Dear Readers,

How do you design a future that’s both intelligent and alive? This week’s stories draw a line between two revolutions quietly merging — the one reshaping machines, and the one rewriting biology. AI is not just thinking faster; it’s starting to think differently. And medicine, once confined to treating decay, is learning to reverse it. Somewhere between Extropic’s thermodynamic chips and Lilly’s bet on cellular reprogramming, the boundary between digital and organic evolution begins to blur.

In Today’s Issue:

🧠 Extropic unveils a probabilistic hardware

💊 Eli Lilly invests $45M in NewLimit

🍎 Fasting cycles are found to trigger immune system renewal

🚶‍♀️ A new study suggests longer continuous walks are more beneficial

✨ And more AI goodness…

All the best,

The Google DeepMind and Google.org initiative “AI for Math” brings together top research institutions like Imperial College London, Institute for Advanced Study, Institut des Hautes Études Scientifiques, Simons Institute for the Theory of Computing and Tata Institute of Fundamental Research to accelerate mathematical discovery using advanced AI tools. They’ve already achieved impressive results — their system improved solutions in ~20% of open problems and broke a longstanding record in matrix-multiplication algorithm complexity.

Bill Gates: AI is the biggest technical thing ever in my lifetime

Extropic has unveiled a bold vision for the post-GPU era: a new hardware architecture called the Thermodynamic Sampling Unit (TSU). Instead of crunching endless matrix multiplications like today’s GPUs, TSUs work by sampling directly from probability distributions—essentially mimicking how generative AI models think. Each processing element, or “pbit,” acts like a biased coin, flipping between states according to thermodynamic laws.

This design could make generative workloads exponentially more efficient. Extropic claims their probabilistic chips could use up to 10,000× less energy for certain inference tasks, aligning computation with the inherently stochastic nature of large AI models. While the company is still building the first silicon prototypes, the implications are clear: if TSUs deliver, they could fundamentally change the economics of AI hardware, from datacenter energy use to on-device inference.

Why it matters: AI hardware is finally being redesigned for how generative models actually function. This could lead to massive efficiency gains, lower costs, and entirely new architectures optimized for probabilistic reasoning.

Sources:
🔗 https://extropic.ai/writing/tsu-101-an-entirely-new-type-of-computing-hardware

Simplify Training with AI-Generated Video Guides

Are you tired of repeating the same instructions to your team? Guidde revolutionizes how you document and share processes with AI-powered how-to videos.

Here’s how:

1️⃣ Instant Creation: Turn complex tasks into stunning step-by-step video guides in seconds.
2️⃣ Fully Automated: Capture workflows with a browser extension that generates visuals, voiceovers, and call-to-actions.
3️⃣ Seamless Sharing: Share or embed guides anywhere effortlessly.

The best part? The browser extension is 100% free.

Check out Guidde

"in the research presented in the paper (...) we can solve simple generative modeling benchmarks using around 10.000 times less energy than the most efficient algorithm running on a GPU.”

Eli Lilly has invested $45 million in NewLimit, a biotech company focused on reversing cellular aging through epigenetic reprogramming. The round values NewLimit at $1.62 billion and follows a $130 million Series B earlier this year, signaling strong momentum in the longevity space. NewLimit’s approach aims to restore youthful function in aged cells across liver, immune, and now vascular systems — pushing the field closer to clinical trials.