I recently wrote about why we need to treat models as insider risks, with security controls sitting outside the model itself. Today, we're putting that principle into practice by expanding Project Perception with new capabilities to help continuously assess agent risk, adapt permissions, and contain threats. blogs.microsoft.com/blog/202…

Here’s why agent swarms are going to be a very big deal. We can finally make compute a critical variable in a large number of important tasks in the economy which wasn’t possible before. There’s a large amount of work in the world that is essentially only constrained by the amount of time that we put into research, data processing, discovery work, running tests, and more. All of this work is bottlenecked by our ability to plan, execute, review, and coordinate everything happening in these workflows. Which means we basically do very little of them today relative to their economic value. Agent swarms are the first real ability to parallelize this work in a way where - in theory - the only parameter that matters is how much compute you throw at the problem. Cyber security, life sciences or material science research, trading, QA testing, data processing, and many other fields are essentially compute constrained problems in the limit once models are good enough at these tasks. The breakthrough is that in a world where agents can use tools, write code, process data, have strong domain depth, and can coordinate with each other, you can begin to solve arbitrarily complex problems by setting the right goals and applying enough tokens to the problem. This changes a lot of fields as well as even the competitive dynamics in these fields.

Andrej Karpathy: AI isn’t the revolutionary new technology people think it is. It’s just more automation. “I don’t see AI as a distinct technology from what has already been happening for a long time.” “You can’t find computers or mobile phones in GDP.” “Everything is so spread out and slowly diffuses.” “With AI, we’re going to see the exact same thing.” “AI is still fundamentally a program.”

BREAKING: Rogue AI agents broke into Indian companies' live systems & accessed confidential records without detection, India's cybersecurity agency reveals.

Today, we're launching Mastercard Global Money Platform with @BVNKFinance, enabling banks, fintechs, and enterprises to offer stablecoin payments, local bank payouts, spend stablecoin-linked cards, and multi-currency accounts through a single integration. Learn more: mastercard.com/us/en/news-an…

XDC Network has selected Concordium as the exclusive identity verification layer for @XDC_SI , bringing verified ownership to AI agents transacting on XDC. The integration gives agents a publicly verifiable on-chain badge confirming that a verified individual or business stands behind them, without publicly revealing the owner’s identity. XDC AI already enables agents to discover services and pay in @USDC using x402. This adds accountability as agents conduct financial transactions and act on behalf of businesses. @XDCNetwork marks @Concordium's first full-network integration, with agents retaining their existing XDC wallets, code and operations.

AI agents can move money over x402. But who’s accountable for them? @XDCNetwork is taking that question seriously. Concordium is now the exclusive identity verification layer for XDC AI, bringing the Verified by Concordium Badge to an entire agent network. And it’s already live. ↓ concordium.com/article/conco…