Do you think the AI/Data center companies would be willing to pay for non-battery paired solar/wind in that scenario? If it would be $ out of their pockets, wouldn't they want their $ to bring something firm online (or at least something paired with storage)?

@katieadkins.bsky.social to validate…

Truly an amazing graph. But even with a boatload of SMRs, we'd need a ton of transmission as well unless all were truly co-located

Who knows if the models are that close (or what even constitutes AGI), but the fact that more people are zeroing in on the arms race aspect with China, says something meaningful...

To many, it felt like we were in the 1st or 2nd inning of this AI game and that mainstream calls for a Manhattan Project to achieve AGI would come, if at all, many, many innings down the road...yet here we are...

Some previously looked at China's long-distance ultra-high voltage transmission lines with longing from a decarbonization/electrification perspective, but their ability to modernize their grid may also prove to be a strategic advantage in a potential AI arms race

To that end, last week OpenAI's head of policy laid out an "AI Infrastructure Blueprint for the US" during an event at CSIS, which was explicitly framed as a pathway to compete with China's race for AI

www.cnbc.com/amp/2024/11/...

www.csis.org/analysis/ens...

China has one acute advantage in the speed in which they build energy assets and transmission--they can measure projects in dozens of months, while we often measure in decades.
www.utilitydive.com/news/interre...

(iii) procuring enough energy, chips, capital, and data to do it before China can (which all assumes the current LLM architecture can get there),

and one heck of a kicker...

(iv), accomplishing the above without wrecking the carbon budget and path to net zero.

Such an arms race to AGI presents several challenges to say the least: (i) needing to win the race to AGI from a national security perspective, (ii) simultaneously ensuring the necessary AI Safety protocols and failsafes (remarkably important & remarkably difficult),...

That said, seeing a bipartisan, Congressionally authorized commission issue a call (in the year of our lord 2024, no less) for a "Manhattan Project" to beat China to AGI is a surprise to many.

But it may be the a sentiment we hear much more in the years to come...

Recently the CEOs of both OpenAI & Anthropic alluded to AGI being potentially achievable by ~2027, and folks like Leopold Aschenbrenner have examined the national security issues of a potential arms race to AGI if we are indeed on the cusp of such a substantial breakthrough situational-awareness.ai

You're doing awesome work coordinating all of this @tnorris.bsky.social!

It's really pretty good. I experimented with leaving it in voice mode on a commute and it worked pretty well for like 25 min. The only challenge was a bit of the timing of the back and forth.

Ethan - your book is excellent and I recently added to the syllabus of a course I teach on technological change and the law. It's a really great on-ramp for attys and non-tech folks to see where things are headed and the different scenarios. (Especially with the CEOs talking of AGI w/in 2 years…)

Hi Miles - can you please add me? Thanks!

Hi Energy Friends - I work on clean energy legal issues as SVP-Legal at Sunrun, and I also teach clean energy law course at U of Missouri. My side quest/deep dive is AI and particularly how energy laws can be used to speed up or slow down the development of AI. Also love me some movies & pinball…