🔥The 9th Round of Easy Loan, Earn $40 Reward is in progress❗️ ⏰ Promotion Period: January 15th - Feburary 15th, 2025 👉 Register now and check more details at gate.io/campaigns/358

Great piece by JS Tan challenging assumptions about data centers driving local growth: -Few permanent jobs (just 30-50 workers per facility) -Weak downstream effects (Loudoun County (VA) has world's most data centers but no tech hub) -Most advanced components still imported

Everyone complains that the language models are stochastic and don't understand mechanisms and thus can't be trusted in mission-critical applications, but I just realized that when I was writing a report last weak I added 3+3 and somehow got 7, so...

global AI leadership will depend on diffusion of the tech into intl economies. ubiquity will be economic power will be global influence. a purely private-sector diffusion strategy cannot compete w a gov-backed export-finance model, which China has shown it does very very well

Britain’s long-term economic growth and strategic heft depend on winning the global contest for top-tier scientists, engineers, and technical founders. Following the government's announcement of a new head-hunting unit for scientists, engineers & founders, I wrote for

Tanay Jaipuria And it’s gotten worse in the 6 months since we first measured. Google is now: 15 scrape for 1 visitor OpenAI is now: 1,200 scrapes for 1 visitor

"The reason why we don't have Bell Labs is because we're unwilling to do what it takes to create Bell Labs — giving smart people radical freedom and autonomy."

Aaron Levie Prompt engineering seems to mean roughly "this thing kind of works, but just barely, so we have to tell it what to do very carefully," and technology often switches rapidly from barely works to just works.

Corollary: The fact that we currently have such a thing as prompt engineering means we don't have AGI yet. And furthermore we can use the care with which we need to construct prompts as an index of how close we're getting to it.

Europe should pay close attention to Saudi Arabia and UAE's massive AI investments. What's striking is how these Gulf state commitments completely dwarf Europe's AI investment. In my view, UAE is already positioned to become #3 globally in AI, behind only the US and China.

new from me over at Works in Progress - a relatively comprehensive look at the many different types of nuclear reactor, their prospects, and what innovation may or may not be coming...

National prosperity and security depend on strategic capabilities – eg vaccine production. Britain’s capabilities have been allowed to wither. I wrote for Centre for British Progress with Ben Johnson, Louise + Andrew Bennett about how the Industrial Strategy can rebuild them.

"Over the past 50 years, the U.S. has created, from scratch, 241 companies with a market capitalization of more than $10 billion, while Europe has created just 14." wsj.com/tech/europe-bi… via The Wall Street Journal

People who don’t/haven’t worked at Google fail to understand the insane amount of resources they have. They have the talent + compute to compete in every ai market if they want to. Product polish is their only weakness

It’s really surprising OpenAI was founded in California, when places like Tennessee and North Carolina have friendlier business climates.

Very big impact: The final version of a randomized, controlled World Bank study finds using a GPT-4 tutor with teacher guidance in a six week after school progam in Nigeria had "more than twice the effect of some of the most effective interventions in education" at very low costs

wow. gemini diffusion is FAST.

Yes, we do. It's ~21GW. You count all the AI chips produced, factor in that they're running most of the time, add some overhead—and you got your answer. It's a lot. And will only get more. But you know what? Probably worth it.

JD Vance told Ross Douthat that he read the AI 2027 paper

1/ Export controls, national supercomputers, poaching scientific talent... are not new! New preprint from me on the global race to build the first supercomputers for numerical weather prediction (NWP) and its relevance to the modern AI race today: