NEW STUDY: It's dogma that fasting is good for you because it increases autophagy, the recycling of cell parts. But in the largest study ever, 6 months of time-restricted eating (TRE) or calorie restriction (CR) did not increase autophagy Huh? Is fasting still worth it? ... 🧵

Introducing Stablecoin Financial Accounts. Hold a stablecoin balance. Send and receive funds with fiat and crypto rails. Accessible from 101 countries: docs.stripe.com/crypto/stabl….

USDC loans on @coinbase, powered by @MorphoLabs: May 2 - $120,000,000 Apr 2 - $45,000,000 Mar 2 - $14,000,000 Feb 2 - $2,000,000 Jan 2 - $0 Just $265M BTC being put to work as collateral >100x more BTC sitting idle

transaction verification, fingerprint reset, private wallet balances, private purchasing, and purchase it with bitcoin. a lot coming to @bitkey starting in may: bitkey.build/building-better…

MySQL comes to Cloudflare Workers and Hyperdrive: your favorite MySQL drivers and ORMs are now compatible with the Workers runtime, and Hyperdrive allows you to connect to your regional database from Workers with optimal performance. cfl.re/3XPb0UI #DeveloperWeek

Physical infrastructure networks — like telecommunications, energy, water, and transportation — are often natural monopolies. The result: little incentive for innovation, not to mention horrible customer experiences, bad user interfaces, lackluster service, and slow response times. We can do better. Introducing: DePIN. @guywuolletjr of @a16zcrypto on the impact DePIN could have on industries including energy, and what the future may hold 👇 a16zcrypto.com/posts/listicl…

I can't stop thinking about this quote from US Secretary of Treasury Scott Bessent: [W]e are going to keep the U.S. the dominant reserve currency in the world, and we will use stablecoins to do that. However big you think stablecoin AUM will get, you're probably thinking too small. So, the question becomes: What does the rest of the crypto market look like if stablecoin AUM is multiple trillions?

(Vibe) "Coded" a weekend project that parses school emails for pertinent dates + action items and sends me an email summary. Took me 30 minutes to build this using @AnthropicAI + @deepseek_ai via @hyperbolic_labs for email summaries + @vercel for hosting Took me 3 hours to navigate @googlecloud to configure oauth correctly :/

Gemma 3 (our open weight LLM) is here and for the first time available on both Google AI Studio and the Gemini API! It is also: - Natively multimodal - Long context (128K tokens) - Can run on a single H100

The vast majority of developers today use #APIs to accelerate software #Innovation — but doing so has created new security risks for their organizations. cfl.re/43kVKm3

Sofia Calado, Cloudflare’s Lead Privacy Counsel, will share her knowledge and experience about privacy online from the legal perspective on March 6 in our Cloudflare Women in Privacy Meetup in our Lisbon office, in Portugal. Join us: cfl.re/3DncBKe

🎙️ Today we launched The Connectivity Cloud Podcast on Spotify, Apple Podcasts and YouTube… your trusted source for insights into the latest trends, strategies, and technologies shaping cloud security and infrastructure. Each month, we look into the challenges faced by decision-makers—whether it's connecting users, networks, apps or clouds, globally.

GPT-4.5 is the first time people have been emailing with such passion asking us to promise to never stop offering a specific model or even replace it with an update great work @kaicathyc @rapha_gl @mia_glaese

BREAKING: 🇺🇸 President Trump says “obviously” Bitcoin will be included in the strategic reserve. “I also love Bitcoin”

Say goodbye to Chain-of-Thought. Say hello to Chain-of-Draft. To address the issue of latency in reasoning LLMs, this work introduces Chain-of-Draft (CoD). Read on for more:

GPT-4.5 is ready! good news: it is the first model that feels like talking to a thoughtful person to me. i have had several moments where i've sat back in my chair and been astonished at getting actually good advice from an AI. bad news: it is a giant, expensive model. we really wanted to launch it to plus and pro at the same time, but we've been growing a lot and are out of GPUs. we will add tens of thousands of GPUs next week and roll it out to the plus tier then. (hundreds of thousands coming soon, and i'm pretty sure y'all will use every one we can rack up.) this isn't how we want to operate, but it's hard to perfectly predict growth surges that lead to GPU shortages. a heads up: this isn’t a reasoning model and won’t crush benchmarks. it’s a different kind of intelligence and there’s a magic to it i haven’t felt before. really excited for people to try it!

Try Grok DeepSearch. Tell it to use more time for a better result.

I really like MCP by @AnthropicAI. It aims to be the USB-C of llm tools. However, it's incomplete right now: To follow the analogy, right now its cable and wire protocol. It's missing the plug. I think MCP should suggest a textual standard of pointing the LLM to a MCP endpoint from the chat window. Maybe use the + symbol, as in +mymcp.tools/endpoint.mcp for SSE or +"npx @modelcontextprotocol/server-filesystem /tmp/files" for local cli tools as syntax? Obviously there would have to be appropriate security disclaimers when used for the first time. I think having a syntactic standard now would allow all the tools that are implementing MCP a common approach to trigger the install process on top of whatever else they do. That would be good.

To build product, “make something people want.” To create art, make something *you* want. (The best do both).