New Week #132
New Week #132
The US wants to block Chinese access to advanced AI models. Plus more news and analysis from this week.
Welcome to this update from New World Same Humans, a newsletter on trends, technology, and society by David Mattin.
If you’re reading this and haven’t yet subscribed, join 25,000+ curious souls on a journey to build a better future 🚀🔮
To Begin
It’s been a busy week when it comes to the new world taking shape around us.
In this instalment, an update on the ongoing clash between the US and China to win the quest for advanced AI.
Also, a postcard from my travels as a speaker and consultant.
And new research shows solar and wind are the fastest scaling sources of electric power in history.
I hope you enjoyed the Singularity essay I posted earlier this week. More long form essays coming soon.
Let’s get into it.
🇨🇳 Intelligence wars
This week, further news on US efforts to contain Chinese access to AI.
Not content with restrictions on the export of advanced AI chips to China, the US Commerce Department is now considering a lockdown on the most advanced models, too.
Currently, there’s nothing stopping AI leaders such as OpenAI or Anthropic from licensing their most sophisticated models to Chinese companies. It looks as though that could soon change.
Rumours of the proposals came as the Secretary of Commerce, Gina Raimondo, said this week that if China were to invade Taiwan and cut off US access to the high-end AI chips made in that country, that would be ‘absolutely devastating’ for the US economy.
Last week I looked at the evolving picture on Taiwan and the crucial role of its flagship chip-maker TSMC. All AI roads currently lead to TSMC; it manufactures 90% of the most advanced AI chips on the planet, including the Nvidia GPUs that are used to train large models such as GPT-4 and Meta’s Llama 3.
In that instalment, I floated the idea that some kind of Grand Bargain may be taking between the US and China when it comes to compute and AI.
Meanwhile, the US isn’t going to stop at limiting China’s quest for machine intelligence. It’s also racing hard to maintain its lead. This week it announced $285 million in funding to accelerate chip simulation technologies.
This money comes straight from the US CHIPS Act. That act is intended to pour $280 billion into US research on AI and compute across the coming years. Adjusted for inflation, that’s more than the US government spent on the entire Apollo moon programme.
A true moonshot.
⚡ NWSH Take:
Another sign, then, of how seriously the US government is taking the race for compute and machine intelligence.
Perhaps they’re listening to Sam Altman, who this week doubled down on OpenAI’s commitment to AGI:
‘Whether we burn $500 million a year or $5 billion—or $50 billion a year—I don’t care, I genuinely don’t…As long as we can figure out a way to pay the bills, we’re making AGI. It’s going to be expensive.’
There are still huge questions, of course, around how close OpenAI — or anyone else — is to something we can meaningfully call AGI. But this kind of talk surely makes the US government nervous. Indeed, there’s an argument to be made for the idea that government officials should walk into OpenAI and see how close they really are to general intelligence. We can journey down that rabbit hole another time.
My perspective on this news, though? Blocking Chinese access to the most sophisticated models is a necessary move. The US has the lead when it comes to advanced AI research and development. But it’s a narrow one, and it needs to be protected.
Yes, these kinds of moves run counter to the globalised free trade nirvana that people my age grew up believing we’d inhabit by now. But the world has changed. The Chinese are developing a form of techno-authoritarianism the like of which we’ve never seen. We’re amid a clash of two Great Systems, and machine intelligence is the faultline along which that clash will play out most energetically.
For all its flaws — despite the messy, noisy, sometimes absurd reality that is the democratic system — I’d still prefer it if the Global North prevailed on this one.
A final thought: in the race to own the exponential future, Europe is currently nowhere. This week former Google CEO Eric Schmidt painted a doleful picture of the EU’s tech prospects:
‘It’s going to be rare to see the successes in Europe that you see in America and the UK. The fundamental reason is that they can’t get the capital, and the regulatory structure is so prohibitive. It’s a shame, because the talent pool is amazing.’
It’s hard to argue:
Right now, the EU leads the world on AI regulation.
This newsletter is built on the idea that if we’re to make sense of the technological revolution unfolding around us, we need to see it through the lens of fundamental human needs and collective human systems. No one is more persuaded than me of the idea that we’ll need regulatory frameworks around machine intelligence.
There are huge collective choices to be made around our future relationship with AI. That means doing politics and, yes, eventually regulation.
But if Europe can only regulate, it’s going to get rolled over. It will become a rule-taker in the coming age of intelligent machines. It will live as an afterthought in a world shaped by the two great AI powers.
No wonder a new movement is taking shape on Twitter: European acceleration, or eu/acc. European AI startups, advance!
🖐️ Let’s talk
I’m writing this from JFK airport.
That’s because I’m on the way back from Nashville, where I spoke at THINK: an annual conference for 800 credit union professionals from across the USA.
The subject? How to put AI to work inside their organisations, to supercharge their own people and serve tens of millions of customers in new ways.
I’m working with all kinds of organisations — consumer-facing giants, global consultancies, creative agencies — to help them make sense of this AI moment. That means inspiring, deeply actionable keynotes and interactive sessions on the innovation opportunities that must be on their radar.
Want to talk AI? Just drop me a line at mattin@hey.com.
🗓️ Also this week
💥 A new report from leading energy consultancy Ember says the world reached an energy milestone in 2023.
Renewable energy accounted for 30% of global electricity generation for the first time. That’s mainly because solar and wind (especially solar) are scaling far faster than anyone thought possible.
Ember are forecasting that fossil-fuelled electricity generation will start to decline globally. That view aligns with the consensus narrative: fossil fuels are on the way out.
But contrarians argue that we’re massively overplaying our ability to wean ourselves off carbon. They point to unresolved issues with solar, including intermittency (the sun doesn’t always shine, etc) and challenges associated with transmitting power across long distances from solar plants to point of use.
And then there’s the AI question. The data centres we’re building to support AI workloads are hugely energy intensive. Cue a renewed conversation on nuclear.
I’ll dive deep on all this soon.
🥼 Google DeepMind released AlphaFold 3, a major upgrade on its flagship AI model for the life sciences.
The original AlphaFold made headlines when it was unveiled in 2020. It was able to predict with a high degree of accuracy the final 3D folded structure of any protein given only its amino acid sequence.
That meant it had essentially solved the iconic protein folding problem, which had foiled biologists for more than 50 years.
AlphaFold 3.0 is even more powerful. Now, it can handle protein-molecule complexes containing DNA, RNA and more.
Proteins are the building blocks of life, implicated in every system in the human body, and pretty much every disease.
This new model addresses significant limitations. And it edges us closer to a revolution in our ability to discover new therapeutic drugs, a process that’s currently slow (think ten years) and insanely expensive (think $3 billion).
I went deep on all this for The Exponentialist this month; there’s still a chance to grab a six day trial for just $1.
Machines of Loving Grace
Thanks for reading this week.
The epic quest between the US and China for world-beating machine intelligence is one of the great stories of our lifetimes.
This newsletter will keep watching, and working to make sense of it all.
Now you’ve reached the end of this week’s instalment, why not forward the email to someone who’d also enjoy it? Or share it across one of your social networks, with a note on why you found it valuable. Remember: the larger and more diverse the NWSH community becomes, the better for all of us.
I’ll be back next week with another postcard from the new world. Until then, be well,
David.