本文探讨了人工智能（AI）发展与能源之间的密切关系，特别是石油供应问题。文章指出，AI的快速发展依赖于能源，包括数据中心的运营和建设所需的能源。文章引用了相关研究，讨论了关于高能源回报率（EROI）石油耗尽、以及对AI发展可能造成的潜在影响。文章还提到了三种不同的观点，即文明倒退、无需担忧以及简化与发展可再生能源。文章最后提出了一个问题：未来AI发展是否会因能源短缺而受到严重延误或停滞？

💡 AI发展高度依赖能源: 数据中心、材料开采、运输和加工都需要大量能源，而这些都与石油息息相关。

⛽️ 石油供应的挑战: 部分观点认为，高质量、低成本的石油正在耗尽，这可能对AI发展构成潜在威胁。

⚠️ 不同的观点: 关于石油短缺对AI的影响存在多种观点，包括文明倒退、无需担忧以及通过发展可再生能源来应对等。

Published on July 16, 2025 7:08 PM GMT

( Any feedback from the mods would be greatly appreciated!) 

  This is both my first post here and a question I've been mulling over for quite a while. I came across some discussion adjacent to this topic here on Lesswrong, though I couldn't find any more. This is what I found: 

https://www.lesswrong.com/posts/N7KYWJPmyzB6bJSYT/the-next-ai-winter-will-be-due-to-energy-costs-1https://www.lesswrong.com/posts/xrxh3usuoYMckkKom/preserving-and-continuing-alignment-research-through-ahttps://www.lesswrong.com/w/civilization-scale-energy 

  As it appears to me, many predictions of AI development seem to ignore energy (specifically oil).  It appears to me that energy is the lifeblood of AI development, particularly due to the fact that data-centers themselves require energy to operate ( whether by being plugged into the grid or being plugged into SMRs, or a hydroelectric plant, or a billion very motivated hamsters on wheels). Furthermore, the very construction of data centers requires inputs from oil. Particularly in mining the required materials needed, as well as shipping and processing them.  

  There are those who argue that we are running out of energetically cheap, high quality oil. Some relevant literature is below:

[ Energy Return on Investment, Peak Oil, and the End of Economic Growth]

( 📘 DOI: 10.1016/j.ecolecon.2011.01.021)

[ Assessing the feasibility of the energy transition with the MEDEAS model]

( 📘 DOI: 10.1016/j.enpol.2020.111247 )

[ Comparative net energy analysis of renewable transition pathways]

(📘 DOI: 10.1016/j.energy.2016.03.089)

[ Resource constraints for clean energy technologies ]

(📘 DOI: 10.1016/j.resourpol.2020.101529 )

                                                                                                                                                    

There appear to be multiple camps on this issue

1) " When we run out of high EROI ( Energy return on investment ) oil, our civilization will inevitably regress back to a pre-industrial standard of living"

2) " There is no need to worry, oil companies don't care about EROI, and peak oilers have been wrong before"

3) "Running out of oil may force us to simplify certain aspects of our civilization, while building up renewables and nuclear early" 

So taking all that into consideration, I'm merely curious as to the probability that future AI development will be seriously delayed or ended due to energy decline. Camp 1) would argue that we should be learning how to farm right now and I wouldn't assign a 0% probability to their view, though I don't know what to assign to it either...

Discuss