Essay by Eric Worrall
“The real existential threat right now is not a degree of climate change. It’s the fact that we could lose the arms race if we don’t have enough power.”
Fire up the gas turbines, says US Interior Secretary: We gotta win the AI arms race
Climate change? No worry – we can solve that later, argues Doug Burgum
Brandon Vigliarolo
Fri 12 Sep 2025 // 20:59 UTCYou would think that the government official responsible for safeguarding the US’ natural resources would be opposed to abandoning climate change mitigation pledges in favor of firing up fossil fuels to power AI development.
This is the Trump administration, however, so you’d be wrong.
Speaking at a natural gas industry event in Italy this week, US Interior Secretary and former software exec Doug Burgum said that the true existential threat facing the world is America losing the AI arms race. Sure, climate change and its effect on future generations worry him, but “that’s all solvable,” Burgum said.
“Yes, I’m worried about the future,” said Burgum, who sold the accounting software company Great Plains to Microsoft for $1.1 billion in 2001, well before starting his political career. As he put it, “The real existential threat right now is not a degree of climate change. It’s the fact that we could lose the arms race if we don’t have enough power.”
…
Read more: https://www.theregister.com/2025/09/12/fire_up_gas_turbines_ai_race/
A video of the speech is available here.
The race for the ultimate AI is the race for cultural supremacy. The winner of the AI race will have the power to outsmart opponents on every level, they will know every weakness of their opponents and exactly how they should exploit that weakness to achieve geopolitical goals.
If the USA wins the race, US cultural values will dominate the world for the foreseeable future, and other societal systems like Chinese communism and European and Canadian socialism will be overthrown. If another power wins the AI race, their cultural values will dominate the world – it will be the USA and US constitution which will be overthrown, as irresistibly well prepared agents of influence sow confusion and tear away at the foundations of the Republic, until the structure collapses and the nation reforms around new value systems.
The cold war arms race is back on, except this time everyone is building ever more powerful AIs, and the power infrastructure to keep them running. What is left of the “green revolution” will be trampled in the rush to build real energy assets.
AI:
According to Nvidia CEO Jensen Huang, the next generation of AI will require 100 to 1,000 times more computing power than current systems. This means a significant increase in the power needed to run the massive AI data centers using his company’s chips.
NVDA leads the pack in terms of chips. If 100 times more power is needed, how can oil and gas supply that?
Also AI:
Building Small Modular Reactors (SMRs) for AI data centers can take anywhere from a few years to more than a decade, with commercial availability not expected until the 2030s. The total timeline depends on various factors, including regulatory approval, site preparation, construction, and licensing.
“commercial availability not expected until the 2030s”
I heard about SMRs 1n the 1990s, when they also weren’t new. Starts to sound like the market isn’t asking.
Sounds a bit like of wasted time and energy to me, despite the fact it will bring us back to “normal” and free us from the “green” intermittent BS.
Time will tell if energy prices will truly fall, higher demand normally induces the contrary and the question rises: “what good will we receive in exchange for an AI arms race?” Looking back to the past cold war arms race I’m a bit sceptical (yes despite thank you internet and GPS – analogue sarcast has spoken)
I haven’t caught the AI bug. I understand that knowledge is power but how powerful is bad knowledge? One nation or group of nations controlling the rest of us because they spent the most on AI seems overboard. This fella talks like there is only a limited amount of AI and the first to develope it are the winners where as the rest of us are losers. I don’t believe that. It makes me think of persons or groups that have written a song or music and went nowhere with it. Someone else comes along and tweaks their work and has a number one hit. Knowledge is one thing, what you do with that knowledge is what matters.
AI is the new climate scam.
30 years ago, at the generic USA state school I went to, most of the grad students were from non USA places. American kids were doing undergrad in business and social sciences. A quick Internet stats check shows most engineering grad students today are from “somewhere else”. I told my own kids don’t do engineering because the jobs mostly won’t be in the USA. IF AI is the new space race, USA gave the competition a head start again.
I thought it was the first to get a nuclear bomb that was going to rule the world?
No, no, no, its the first to build out the internet.
Well not the internet per se but dot com, put your money in my hand.
It all got destroyed by Y2K and we had to start civilization over again.
Then we promptly all died of Covid.
Don’t get me wrong, AI will change a lot of things. But the hype is getting retarded.
Well, if its the excuse to kill the Green Monster that works, I can l live with it.
My assessment of artificial intelligence to this point is that it’s garbage in, garbage out on steroids.
Agreed. AI can only be trained on what is known, not “everything”. Therefor it does not “know” anything about potential / possible future discoveries.
Yes, and who chooses what material is used in the training? “Quis custodiet ipsos custodes“
Artificial Intelligence has been known to tell lies/untruths.
I hope AI doesn’t lie to our military.
Or our banks.
I think Willis Eschenbach’s recent post where he outlined the rules he set for his AI search shows how it can be made to help but only if you set very strict rules.
True. Just consider where it gets the data.
This sounds exactly like the sort of thing someone who recently bought NVidia shares would say.
Meanwhile, back in the real world, we have people losing their minds and self-deleting because AI chatbots are fuelling their delusions, entry-level jobs for young adults drying up faster than an actress in rehab, and lawyers that are using AI to do their grunt work are finding out first hand that hallucinations cost money.
AI is a poisoned chalice and no mistake.
Your first example is AI “sowing confusion”, your second example is AI boosting the economy, albeit at the expense of people whose economic usefulness has been diminished, and your third example is a defect which will be corrected.
“which will be corrected.”
Oh yea? It will only be corrected when new discoveries are made, and NOT by AI.
AI will get better at grunt work. There are already AI companies that specialize as assistants for lawyers and because they are purpose built, those kinds of mistakes don’t happen.
Don’t for one second believe that grunt work won’t be done by AI because of some famous mistakes. Would you fire a new employee for making a mistake? AI is already getting pretty good at grunt work, it will only get better.
Depends on the mistake.
a defect which will be corrected?
Consider the data sources. AI is a court of public opinion based on the lowest evidentiary standard – preponderance of the evidence.
Until peer review and internet biases are cleaned up, AI will be destructive.
Right, Eric, who will do the correcting? Another AI?
Is that a euphemism for AI-induced psychosis?
Surely you’re not serious? People without jobs don’t spend much money. What’s the point in having AI be your call centre if no-one has the money to buy your products and complain about them? You’ve no doubt heard about that MIT report that says 95% of business are getting zero ROI on their AI investments, so aside from the companies selling AI, I don’t think there’s much of a boost going on.
As for the article, not sure that anything in that is going to a) make a difference or b) work. Chain of reasoning may as well be called chain of hallucination. LLM’s have zero reasoning ability. They just fool you into thinking that they do. And putting a human in the loop defeats the whole point of AI, surely?
Finally, cherry on the top, dead internet theory. AI will generate more garbage more quickly than we mere mortals, and having AI pick up new info on the fly – more and more likely to have been generated by AI – instead of discrete data sets will only compound the hallucination problem.
The reason why entry level jobs are drying up is mostly due to rising minimum wage.
I would agree with you that AI is rather overhyped but the rush to deliver it requires lots of power and that is incompatible with Net Zero so is to pose a dilemma. It could help collapse the Net Zero fantasy.
It does not bode well for future employment for those starting at the bottom. Your mention of the law industry using it instead of employing trainees will eventually result in a dearth of trained people. I suspect most people encounter AI via a chatbot which is a front companies put up to stop you bothering them with queries. I have found them to be utterly useless and inclined to take my business elsewhere.
People have been worrying about technology/automation taking away jobs for well over 300 years. (Look up where the word sabotage comes from.)
Old jobs go away, new jobs are created. Increased efficiency allows people to earn more while working less. (200 years ago, most people worked dawn to dusk 6 days a week.)
Will things change? Yes.
Will society end? No
Look, the key is energy – affordable, plentiful and reliable energy, not pixie dust and unicorn farts. Everything else, including AI will benefit from that. Eyes on the prize. Without it, our enemies, especially China, will eat our lunch.
Burgum is still confused about CO2 and the Earth’s climate. He thinks there is a connection. There is no evidence for this, Doug.
“Sure, climate change and its effect on future generations worry him, but “that’s all solvable,” Burgum said.”
There is no reason to worry about emissions of CO2, and there is no “climate change” problem to solve. There never was a well-supported scientific basis to expect rising pCO2 to drive sensible heat gain down here, or to influence any trend of any climate variable. In short, this is because the minor radiative effect is massively overwhelmed by energy conversion within the general circulation. The modelers know this, as the ERA5 reanalysis demonstrates through its hourly parameter, “vertical integral of energy conversion.” More here about that:
https://drive.google.com/drive/folders/1PDJP3F3rteoP99lR53YKp2fzuaza7Niz?usp=sharing
Our U.S. officials are doing a great job right now of correcting the policy errors of past administrations, but they really need to get serious about correcting the error of the core claim.
We all gotta remember that Artificial Intelligence is ARTIFICIAL!
… and it is not intelligent.
I have wondered when the definition of Artificial Intelligence changed. My, admittedly hazy, memory is that it used to mean machine intelligence with the ability to reason; not a super duper search engine combined with an expert system that can mimic human communications
Being an old fart, I have certain kinds of ads that follow me around on the internet: incontinence, sexual dysfunction and so on. Yesterday, for about a minute, I watched the avatar for a gray-haired actor who was pretending to be a doctor. He explains why none of the conventional cures for arthritis are any good, and promises to share the single, common sense dietary cure. It took me a while, watching his features and eye-movements to realize it was an animation, a product.
Without really knowing, I would guess there will be answers to these nuisances (and serious fraud) provided by the tool itself. But I could be wrong. Query: Does AI police itself?
I don’t know where these answers are coming from, but people with common sense are going to make use of them to their benefit.
My wife has been following the World Athletic Championships in Tokyo. This evening she asked Google for the medals table. The system brought up an AI which informed her that the games were over. No, they go on until the the 21st, or so. So, the AI is really good at quickly giving a completely incorrect answer.
Ask the AI when it’s database was last updated. In the case of Copilot, apparently it accesses info dynamically. It was able to immediately give me a table of the college football games from last weekend.
Here was the explanation it gave me:
I stay current by pulling in the latest information from the web—including sports scores, news headlines, financial data, and more. So whether it’s a coaching change in college football, a diplomatic development in the Baltics, or a tech update relevant to ASP.NET, I can usually give you the most recent and relevant details.
That said, I don’t store a static “database” in the traditional sense. Instead, I access and synthesize fresh information dynamically. So if you ask about something that happened yesterday—or even earlier today—I can typically find it and summarize it for you.
If you ever want to test how current I am, just toss me a topic and I’ll show you what I’ve got.
Well, at least he has his priorities in proper order.
“The race for the ultimate AI is the race for cultural supremacy. The winner of the AI race will have the power to outsmart opponents on every level, they will know every weakness of their opponents and exactly how they should exploit that weakness to achieve geopolitical goals.”
Really! Wow. You are talking about AI? Which is currently not particularly unbiased – but “we’ll do it better next time”.