The fight for global AI supremacy, and Steve Bannon versus the Tech Right
Welcome to Cautious Optimism, a newsletter on tech, business, and power.
📈 Trending Up: +16.1 million … the cost of Apple’s App Store tax … o3-mini, shortly … data backing a nation of immigrants … vertical AI? …
📉 Trending Down: Saudi oil in the United States … Russian banking health … Pinterest’s share price after a downgrade … the Russian economy … chances of domestic rate cuts at the next Fed meeting …
Crypto prices are sharply falling, with bitcoin dipping to the $90,000 mark in recent trading.
The fight for global AI supremacy
To borrow from sport: it’s getting chippy out there.
This morning, the Biden-Harris administration released an Interim Final Rule on Artificial Intelligence Diffusion, a fancy way of saying that there are new rules in town for AI-related chip exports.
The gist is a three-tier system, with 18 countries granted a no-restrictions license to buy as much US AI chips as they want. A second tier will face caps on their purchases, with possible exceptions to allow for greater exports to those nations. And then there’s the third tier — China, Russia, etc — who face, per a FT analysis of the new rules, an effective ban on US chip buys.
The new ruleset, the current administration says, “builds on previous chip controls by thwarting smuggling, closing other loopholes, and raising AI security standards.”
And everyone seems to hate it. Nvidia went scorched-Earth over the new rules, name-checking the first Trump administration as a positive example of AI policymaking, and saying (emphasis added):
In its last days in office, the Biden Administration seeks to undermine America’s leadership with a 200+ page regulatory morass, drafted in secret and without proper legislative review. This sweeping overreach would impose bureaucratic control over how America’s leading semiconductors, computers, systems and even software are designed and marketed globally. And by attempting to rig market outcomes and stifle competition — the lifeblood of innovation — the Biden Administration’s new rule threatens to squander America’s hard-won technological advantage.
While cloaked in the guise of an “anti-China” measure, these rules would do nothing to enhance U.S. security. The new rules would control technology worldwide, including technology that is already widely available in mainstream gaming PCs and consumer hardware. Rather than mitigate any threat, the new Biden rules would only weaken America’s global competitiveness, undermining the innovation that has kept the U.S. ahead.
Well, shit.
Rhode Island’s own Commerce Secretary Gina Raimondo said a few weeks back that chip export controls only do a little to slow China’s AI development. So, what gives? I’m honestly not sure. But I suspect that Nvidia’s clear signaling for relief from the next administration will be heard.
Recall that venture investment in China has cratered, while investment in US AI startups is world crushing. In short: Before this new ruleset got downloaded onto the market, things were looking pretty good if you care about China-US competition and want the Stars and Bars to win?
My most charitable view here is that the Biden-Harris admin have a very particular view of the best way to compete with China and its various allies of ill repute (the governments of Russia, Iran, North Korea etc) and wanted to leave their mark on the world before exiting the stage.
Fair enough, but if you are going to lame-duck something this big, don’t expect much deference from aggrieved parties who feel like you are crimping their future cash flows.
Also this morning, OpenAI dropped a “a new Economic Blueprint” for the nation “as we enter the Intelligence Age.” You can read the full text here. I’m still digesting its entire heft, but a few policy proposals stand out even in a first read:
Regarding “Competitiveness and Security.” OpenAI has a host of proposals that would improve information sharing between the government and private industry. The gist is that without robust two-way communication, AI companies are going to be partially blind, and the US government would not know as much as it needs to. This section is a little technocratic, but seems pretty damned reasonable to my eyes.
Regarding “Rules of the Road.” OpenAI proposes that AI companies do a lot of work to prevent their models from being used to generate CSAM/CSEM. It also makes a very interesting pitch that "those building with AI, from large companies to start-ups to developers and entrepreneurs, should [apply] provenance data to all AI-generated audio-visual content.”
In the same section, OpenAI writes that AI models should “aim to be politically unbiased by default, supplemented by users having the ability personalize the tools.”
And in a final section entitled “Infrastructure as Destiny.” OpenAI unloads the cannons, pitching:
Work to ensure that “AI has the ability to learn from universal, publicly available information, just like humans do, while also protecting creators from unauthorized digital replicas.”
A new “Compact for AI” between “US allies and partner nations that streamlines access to capital and supply chains in ways that support AI infrastructure and a robust AI ecosystem.”
The creation of “AI Economic Zones” by “local, state and the federal government together with industry, that significantly speed up the permitting processes for building AI infrastructure like new solar arrays, wind farms and nuclear reactors.”
The feeling you get from OpenAI’s document is not that it has radical or unexpected policy proposals. Instead, it seems that OpenAI — and, presumably, many of its peers — feel that there’s a lot more that we could do as a nation to accelerate our progress towards an AI-infused future — Sam’s Age of Intelligence — without sacrificing any of our ideals or goals.
I agree. Frankly, OpenAI’s pitch strikes a chord with my own view of the future. So, I am biased in its favor.
Putting aside my myriad issues with former and future President Trump, if the technology wing of his second administration can implement half of what OpenAI wants, it would earn long-held favor from Silicon Valley. Unlike the Biden administration, which, despite what I presume are the best of intentions, just tripped itself heading out of the door.
Good luck, Democratic bundlers in 2028 visiting venture offices.
One more thing
As a final nibble, the U.K. is seemingly serious in its view that it can have a material position in the future AI pantheon. Good on it. Here’s its PM on AI policy, including the following riff that will make a lot of folks happy to read:
I am determined the UK will become the best place to start and scale an AI business. I know growth in this area cannot be state-led. But it is absolutely the job of government to make sure the right conditions are in place.
Even more here from Keir.
Steve Bannon vs. The Tech Right
Trump gets sworn in in a week, which means that who will hold power inside of his second administration is still an open question. And while many in Washington turn to upcoming confirmation fights for key roles, there’s a far scrappier fight going on: The MAGA Right versus the Technology Right.
We’ve written about this at CO often, so I won’t retread apart from saying that there’s a host of policy positions that the Musk/a16z wing of Trump 2.0 and the Bannon wing violently disagree about — including high-skill immigration. And tax policy. You get the idea.
So, when former Trump Senior Counselor to the President and White House Chief Strategist says the following, I take note (via a Breitbart transcription):
“I will have Elon Musk run out of here by Inauguration Day,” Bannon told the Italian daily Corriere della Sera this week. “He will not have a blue pass to the White House, he will not have full access to the White House, he will be like any other person.”
And:
“He is a truly evil guy, a very bad guy. I made it my personal thing to take this guy down,” Bannon added. “Before, because he put money in, I was prepared to tolerate it; I’m not prepared to tolerate it anymore.”
And, via the original Corriere article:
“Peter Thiel, David [Sacks], Elon Musk are all white South Africans... He should go back to South Africa. Why do we have white South Africans, the most racist people in the world, commenting on everything that happens in the United States? Sure, I love Bernie who is only now joining the H1B visa party. We have been fighting this battle for 10 years and Bernie hasn't said a word because of all the money the Democratic Party takes from techno-feudalists. We will expose the corruption of the American system, how money controls everything, and hopefully inspire Italy to wake up.”
That’s stark. The same interview contained even more Bannonisms, but the underlying vibe is that there is a promised power struggle coming in Trumpland. How you handicap Musk’s DOGE and its chances of success is your own call, but it will be harder for the cost-cutters to do their work sans access to the returning President.
So, how much sway does Bannon still have? More than Musk? We’re going to find out pretty quickly. Notably, over on the Tech Right broadsheet X, folks seem to expect Musk to win out over Bannon. I say never bet on what Trump will do until he does it. After all, right?