Malte Ubl✓
@cramforce
Self-driving infrastructure @Vercel CTO. Immigrant 🇺🇸/🇩🇪/acc
940 Following52.1K Followers61.5K Posts
- Vercel CI is 1.6x faster in the worst case. The real magic is coming when task caching can prune down the work further. Also: This is the worst it ever will be. We'll be grinding down every millisecond that can be savedJono @jonoalford@cramforce Tests from @studioroboto so far show it’s pretty god damn fast, going to see what it benchmarks on turbo-start-sanity42382.7K
Posts @cramforce@cramforce Tests from @studioroboto so far show it’s pretty god damn fast, going to see what it benchmarks on turbo-start-sanity
02132.7K- Agent: Look back at our session and tell Vercel what they could have done better 🖤Vercel Developers @vercel_devYour agent can now send feedback to Vercel. Use plain language to explain a bug or feature request, and your agent will use 𝚜𝚞𝚋𝚖𝚒𝚝_𝚏𝚎𝚎𝚍𝚋𝚊𝚌𝚔 via Vercel MCP to attach structured context and submit the comment.61232.6K
Your agent can now send feedback to Vercel. Use plain language to explain a bug or feature request, and your agent will use 𝚜𝚞𝚋𝚖𝚒𝚝_𝚏𝚎𝚎𝚍𝚋𝚊𝚌𝚔 via Vercel MCP to attach structured context and submit the comment.9213442.4K- With Vercel CI, 1 out of the top 3 most exciting things that we have in the oven shipped. Stay tuned.60534.1K
- Vercel CI isn't just faster CI. It's CI that does strictly less work: Tests and build tasks that couldn't possibly break by a change are automatically skipped. Only changed the backend? No need to run the linter on the frontend. But it will run your e2e tests. Vercel CI determines the minimum amount of work automatically with zero config. There are no magic triggering lists to maintain that quickly get out of date. Try it out today with this prompt: Set up Vercel CI for this repository. Run `npx skills use vercel/ci@setup` and follow the skill's instructions to select initial tasks, configure Vercel CI, and open a pull request.Vercel @vercelIntroducing Vercel CI • Intelligently skips tasks with Turborepo • 2-6× faster CI in our internal repositories • Optimized for agents via Vercel MCP and CLI After nearly 8 million tasks, we're excited for the future of agentic CI. In beta on all plans.21314312.9K
Vercel@vercelIntroducing Vercel CI • Intelligently skips tasks with Turborepo • 2-6× faster CI in our internal repositories • Optimized for agents via Vercel MCP and CLI After nearly 8 million tasks, we're excited for the future of agentic CI. In beta on all plans.102781.4K234K- True at Vercel since 2015Theo - t3.gg @theoIt is easier to build features than it is to announce them60864.1K
- What I learned today: - deepseek/deepseek-v4.1-flash-1017 works great in Claude Code for sub agents (just ask it to write an .claude/agents/ file as the sub agent tool doesn't take a tool arg) - With Opus 5.5 doing orchestration everything just works wonderfully - It's extremely cheap. I currently have 24 sub agents cooking (all with their own Vercel Sandbox to leave my laptop alone) and until now I have spent like $9 on deepseek which does 99% of the work 🤯 - Previously I averaged about $2K per day in list-price tokens (Sorry @mabrahamsen) Even if you have a Max subscription, just imagine how much more you could do552043429.3K
- The new normal: Dev: I built a new multi-writer distributed-queue transactional zero-proof foo system. Me: How does it work? Dev: I don't know yet492959423.1K
- Malte UblRT @shuding: UI and math time! UI = f(props) is the most famous line in React. But that f is JavaScript, not math. Inspired by @_chenglou’s Handwired project, I compiled a <Dashboard/> component into a real math function, and asked what becomes possible.02150124
- Malte UblRT @browserbase: Introducing BELL, the complete OSS stack to build your own personal assistant. @browserbase for browser infra @eve by Vercel for harness @thelinqapp for iMessage @link by Stripe for payments We're giving it $10,000 to make agentic payments, join the stream to spend our money.0360158
- We are rolling out 2fa-confirmation or email-re-login to many sensitive actions, such as reading environment variables, in the Vercel dashboard. Consider enrolling a passkey today to make that a quick fingerprint sensor thing.13211111.7K
- Malte UblRT @vercel_dev: Manage spend from the terminal. Set a team on-demand budget, pause production deployments at 100%, get notified on email or a webhook, and review your alert history. ▲ ~/ vercel budgets set030122
- The recent node.js RCE related to brotli compression is broader than just node.js and applies to essentially all software using brotli[1] Vercel’s firewall protection against the attack is effective independent of the backend technology. Python is vulnerable. However, unlike Node.js it requires using the pypi/brotli package explicitly[2]. It is likely that other software is impacted. Attacks can happen both through decoding incoming HTTP requests (which the Vercel Firewall protects against) as well as downloading attacker-controller payloads via HTTP. [1] [2]Vercel Developers @vercel_devNode.js Security Advisory: a high-severity RCE flaw in Node's Brotli decoder affects Node.js 22, 24, and 26. Vercel WAF rules are live for every project and block known exploit patterns.201716219.9K
- Node.js Security Advisory: a high-severity RCE flaw in Node's Brotli decoder affects Node.js 22, 24, and 26. Vercel WAF rules are live for every project and block known exploit patterns.743122.8K613K
- I'd love the industry to pick up this pattern for agent permissions: Combine planning with permission requests. Agent thinks through what permissions it will need and asks the user once with context – rather than asking one-by-one as it encounters permission prompts.Vercel Developers @vercel_dev"Fix this deployment" shouldn't mean "here are the keys to everything." We designed Vercel Agent to investigate with read-only access, then propose a plan with the permissions for the fix.231485.2K
- "Fix this deployment" shouldn't mean "here are the keys to everything." We designed Vercel Agent to investigate with read-only access, then propose a plan with the permissions for the fix.63749.1K
- Malte UblRT @cramforce: Given the state of AI we must internalize: Everything hackable will get hacked. Two forces are at work: ▪︎ Near-frontier open weight models hack without refusal ▪︎ Frontier models help you defend yourself How we're approaching this new reality ↓036068.4K
- One of the largest barriers to OIDC adoption is that setting it up can have awkward UX. And every provider has to special-case every other provider to optimize the UX of linking connections between them leading to NxM work. This is why we've invested in shareable UI components for setting up OIDC connections. Thanks @shadcn!shadcn @shadcnSpeaking of security, we created a registry for building OIDC configuration components. Pick your providers, form and validation libraries, and token verifier (powered by jose). Then run one command to copy everything into your app. I've built these a few times before. I know how annoying they can be to get right. Hope this helps.9615715.1K
Speaking of security, we created a registry for building OIDC configuration components. Pick your providers, form and validation libraries, and token verifier (powered by jose). Then run one command to copy everything into your app. I've built these a few times before. I know how annoying they can be to get right. Hope this helps.46561.1K73K- I've been cooking on this post for a bit, but with the cyber security events of this week I had to ship it: Long-live static credentials must die. We never want to rotate API keys again. Turns out only 8% of cloud and SaaS vendors support secure alternative. This is our call to action to change this AND our commitment to do our part: - We are requiring vendors on our marketplace to support OIDC - We shipped @shadcn components to make it easier to ship OIDC configuration - We are expanding OIDC support on our own platform The last generation got the web from HTTP to HTTPS. It is on us to move past API keys.699295771.4K
- Show more posts
