After 2 years of iteration at scale, here's everything we learned about engineering Fluid. The most desirable (yet seemingly hardest to attain) property of a compute system is *efficient and predictable performance at any scale*. Serverless as pioneered by Lambda had an unusual set of tradeoffs. It has the potential for cold starts in production, which is a no-go. What's lesser known though is that it solved for the achilles heel of servers: noisy neighbors, congestion, and inadequate load balancing. It did so at tremendous expense (1 "computer" per concurrent request), but it was glorious. I sometimes compare this to you and your friend leave work to go to the same exact restaurant, at the same time, but you both order Uber XLs. You'll have an amazing experience, but it's wasteful. When we built Fluid, we wanted to have the cake and eat it too. I'm particularly happy with our algorithm that uses 'every available seat in the car' (great for AI apps that wait a lot on tokens), but spins up more computers horizontally as needed. This is something exceedingly hard to get right, even for experienced DevOps teams.

Consumption is the only sustainable pricing model for infrastructure, and now intelligence, at scale. Kudos to @aws for pioneering the model where, the more you use of an energy-consuming resource, the more you pay, with full transparency and incentive alignment. What's become evident is that when you use tools like @v0 or @claudeai Code, you're actually incurring in highly fluid infrastructure spend, unlike traditional SaaS products. The 'digital workers' you're orchestrating in the form of agents are not dissimilar to spinning up a big compute cluster on demand. And the usage patterns vary greatly by person to person, just like cloud consumption varies from app to app.

Be guided by beauty

Just saw a great PR come in. Killer feature, beautiful description, commit messages, comprehensively tested code (40% of code is tests). Ask the engineer what % of this is AI. The answer: 100%

MCP is like the HTTP for agents, and feels like the foundation of a new internet. We shipped the blog post that answers the most common questions i've been getting from developers, partners and enterprises. Shoutout to MCP's creator @dsp_ for collaborating on this!

AGENT=1 env would be a cool convention, like CI=1

A large e-commerce platform went live on Vercel with a %-based traffic rollout. The resulting performance was so wild they ramped up immediately. 🤑 Within 24 hours, metrics for the Vercel cohort reported by the customer: ▪️ Conversion Rate: +12% ▪️ Revenue Per Visitor: +34% ▪️ Average Order Value: +19% ▪️ Bounce Rate: -5% + immaculate vibes on DX, migration speed, AI support, and all the new features they’re excited to ship

Make something people love

Terminals are amazing. You'll never regret mastering the shell & unix fundamentals (pipes, processes, filesystem), and key tools (curl, dig, jq, ssh, tmux, apt, nvim…) Especially given agents' propensity to wield these

It's been fun collaborating with @superwhisperapp, a blazing fast AI speech-to-text app, with support for local and cloud models. By putting @vercel CDN in front of their model API, they're seeing up to 350ms gains in some geos 🤯 To be clear, this is 350ms+ faster by *just* putting a proxy in front backed by our global network. They haven't even replicated their db / compute yet! It's free real estate. When we talk about 'fast data/origin transfer', this is what we mean. We partnered with @awscloud to back vercel with a fully redundant 400 GbE fiber network backbone. No noisy neighbors, top notch security, no last-mile unexpected rerouting, no free-tier throttling, no enterprise-only POPs or TCP optimizations, no Telekom peering issues…