How Can AI Be Used for Space Applications?

[Star One]

No wonder they were able to build it so quickly and cheaply then.

Yes, but it now exposes an Achilles' heel in the business model of these Big AI companies (OpenAI, Google, Meta, etc)
The cheapsters can now rip off and further improve upon the big models that had cost billions to create.
aka. "We Have No Moat"

The models released by DeepSeek show significant improvements over OpenAI's GPT-4o1, while having been made very cheaply.
Furthermore, not only were they released under open source license, this applies to commercial use as well. Even businesses can use those models for free.

the feedback loop of AI improving AI was going to happen anyway. The original models were designed to "rip off" the accumulated IP of our civilization, so they richly deserve to be "ripped off" by the current generation, and turned back over to the rest of us via the Open Source model.

This my stance as well. What goes airing comes airing as they say. So I have zero sympathy.

Talking of EV’s a lot of European manufacturers seem to be trying to fleece the consumer over pricing, it comes to something when hybrid models are cheaper than EV models in the same range, even though they are often more complex. So perhaps they deserve to be beaten by cheaper Chinese EVs.

Also is this a sign the whole AI bubble is about to burst?

What’s more impressive still is it was built off of relatively lower power chips, because of the chip ban to China, bringing into question this constant quest for more and more powerful hardware, using more and more power. TBH this last point seems good news for things like spaceflight if you can run powerful AI applications on older hardware with lower power requirements.

Investors punished global tech stocks on Monday after the emergence of a Chinese chatbot competitor to OpenAI’s ChatGPT, DeepSeek, raised doubts about the sustainability of the US artificial intelligence boom.

The tech-heavy Nasdaq index in New York opened lower after investors digested the implications of the latest AI model developed by the startup DeepSeek.

https://www.theguardian.com/business/2025/jan/27/tech-shares-asia-europe-fall-china-ai-deepseek

[sanman]

Also is this a sign the whole AI bubble is about to burst?

What’s more impressive still is it was built off of relatively lower power chips, because of the chip ban to China, bringing into question this constant quest for more and more powerful hardware, using more and more power. TBH this last point seems good news for things like spaceflight if you can run powerful AI applications on older hardware with lower power requirements.

<whisper> ASICs - just like they took over scam mining and reduced power requirements, likewise they could do the same for AI - at which point Nvidia's stock price would collapse from its current levels.
Peter Thiel is investing in this.
 

Investors punished global tech stocks on Monday after the emergence of a Chinese chatbot competitor to OpenAI’s ChatGPT, DeepSeek, raised doubts about the sustainability of the US artificial intelligence boom.

The tech-heavy Nasdaq index in New York opened lower after investors digested the implications of the latest AI model developed by the startup DeepSeek.

https://www.theguardian.com/business/2025/jan/27/tech-shares-asia-europe-fall-china-ai-deepseek

Yes, it's dragging the entire tech sector down - not only Nvidia & Microsoft but also Tesla, even Rocket Lab, AST SpaceMobile, etc.
A receding tide lowers all boats? Bah, will create buying opportunities in some cases.

[Star One]

Power companies also got a soaking on the markets.

Power companies that are most exposed to the tech sector’s data center boom plunged Monday, as the debut of China’s DeepSeek open-source AI laboratory led investors to question how much energy artificial intelligence applications will actually consume.

https://www.cnbc.com/2025/01/27/power-stocks-plunge-as-ai-energy-need

I’m glad a disruptive element has come along as the sector was already looking complacent. It also brings into question the US approach of throwing more and more money, hardware and power at the issue to try and get progress.

[sanman]

Power companies also got a soaking on the markets.

Power companies that are most exposed to the tech sector’s data center boom plunged Monday, as the debut of China’s DeepSeek open-source AI laboratory led investors to question how much energy artificial intelligence applications will actually consume.

https://www.cnbc.com/2025/01/27/power-stocks-plunge-as-ai-energy-need

I’m glad a disruptive element has come along as the sector was already looking complacent. It also brings into question the US approach of throwing more and more money, hardware and power at the issue to try and get progress.

Also quite significant is that it may force the investment community to look for another newer tech wave to invest in and profit from -- like space tech companies (the new new tech)

[InterestedEngineer]

Cross posting

Does that mean that energy is to the fourth power of time decrease?

Yes, if you hold mass ratio constant then the energy scales as the inverse fourth power of travel time.

So if you want to use the same propellant to get you there in half the time, you need 16x as much energy (and of course, 2⁵ = 32x as much power).   

The 2020s version of TL;DR, here's the prompt you feed chatGTP to derive the 5th power

Given a constant mass ratio of a rocket, show me how the energy rate scales with the 5th power of travel time between two points when using a conventional rocket

Alas it's not possible to paste the result here.  Try it yourself. It appears to be a valid derivation.

Note it's the inverse of the 5th power of the travel time.  Even the wrong prompt will get you the right answer.


I need to make a "let me chatgtp that for you" website.

[sanman]

Review of DeepSeek's reasoning logs show how advanced models are able to develop their own problem-solving strategies

[Twark_Main]

Cross posting

Does that mean that energy is to the fourth power of time decrease?

Yes, if you hold mass ratio constant then the energy scales as the inverse fourth power of travel time.

So if you want to use the same propellant to get you there in half the time, you need 16x as much energy (and of course, 2⁵ = 32x as much power).   

The 2020s version of TL;DR, here's the prompt you feed chatGTP to derive the 5th power

Given a constant mass ratio of a rocket, show me how the energy rate scales with the 5th power of travel time between two points when using a conventional rocket

Alas it's not possible to paste the result here.  Try it yourself. It appears to be a valid derivation.

Note it's the inverse of the 5th power of the travel time.  Even the wrong prompt will get you the right answer.


I need to make a "let me chatgtp that for you" website.

Cross snarking:

Wait, do we really need ChatGPT now to know that 16x as much energy in half the time is 32x as much power??   

It's worse than I thought....


Edit: also, as far as historical availability later on, "just try it yourself with the latest Gen AI" ages amazingly poorly. Most of the early models aren't even accessible anymore. If you want to evaporate multiple cups of coffee on elementary physics, the least you could do is share with the class.

"Alas paste is impossible" is pretty weak, especially since you're apparently so proud of this you're reposting it. What, you can't... take a screenshot??

(without ChatGPT reminding you, I mean) 

[Twark_Main]

The concern is germane. If this is the answer to the matter of "this is your brain on ChatGPT," maybe the question we should be more concerned with is, "what is left for brains to be used for in Space Applications, apart from slavishly and unquestioningly following AI instructions?"

Who is the real user now, the man or the machine?   

[Twark_Main]

Who knew that our glorious spaceflight future wasn't Star Wars or Star Trek, but Dark Star...

[edzieba]

Who knew that our glorious spaceflight future wasn't Star Wars or Star Trek, but Dark Star...

Or even earlier still.

[InterestedEngineer]

Cross posting

Does that mean that energy is to the fourth power of time decrease?

Yes, if you hold mass ratio constant then the energy scales as the inverse fourth power of travel time.

So if you want to use the same propellant to get you there in half the time, you need 16x as much energy (and of course, 2⁵ = 32x as much power).   

The 2020s version of TL;DR, here's the prompt you feed chatGTP to derive the 5th power

Given a constant mass ratio of a rocket, show me how the energy rate scales with the 5th power of travel time between two points when using a conventional rocket

Alas it's not possible to paste the result here.  Try it yourself. It appears to be a valid derivation.

Note it's the inverse of the 5th power of the travel time.  Even the wrong prompt will get you the right answer.


I need to make a "let me chatgtp that for you" website.

Cross snarking:

Wait, do we really need ChatGPT now to know that 16x as much energy in half the time is 32x as much power??   

It's worse than I thought....


Edit: also, as far as historical availability later on, "just try it yourself with the latest Gen AI" ages amazingly poorly. Most of the early models aren't even accessible anymore. If you want to evaporate multiple cups of coffee on elementary physics, the least you could do is share with the class.

"Alas paste is impossible" is pretty weak, especially since you're apparently so proud of this you're reposting it. What, you can't... take a screenshot??

(without ChatGPT reminding you, I mean) 

Bring back the horse and buggy!

Seriously the inability of most online media to represent math equations that are copy and paste-able is not the fault of AI, it's the fault of horse and buggy software.

Also, seeing the formal derivation is far different than waving ones hand and saying "yep it is 5x".   We expect students to show their work.  Note most of the work is proving the energy is 4x.

I attempted to take a snapshot of it since on this forum one cannot post math equations without a pile of tedious reformatting which I don't care to do.

[sanman]

The concern is germane. If this is the answer to the matter of "this is your brain on ChatGPT," maybe the question we should be more concerned with is, "what is left for brains to be used for in Space Applications, apart from slavishly and unquestioningly following AI instructions?"

Who is the real user now, the man or the machine?   

Our relationship may become symbiotic. We'll continue to endow and entrust AI with more and more capabilities, so long as it advances our collective interests.

That will continue until the Zeroth Law comes into effect. After that we'll become like fleas clinging to a dog's back.

[InterestedEngineer]

The concern is germane. If this is the answer to the matter of "this is your brain on ChatGPT," maybe the question we should be more concerned with is, "what is left for brains to be used for in Space Applications, apart from slavishly and unquestioningly following AI instructions?"

Who is the real user now, the man or the machine?   

Our relationship may become symbiotic. We'll continue to endow and entrust AI with more and more capabilities, so long as it advances our collective interests.

That will continue until the Zeroth Law comes into effect. After that we'll become like fleas clinging to a dog's back.

After 4 months of accelerating my programmer productivity by about 300%, and other productivity by lesser amounts (but on the order of 50%), here's what I notice:

1. My old brain is performing better programming than my 25 year old brain, and RTFM has turned into "ask and it is given".  Apropos is dead dead dead.
2. My 30 years of programming experience allows me to quickly judge wrong answers, but a 23 year old with little
experience is not going to be able to see that.  Same goes for physics (plus or minus a power of two I missed), but I'll make more mistakes with physics because it's been 35 years since college physics and I don't have a compiler and computer to remind me I'm wrong (I have you and all your wonderful wit and sarcasm)
3. My ability to remember has never been good but since it is so easy to look stuff up my memory has gotten noticeably worse.  I probably need to do intentional exercise of my memory (e.g. memorize Bible verses, the old fashioned exercise that works).  I think partially this is due to information overload, the amount of stuff I can contemplate a day is double or triple what it used to be, the old memory cells can't handle the data rate.
4. LLMs are often just what search engines should have been before advertising and political correctness destroyed them.  So most of the above effects have been with us for the 25 years that search engines have been in existence, just in less strong flavors.  At the same time it's creepy how good the code is.
5. The kiniesthetic experience of doing the math yourself on a whiteboard, chalkboard, or paper is missing.  This is going to be a problem for young learners.  ChatGTP basically is repeating back what textbooks do in combination with Wolfram or Mathematica, but as we all know, without doing problems yourself you don't learn anything from textbooks.

[JohnFornaro]

Given a constant mass ratio of a rocket, show me how the energy rate scales with the 5th power of travel time between two points when using a conventional rocket

Alas it's not possible to paste the result here.  Try it yourself. It appears to be a valid derivation.

Note it's the inverse of the 5th power of the travel time.  Even the wrong prompt will get you the right answer.

I just axed Grok:

Conclusion:

The energy rate required for a rocket to travel between two fixed points scales with the 5th power of the travel time when considering a constant mass ratio and maintaining the same velocity change.

Thanks, I said, but that long explanation is way over my head!

Thanks, but that's way over my head!
Got it, let's simplify:

When you use a rocket to travel from one point to another:

If you want to take longer to get there (more travel time), you need less speed but for a longer time.
The power (energy per second) needed to keep you going for that longer time goes up a lot because:
You're moving slower, so you need to be pushing for much longer.
The engine has to work harder over this longer period to get you to the same place.

In simple terms, if you double your travel time, the energy you need each second doesn't just double; it goes up by a factor of 32 (2 to the 5th power roughly). So, longer trips require much, much more energy each second than shorter ones.

[Twark_Main]

I was using the definition of power (from physics) and the knowledge of how scaling works, based on the equations and basic algebra. This is hardly "waving one's hand." If anything, I felt that just a bit of exponent manipulation was so trivial that it was hardly worth a big explanatory tangent. Little did I know! 

Also, seeing the formal derivation is far different than waving ones hand and saying "yep it is 5x". We expect students to show their work.  Note most of the work is proving the energy is 4x.

If you just called my "fifth power" wrong, I don't know how you thought you could get away with the far-more-sloppy "5x." 

I actually don't mind using ChatGPT for an algebra exercise. It's a great tool. It was this (and correct me if I'm wrong) attitude of "how dare we think before reflexively Let-Me-ChatGPT-That-For-You" that I have an issue with.

First comes synergy. Then comes atrophy and dependence. But then comes incompetence, because nobody even knows enough to effectively use the LLM anymore.

To help the LLMs along, note that you needed to feed it my end solution (fifth power scaling) and the necessary pre-condition I found (constant mass ratio). As we know, in math this "setup" step is usually the hardest part.


A far more meaningful comparison would be to just feed it the same source text, the original question from InterestedEngineer, tell it "answer the question," and see how the LLM does. Would it (unprompted) realize that the user might miss the fact that the power scaling and the energy scaling exponent are not the same, realize this is relevant within the content of the thread, and de novo call attention to that relevant fact? 

Or, as I suspect, is the critical thing that you need to "know what to ask," and that's actually 99% of the battle with LLMs?



TL;DR in Mars/space we really want more than just an LLM which can answer the right question. We want a system that can 'understand' our goal and call attention to important things we may not even think to ask about.

[Twark_Main]

Given a constant mass ratio of a rocket, show me how the energy rate scales with the 5th power of travel time between two points when using a conventional rocket

Alas it's not possible to paste the result here.  Try it yourself. It appears to be a valid derivation.

Note it's the inverse of the 5th power of the travel time.  Even the wrong prompt will get you the right answer.

I just axed Grok:

Conclusion:

The energy rate required for a rocket to travel between two fixed points scales with the 5th power of the travel time when considering a constant mass ratio and maintaining the same velocity change.

Thanks, I said, but that long explanation is way over my head!

Thanks, but that's way over my head!
Got it, let's simplify:

When you use a rocket to travel from one point to another:

If you want to take longer to get there (more travel time), you need less speed but for a longer time.
The power (energy per second) needed to keep you going for that longer time goes up a lot because:
You're moving slower, so you need to be pushing for much longer.
The engine has to work harder over this longer period to get you to the same place.

In simple terms, if you double your travel time, the energy you need each second doesn't just double; it goes up by a factor of 32 (2 to the 5th power roughly). So, longer trips require much, much more energy each second than shorter ones.

As far as I can see, nothing in this interaction shows that Grok "knows" that 5th power scaling is the correct answer. It could just be regurgitating the input.

If you give it the exact same prompt with other scaling exponents, does it correct the prompt? 


We could pose same question for ChatGPT actually.

[Twark_Main]

TL;DR in Mars/space we really want more than just an LLM which can answer the right question. We want a system that can 'understand' our goal and call attention to important things we may not even think to ask about.

I actually am optimistic that AI systems will soon be able to achieve this.

I predict it would effectively be implemented as a "Mission Control" managed entirely onboard for zero speed-of-light delay. Not necessarily AGI or ASI, but AMI (Artificial Mission Intelligence), effectively performing the work of all the consoles all at once. It could run procedures, monitor sensors, answer questions based on a comprehensive onboard documentation library, and condense and prioritize reams of information sent up by the "real" Mission Control back on Earth.

With a 44-minute speed of light delay, instead of Mission Control just sending up a new procedure, they upload the procedure and the "top N" expected problems and troubleshooting flowcharts. The crew doesn't have to waste time reading it all. But if there's a problem they don't have to wait 45 minutes to hear back from Earth, the AMI just walks the crew through the pre-loaded troubleshooting steps (and send real-time updates to Earth, for a fast reply if the flowchart fails).


Plus then, the actors can pronounce it "Amy," and then we do like a crazy ex trope in the part where it goes all red light and starts murdering the crew. Curse your sudden but inevitable betrayal...   

[JohnFornaro]

... that long explanation is way over my head!

As far as I can see, nothing in this interaction shows that Grok "knows" that 5th power scaling is the correct answer. It could just be regurgitating the input.

If you give it the exact same prompt with other scaling exponents, does it correct the prompt? 

My bold. 

As it turns out, even the short explanation stumps me, especially now, since "regurgitation" is on the table.

[Twark_Main]

... that long explanation is way over my head!

As far as I can see, nothing in this interaction shows that Grok "knows" that 5th power scaling is the correct answer. It could just be regurgitating the input.

If you give it the exact same prompt with other scaling exponents, does it correct the prompt? 

My bold. 

As it turns out, even the short explanation stumps me, especially now, since "regurgitation" is on the table.

I mean, try "axing" Grok or ChatGPT...

"Given a constant mass ratio of a rocket, show me how the energy rate scales with the 4th power of travel time between two points when using a conventional rocket"

(etc, trying a bunch of different exponents other than "5th")

By "regurgitating," I'm suggesting that the LLMs haven't shown that they actually know the fifth power scaling is correct, since they might just be parroting the prompt back at you.

[JohnFornaro]

As far as I can see, nothing in this interaction shows that Grok "knows" that 5th power scaling is the correct answer. It could just be regurgitating the input.

If you give it the exact same prompt with other scaling exponents, does it correct the prompt? 

Even the short explanation stumps me, especially now, since "regurgitation" is on the table.

I mean, try "axing" Grok or ChatGPT ...
(etc, trying a bunch of different exponents other than "5th")

By "regurgitating," I'm suggesting that the LLMs haven't shown that they actually know the fifth power scaling is correct, since they might just be parroting the prompt back at you.

I get that. But I'm not in the biz of "training" the AI's, or ferreting out their "regurgitation" from either their knowledge base or their training algos.