8 countries. One look at how the world actually moves money.
UPI, FedNow, Pix, SEPA, AANI, and Alipay, every market runs its own rails, speeds and settlement.
The world isn't converging on one rail. We connect all of them.
$500M processed · 40 markets · settled in stablecoins.
→ credible.finance
Today I'm publishing a new essay, Policy on the AI Exponential. AI is progressing extremely fast—much faster than the policy process was built to handle. The essay lays out where I think the technology is now, and the action needed to close the gap: darioamodei.com/post/policy-…
We aren't born with the perfect storage architecture. Experience actively swaps out a noisy, localized network for a quiet, hyper-efficient memory machine.
nature.com/articles/s41467-0…
The brain remains the ultimate hardware benchmark: it is 100,000x more energy-efficient than today’s digital compute.
The Intellectual Revolution is getting an organic hardware upgrade. 🧵
The Reality: If you asked today's most advanced Al models to design the most energy-efficient, powerful computing architecture in existence-the math leads them straight back to the human brain.
We aren't just building better software; we are re-engineering the physical interface of intelligence. We are replicating the brain with machines, minerals and energy.
#OI
A 75 kg humanoid robot from KAIST can run at 13 km/h while carrying 20 kilos.
It also performs jumps, moonwalks, and soccer skills on grass with strong balance.
The robot even pushes heavy objects using human-like full-body movements.
Cognition is Scaling. Energy Must Follow—or It Will Bottleneck Superintelligence. Indeed, we’re entering the Intellectual Revolution: 99% of Earth’s cognition will soon run on machines, externalizing memory and reasoning to neural networks.
Is silicon the only material that can meet this demand? An in a world with nuclear, fusion, solar, wind, tidal, and space-based energy, why is power still the constraint?
Energy production is gatekept n outdated, stuck on legacy infrastructure that simply cannot keep pace with the hyper-growth of machine AI. Software will write the future. Hardware will power the pen. How are you factoring AI-scale energy & compute into your investment thesis?