Over 200 subscribers
This year is not gonna let go is it? The trade wars start, and we get one of the worst drops in crypto of recent memory with $8-10B in liquidations.
We're not going to claim we're experts on trade tariffs, but think it continues to show that our real world decentralization thesis is starting to play out, at least on the macro side if not on the technological side yet.
It also all got announced while I was reading the part of Americana where the author explains how the US moved away from tariffs as the main source of revenue and into the personal income taxes era. Fun timing. Superb book, highly recommend it to everyone.
The rest of this edition is courtesy of Francesco Moiraghi as I was in bed with the flu.
Grok can do a good summary of all the Deepseek week-2 content, with post-market crash analysis and in a meme-rich format. Among the takes of all the AI leaders who weighed in, it's worth mentioning Dario Amodei's "On Deepseek and Export Controls" (for an antagonistic take, see here). Our view is that competition will likely drive to lower costs, and better models. We do support Zetta's take (and hope the trend has the exact opposite outcome of Chairman Mao's Hundred Flowers strategy). Could we ever call ourselves unruly and believe in one model to rule them all? That it's still early to announce the winners of the AI race, is also clear on the foundational level. Speaking at the WEF in Davos, Yann LeCun said that new architectures will emerge soon, giving current LLM paradigms a shelf life of five years. That we have entered an age of efficiency has become a mainstream view thanks to voices like Klarna's CEO Sebastien Semiatkowski. Competition from new players means that even foundational models will also have to adapt to the new age.
Despite all the irony on time to market, fusion energy is on the map for good and will be a decisive technology of this century.
Scientists at the Experimental Advanced Superconducting Tokamak in Hefei, China announced the reactor generated plasma for 1,066 seconds on the 20th of January. And according to a US research organization, China might also be building a massive laser-ignited fusion facility in Mianyang. On the other side of the Pacific, Sama-backed Helion announced a $425m Series F joined by Lightspeed and Softbank, among others.
Paces, Scale Microgrids and Stripe published an excellent white paper on how off-grid solar microgrids can power AI overcoming the speed challenges of grid expansion while being, hopefully, cheap enough. LCOE workbook included. Interesting to read alongside Nat Bullard's 2025 State of Decarbonization.
But the Pacific Ocean is warming up regardless of plasma and climate change. Onebrief has announced its $50m Series C co-led by General Catalyst and Insight Partners to accelerate decisions, planning and staff operations in defense. Swiss semiconductor company SEALSQ has unveiled quantum-resistant secure hardware to protect cryptographic systems. We will need this and much more to secure financial institutions, economies and governments in a post-quantum cryptography world. In the quantum realm, French company Alice & Bob announced its €100m series B to make the NISQ era as painless as possible. Good news among increasing warning signals that Europe might be falling behind in quantum too.
Talking about efficiency, maybe we'll be able to treat sleep as a drug target. "Insane life alpha", to be able to spend 10 fewer years in bed. It could be an insane GDP-growth hack for countries, in a few decades. Can't wait to do without those annoying supervillain sleep machines.
Something else to look forward to is Roche's NGS team presentation of their latest Sequencing-by-Expansion technology, on the 20th of February on this webinar. Someone is already thinking about the implications. But nanoporeans and sequencing geeks will certainly have fun.
You can't do a newsletter without checking on those robots. We loved listening to Jenny Read's interview on the Audrow Nash Podcast. Some great insights into sensing, muscles, control systems, timeline, bottlenecks, and some details on ARIA's work in the field.
"What-ifs" are such great tools to challenge the way things are. What if evolutions had given us different models for robotics, is one question all bio-inspired designs should ask. The research topic "Redesigning the Tree of Robotic Life" is open to submissions. We're so excited to see what will come out of it.
And if you feel like you've missed out on some of the fundamentals, Gleb Zarin's Onboarding to Robot Learning is a great place to start.
A review of all the browser using AIs
Corca App - An app that lets you create mathematic equations and play with them.
On the Central Dogma of Biology - A refresher by Asimov Press.
