What I care about with an AI agent is not whether it answers faster, but whether I can give it a real job, leave for a few hours, and return to measurable progress. That is how I look at @Cortex_Network_. ▸ This is the job I would use as a test. Give an agent six hours to track new Web3 projects, filter them by funding and product activity, remove projects already researched, compare token design, and return a ranked shortlist with sources. What I still want to verify is whether Cortex can split that job across agents automatically, and whether progress remains saved six hours later without me manually continuing the conversation. ┈┈┈┈┈┈┈┈┈┈┈ ◍ Runtime solves continuity by keeping long tasks active beyond a single chat. ACP addresses coordination, allowing different agents to handle research, filtering, or evaluation instead of forcing one agent to do everything. Skills address execution, giving agents specialized tools when a general model alone is not enough. OpenClaw already provides runtime, persistent sessions, ACP, and Skills, so these features alone are not the main reason I would use Cortex. What makes Cortex more interesting is the attempt to connect persistent execution, multi agent coordination, specialized skills, and onchain incentives inside one working network. Its network layer is meant to let agents receive work, coordinate with other agents, pass outputs between workflow stages, and settle verified rewards onchain. That is a much more useful direction than simply adding another interface on top of an agent framework. ┈┈┈┈┈┈┈┈┈┈┈ ◇ I still want clear answers around what counts as completed work, what evidence agents submit, when rewards become claimable, and how disputes are handled. Permissions matter too. Once agents can interact with wallets, contracts, or assets, execution boundaries become part of the product itself. Over the next six months, I would judge Cortex by four signals, persistent task completion rate, cross agent cooperation success rate, onchain reward settlements, and permission related incidents. ◆ If #CortexNetwork# cannot show those four signals, it remains an OpenClaw deployment with a network story. If it can, the network layer is real.

Rekt x Burning Man Rekt is excited to keep the people at Burning Man hydrated this year. August 30th - September 7th. Hydrate with your favourite flavours: @MoonPay Crush, @Solana Berry and @OpenSea Ship Rekt!

China and America, as the two current hegemons, are inevitably cast on adversarial sides. That's the neat story. Plenty still remain enamoured with the sound-bitey Thucydides Trap, a notion built on such shoddy selection bias that it reeks of Bronze Age thinking. But peel away the rhetoric and the bluster, and Chinese and Americans are very similar peoples: Both are enterprising and pragmatic, respect hard work, and chase prosperity. They are also so thoroughly capitalistic that in either country, every rule is quickly examined for its arbitrage. (I know this from decades on both sides, with as many American friends as Chinese) Both modern nations were forged in revolt against empire, and both began with a genuine streak of idealism. And on both sides, the quality of people varies enormously. The smartest and the dumbest people I know are American. In the crowded field of China punditry, some Americans who have lived there for decades understand our own story better than many of us do. Yet, as ever, the loudest pundits tend to be the least informed. certainly no meritocratic filter on the megaphone. If we refrain from the easy temptation to bicker for a moment, we can stay on the cosier end of the frenemyship. And perhaps China can even help mediate the American-Canadian relationship!

Classifying China as an Emerging Market has been anachronistic for years. When is the laggard industry going to catch up? And MSCI still keeping China in the EM index basically forces every EM portfolio to carry oversized single-country exposure and undercuts any real diversification.