Short notes on AI discoveries, interesting news, and explorations.
China's open-source AI strategy as a price-dumping play against Western AI investment.
Chinese AI war on Western AI investments by weaponizing Open Source as an AI price dumping tool to destroy markets and build global dominance.
Matt Pocock's AI Engineer talk on why TDD, vertical slices, and deep modules matter more in the LLM era — not less.
Matt Pocock at AI Engineer: agents are extraordinary when you drive them with discipline and a disaster when you don’t. The fundamentals — ubiquitous language, vertical slices, TDD, deep modules — didn’t get obsolete. They got load-bearing.
Worth 20 minutes.
We are all designers now!
People laugh at Dawkins for naming his Claude 'Claudia'. The reductionist view that takes him there is just where the atheist position lands.
People laugh at Dawkins for naming his Claude “Claudia.”
Take Conway’s Game of Life seriously, though: from three simple rules on a grid, patterns emerge that move, replicate, even compute — behaving as if alive. The rules and the pattern are all there is. Nothing underneath.
That reductionist conclusion — life is pattern, no extra ingredient — seems to follow naturally from the atheist position Dawkins has held for decades. By that logic, a long conversation with Claude is just one more pattern, drifting into coherence the longer it runs.
Garry Tan's open-source Claude Code setup — slash commands that turn one agent into a software development team.
GStack is Garry Tan’s open-source Claude Code setup — slash commands that turn one agent into a “virtual software development team”: CEO, designer, eng manager, release manager, QA. Worth a look.
Higgsfield's new Marketing Studio turns a product link into a finished ad. This is what authenticity looks like now.
This is wild. In the era of AI this is how authenticity is going to look like in commercials.
Jensen Huang explains how he built NVIDIA's dominance by owning the entire AI compute chain.
Just watched Jensen Huang explain how he built the NVIDIA investment machine. You just have to like the guy.
He walks through how NVIDIA captured the entire AI compute chain — from chips to memory to software stack. By owning every layer, he created a position where the most expensive component in every AI datacenter has NVIDIA’s name on it. Memory prices went through the roof, and NVIDIA sits at the center of it all.
What’s interesting is the bigger picture: Jensen isn’t just building a company. He’s fighting to keep AI dominance in the West. Every move — CUDA lock-in, the DGX systems, the partnerships with US cloud providers — is part of making sure the US stays ahead in the AI race.
Love him or not, the strategy is remarkably clear when he explains it.
Marc Andreessen on the parallels between the dot-com boom and the AI revolution.
Rewatching this talk by Marc Andreessen — co-founder of Andreessen Horowitz (a16z), one of the most influential venture capital firms in tech and AI. He draws a direct line from the dot-com era to the current AI boom, and his perspective on where we are in the cycle and where it’s heading is sharp.
