Very surprised that the GPT Router isn’t discussed as much as it should. In my view, THIS is the main announcement and not GPT5. It’s a major unlock in so many levels for OAI: > allows them to optimize costs by routing dumb queries to less expensive models > enables them to really become an agentic super app > hence, leave behind the unsustainable subscription business model and turn on the “cut per agentic transaction” model potentially taking down Google and showing that Search is irrelevant

Nobody does Regulatory Capture better than banks. This is some Tony Soprano level of arm twisting.

Some good insights from @dLocalPayments Earnings Report: > the S-curve of digital merchants adopting emerging markets payment localization. > all areas of product innovation (credentials, BNPL, stables) can either directly on indirectly be enabled by blockchain > despite the inherent global nature of stables, local fiscal policies/regulation/macro trends will always create challenges in adoption and GTM strategies (what I call Glocalization)

V29 upgrade for @zksync has now been posted by the Matter Labs team. This upgrade set us up very nicely for two of our stated goals: ↔️ Interop. First step is to enable messaging between ZKsync Layer 2s that are connected to the Gateway 🏎️ Fast Finality. Thanks to having the fasted ZK prover in the game, Airbender, we can set our eyes on L1 finality for ZKsync Layer 2s within seconds. This will be a game changer for third party bridges, solvers and market makers, CEX withdrawal/deposits, and overall user experience

The TAM is the Internet Economy. $ZK

Reminder: USD stablecoins are the single most important 🇺🇸 geopolitical tool

Institutions aren’t asking “if” anymore, they’re asking “how.” We see the same shift with partners building enterprise chains on ZKsync. Good, pragmatic read by @a16z on moving from pilots to production.

On Stripe/Circle L1s: > from a first principles, it doesn't make a lot of sense that these aren't L2s. the argument "they already have millions of users, they don't care about ETH ecosystem effects" is nonsensical. there isn't any business in the world that thinks in terms of "i have enough users, why go somewhere with a bigger TAM"? > a legit but, imo, short-sighted argument for an L1 is simply bigger valuation/token price > both Stripe/Circle are strong companies with smart people but never underestimate how layers of management/bureaucracy/not good knowledge of an emerging industry can lead organizations to poor decision making > from a first look, this may seem like a miss for Ethereum but it opens up new avenues for product innovation and GTM. as a teaser, I will remind everyone that arguably the biggest innovation from Anthropic isn't Claude but MCP 🤓

You are not working in Crypto. You are working in World Building. That’s the right mindset.

> it's true that not every enterprise will run its own chain. it depends a lot on the use case they are building > since we are in the CeFi era of Crypto, most financial institutions will indeed run a chain since it's the only way they can launch in a secure, private and compliant way > it's also true that running a chain has become easier/cheaper the past 2 years and the logical progress is that it will be commoditized > the main fallacy in👇argument is that running a chain is "context" and not "core". this is wrong. running a chain itself isn't a differentiator but it enables differentiation in many levels through customizing and owning the experience. the same way that owning a store isn't per se a differentiator VS running a store in a mall but it allows a retailer to do all kinds of customer engagement things that they can't do in a mall