from it.

What I don't see a lot of talk about is the scenario where that ... Just doesn't happen. There's no "next industrial revolution", but also no catastrophic collapse.

[Vultur]

I"m confused.  How does using googlewatts of power in space negatively affect the Earth at all?

The fact that we don't have to generate the power on Earth is a positive effect.  What's the negative?  (leaving aside AI armageddon predictions, let's stick with energy)

It doesn't, other than jobs being displaced (which is possibly a larger concern in an overall low growth world than it would have been a couple decades ago).

My concern is about the investment in space AI being wasted if it turns out that the current huge AI demand is basically because it's free to most of the end users (or bundled as Copilot with Microsoft Office, etc) and once the real costs start being charged there turns out to be actually relatively little demand for AI.

[meekGee]

The question was whether the use of solar in space could grow to be comparable with the total solar insolation the Earth receives.

Because exponents will get you there, what's 100,000,000x between friends.

My point was, no.  It'll stop long before that, simply because Earth itself as a consumer is finite.

Coal owners in the 1800s comforted themselves with similar hand-waving stories why the effects of CO2 could never change the total energy balance of Earth. Earth was equally finite then too.

When we first invented plastics, we never imagined that microplastic pollution could ever get so massive that it becomes an issue. Yet here we are.

I, for one, welcome our strange new "driving by looking out the front windshield" overlords.   


Also, if you think humans can't effect the total exergy (heat engine) of Earth, do a quick check on the current impact of a little thing called "agriculture" on the total primary energy (*exergy) available for the biosphere.

This isn't some theoretical / unimaginable scope and scale. We're applying major stress to this huge system already, and this would be on top of that.


Note that impact can greatly exceed the watts of power you generate as a human. Cutting down a rain forest and replacing it with solar panels comes with no guarantee that the solar panels use energy as effectively as the previous rain forest.   

Yup but those are qualitative hand-wavy analogies.  They don't hold up to the 100,000,000x mismatch.

I thought world power consumption since the 1700s would be 100,000x.  Turns out it's "only" less than 100x.

Broadly, even though our great great grandfathers didn't think emissions would be the problem that we later figured it is, it does not follow that solar panels on Earth risk directly affecting Earth's energy balance, even if we don't think so today...  It's a word analogy, but not a logical argument.

Not even in 200 years.

[Twark_Main]

The question was whether the use of solar in space could grow to be comparable with the total solar insolation the Earth receives.

Because exponents will get you there, what's 100,000,000x between friends.

My point was, no.  It'll stop long before that, simply because Earth itself as a consumer is finite.

Coal owners in the 1800s comforted themselves with similar hand-waving stories why the effects of CO2 could never change the total energy balance of Earth. Earth was equally finite then too.

When we first invented plastics, we never imagined that microplastic pollution could ever get so massive that it becomes an issue. Yet here we are.

I, for one, welcome our strange new "driving by looking out the front windshield" overlords.   


Also, if you think humans can't effect the total exergy (heat engine) of Earth, do a quick check on the current impact of a little thing called "agriculture" on the total primary energy (*exergy) available for the biosphere.

This isn't some theoretical / unimaginable scope and scale. We're applying major stress to this huge system already, and this would be on top of that.


Note that impact can greatly exceed the watts of power you generate as a human. Cutting down a rain forest and replacing it with solar panels comes with no guarantee that the solar panels use energy as effectively as the previous rain forest.   

Yup but those are qualitative hand-wavy analogies.  They don't hold up to the 100,000,000x mismatch.

I thought world power consumption since the 1700s would be 100,000x.  Turns out it's "only" less than 100x.

Broadly, even though our great great grandfathers didn't think emissions would be the problem that we later figured it is, it does not follow that solar panels on Earth risk directly affecting Earth's energy balance, even if we don't think so today...  It's a word analogy, but not a logical argument.

Not even in 200 years.

The bold says you're still thinking linearly, not exponentially.

100x in 200 years? That's quite fast! That means 100,000x only takes 500 years. That's a moderately recent pub in certain parts of the world. 

[thespacecow]

Elon replied "Yes" to this post: https://x.com/aaronburnett/status/2000970949856837991

Phil is correct. The space data center arguments generally are talking past each other. If you adjust the time frame or demand scale you can make the numbers support your perspective, whatever it is. Anyone that’s being intellectually honest isn’t saying the next 10 GWs is cheaper per GW in space than on earth, they are saying at scale the only way to build 50-100GWs per year is in space and at that scale is when it becomes cheaper.

I’ve said this many times and will continue to say it. This all comes back to demand for intelligence. If you can’t imagine needing more than 100x the intelligence we have today then you will not understand space data centers. If you can see use for 1000, 10000, or 1000000 times the intelligence then SDCs/ODCs are obvious.

[thespacecow]

Related: https://x.com/StockSavvyShay/status/2001391832241181103

Elon Musk told xAI staff that winning the AI race comes down to surviving the next 2–3 years & scaling compute, power & capital faster than anyone else.

Key points from the meeting:
• xAI could achieve AGI as early as 2026
• xAI has access to ~$30B per year in funding
• Grok 5 carries ~10% probability of achieving AGI
• Optimus humanoid robots could eventually operate data centers
• Plans to scale from ~200k GPUs to ~1M GPUs across its Colossus data centers
• Long-term concepts include data centers in space & Mars-related infrastructure

Edit: I think this tweet is based on this BI article: https://www.businessinsider.com/xai-all-hands-agi-superintelligence-funding-success-optimus-space

[meekGee]

The question was whether the use of solar in space could grow to be comparable with the total solar insolation the Earth receives.

Because exponents will get you there, what's 100,000,000x between friends.

My point was, no.  It'll stop long before that, simply because Earth itself as a consumer is finite.

Coal owners in the 1800s comforted themselves with similar hand-waving stories why the effects of CO2 could never change the total energy balance of Earth. Earth was equally finite then too.

When we first invented plastics, we never imagined that microplastic pollution could ever get so massive that it becomes an issue. Yet here we are.

I, for one, welcome our strange new "driving by looking out the front windshield" overlords.   


Also, if you think humans can't effect the total exergy (heat engine) of Earth, do a quick check on the current impact of a little thing called "agriculture" on the total primary energy (*exergy) available for the biosphere.

This isn't some theoretical / unimaginable scope and scale. We're applying major stress to this huge system already, and this would be on top of that.


Note that impact can greatly exceed the watts of power you generate as a human. Cutting down a rain forest and replacing it with solar panels comes with no guarantee that the solar panels use energy as effectively as the previous rain forest.   

Yup but those are qualitative hand-wavy analogies.  They don't hold up to the 100,000,000x mismatch.

I thought world power consumption since the 1700s would be 100,000x.  Turns out it's "only" less than 100x.

Broadly, even though our great great grandfathers didn't think emissions would be the problem that we later figured it is, it does not follow that solar panels on Earth risk directly affecting Earth's energy balance, even if we don't think so today...  It's a word analogy, but not a logical argument.

Not even in 200 years.

The bold says you're still thinking linearly, not exponentially.

100x in 200 years? That's quite fast! That means 100,000x only takes 500 years. That's a moderately recent pub in certain parts of the world. 

No it does not, since exponential growth in reality doesn't persist like that. Something always kicks in.  Otherwise, each species of bacteria would have outweighed the planet already.

In our case, power consumption on earth can't increase again and again by the factor of 100x (30x actually but rounding up) that it did during the industrial revolution.

It might do so as humans go into space in the next centuries, but that wasn't your premise - you were talking about affection the energy balance of the earth.

It's just that not every extrapolation is true, you neednto always look wt underlying limitations, resource constraints, etc.

[Twark_Main]

The question was whether the use of solar in space could grow to be comparable with the total solar insolation the Earth receives.

Because exponents will get you there, what's 100,000,000x between friends.

My point was, no.  It'll stop long before that, simply because Earth itself as a consumer is finite.

Coal owners in the 1800s comforted themselves with similar hand-waving stories why the effects of CO2 could never change the total energy balance of Earth. Earth was equally finite then too.

When we first invented plastics, we never imagined that microplastic pollution could ever get so massive that it becomes an issue. Yet here we are.

I, for one, welcome our strange new "driving by looking out the front windshield" overlords.   


Also, if you think humans can't effect the total exergy (heat engine) of Earth, do a quick check on the current impact of a little thing called "agriculture" on the total primary energy (*exergy) available for the biosphere.

This isn't some theoretical / unimaginable scope and scale. We're applying major stress to this huge system already, and this would be on top of that.


Note that impact can greatly exceed the watts of power you generate as a human. Cutting down a rain forest and replacing it with solar panels comes with no guarantee that the solar panels use energy as effectively as the previous rain forest.   

Yup but those are qualitative hand-wavy analogies.  They don't hold up to the 100,000,000x mismatch.

I thought world power consumption since the 1700s would be 100,000x.  Turns out it's "only" less than 100x.

Broadly, even though our great great grandfathers didn't think emissions would be the problem that we later figured it is, it does not follow that solar panels on Earth risk directly affecting Earth's energy balance, even if we don't think so today...  It's a word analogy, but not a logical argument.

Not even in 200 years.

The bold says you're still thinking linearly, not exponentially.

100x in 200 years? That's quite fast! That means 100,000x only takes 500 years. That's a moderately recent pub in certain parts of the world. 

No it does not, since exponential growth in reality doesn't persist like that. Something always kicks in.  Otherwise, each species of bacteria would have outweighed the planet already.

In our case, power consumption on earth can't increase again and again by the factor of 100x (30x actually but rounding up) that it did during the industrial revolution.

It might do so as humans go into space in the next centuries, but that wasn't your premise - you were talking about affection the energy balance of the earth.

It's just that not every extrapolation is true, you neednto always look wt underlying limitations, resource constraints, etc.

Again, in earlier times you could've easily argued that "energy use can't grow further" because there isn't enough wood that grows in all the world's forests. We all know why that prediction was wrong (coal).

Similarly, saying "power consumption on earth can't increase" ignores the fact that we can use power in space to serve needs on Earth (which, reminder, is the subject of this thread).


How many times have people predicted the "end of growth?"  How many times have they been wrong, so far?   


The favorable economics of computing in space are independent of any particular AI or compute technology. This isn't like predicting "everyone will own a DeSoto automobile," this is like predicting "wheeled vehicles will be the dominant form of medium-distance transport."

[meekGee]

The question was whether the use of solar in space could grow to be comparable with the total solar insolation the Earth receives.

Because exponents will get you there, what's 100,000,000x between friends.

My point was, no.  It'll stop long before that, simply because Earth itself as a consumer is finite.

Coal owners in the 1800s comforted themselves with similar hand-waving stories why the effects of CO2 could never change the total energy balance of Earth. Earth was equally finite then too.

When we first invented plastics, we never imagined that microplastic pollution could ever get so massive that it becomes an issue. Yet here we are.

I, for one, welcome our strange new "driving by looking out the front windshield" overlords.   


Also, if you think humans can't effect the total exergy (heat engine) of Earth, do a quick check on the current impact of a little thing called "agriculture" on the total primary energy (*exergy) available for the biosphere.

This isn't some theoretical / unimaginable scope and scale. We're applying major stress to this huge system already, and this would be on top of that.


Note that impact can greatly exceed the watts of power you generate as a human. Cutting down a rain forest and replacing it with solar panels comes with no guarantee that the solar panels use energy as effectively as the previous rain forest.   

Yup but those are qualitative hand-wavy analogies.  They don't hold up to the 100,000,000x mismatch.

I thought world power consumption since the 1700s would be 100,000x.  Turns out it's "only" less than 100x.

Broadly, even though our great great grandfathers didn't think emissions would be the problem that we later figured it is, it does not follow that solar panels on Earth risk directly affecting Earth's energy balance, even if we don't think so today...  It's a word analogy, but not a logical argument.

Not even in 200 years.

The bold says you're still thinking linearly, not exponentially.

100x in 200 years? That's quite fast! That means 100,000x only takes 500 years. That's a moderately recent pub in certain parts of the world. 

No it does not, since exponential growth in reality doesn't persist like that. Something always kicks in.  Otherwise, each species of bacteria would have outweighed the planet already.

In our case, power consumption on earth can't increase again and again by the factor of 100x (30x actually but rounding up) that it did during the industrial revolution.

It might do so as humans go into space in the next centuries, but that wasn't your premise - you were talking about affection the energy balance of the earth.

It's just that not every extrapolation is true, you neednto always look wt underlying limitations, resource constraints, etc.

Again, in earlier times you could've easily argued that "energy use can't grow further" because there isn't enough wood that grows in all the world's forests. We all know why that prediction was wrong (coal).

Similarly, saying "power consumption on earth can't increase" ignores the fact that we can use power in space to serve needs on Earth (which, reminder, is the subject of this thread).


How many times have people predicted the "end of growth?"  How many times have they been wrong, so far?   


The favorable economics of computing in space are independent of any particular AI or compute technology. This isn't like predicting "everyone will own a DeSoto automobile," this is like predicting "wheeled vehicles will be the dominant form of medium-distance transport."

You're strawmaning.

I didn't say energy use can't grow.  I said it can't keep growing exponentially, to a factor of 100,000,000.

So the rest of the post is just poking holes in something I never said.

And my comments are all about the statement that we need to take PV to orbit because sooner or later PV use on Earth will affect the planet's energy balance.

Which btw you made with some qualifiers that made me think you're not really invested in that argument, so imagine my surprise...