Sumanth Chinthagunta (సుమంత్ చింతగుంట)

Announcing 🎀 𝒮𝓅𝑒𝒸𝓉𝒶𝒸𝓊𝓁𝒶𝓇 🎀 a full-stack @turborepo template project with 𝐒𝐯𝐞𝐥𝐭𝐞𝐊𝐢𝐭 meta-framework, @tailwindcss @SkeletonUI #GraphQL #Houdini @HasuraHQ #superforms with 𝐿𝒾𝓋𝑒 𝒟𝑒𝓂𝑜𝓈 @SvelteSociety @vercel github.com/xmlking/specta…

@anthonysheww @workspace @Repo @shadcn github.com/shadcn-ui/ui/t…

@xmlking @workspace @Repo @shadcn Makes no difference!

Turborepo templates updated to 16.2. Enjoy! 𝚗𝚙𝚡 𝚌𝚛𝚎𝚊𝚝𝚎-𝚝𝚞𝚛𝚋𝚘@𝚕𝚊𝚝𝚎𝚜𝚝

@anthonysheww @workspace @Repo @shadcn Technically you are right. But "@workspace/ui" feels like local repo them other

Anyone creating @voidzerodev / viteplus monorepo starter template with @tan_stack start and @shadcn ? @haydenbleasel

anyone had success migrating to #vinext and @voidzerodev for existing nextjs projects ?

@livingdevops similar to KinD (Kubernetes IN Docker) but with policies ? kind.sigs.k8s.io

🔥Breaking: NVIDIA just open-sourced the guardrails AI agents should have had from day one. It’s called OpenShell. Announced at GTC yesterday. Your coding agent currently has access to your terminal, files, AWS keys, and network. OpenShell fixes that. What it does: - Filesystem locked at sandbox creation - Network blocked by default. - You whitelist what’s allowed - API keys never touch the filesystem. Injected at runtime only - Policies defined in simple YAML One command to sandbox Claude Code, Codex, or Cursor. The architecture runs a full K3s cluster inside a single Docker container. No separate Kubernetes install. Adobe, Atlassian, Cisco, CrowdStrike, Salesforce are already integrating it. Most teams solve agent security at the application layer. OpenShell solves it at the infrastructure layer. GitHub repo link in comments.

@shadcn No more tsconfg alias ?

We added node subpath import support in the latest canary. Long-requested. If you have a project using package.json imports, help us test: npx shadcn@canary init

@matteocollina @nodejs Will it work like FUSE ?

.@nodejs has always been about I/O. Streams, buffers, sockets, files. But there's a gap that has bugged me for years: you can't virtualize the filesystem. You can't import a module that only exists in memory. You can't bundle assets into a Single Executable without patching half the standard library. That changes now 👇

@matteocollina @nodejs This could be useful for managing AI agents’s context with node native virtual file system @mastra !

@voidzerodev @bunjavascript please make it work

@voidzerodev do you have a template if I want to create a workspace with both monorepo and @tanstack/start or nextjs ? vp create vite:monorepo vp create @tanstack/start

@branchseer @voidzerodev would be nice to have dotenv files loaded automatically for monorepos (load root .env then apps/web/.env) by dev tools. Since environment variables are infra related and .env is predominately used for local developer only, It should not be framework or lib responsibility.

@xmlking @voidzerodev No. It should be loaded inside the task. (If it is, vite-task will fingerprint .env in the cache.)

@voidzerodev Do vite-task support loading .env into context for tasks ?

@TheAlexLichter @vite_js I need monorepo template with react/nextjs so that I can switch over from turborepo to vite+

Can't wait until the @vite_js community discovers the stunning new starter templates 🔥

@tempoimmaterial @Steve8708 Why @bunjavascript is not in the #vp echo system ?

If you don't understand this concept, “eventually” you will be in trouble. How to achieve Exactly-Once delivery? One of the most serious problems a distributed system can have is executing an operation many times due to retries. At first glance, Exactly-Once delivery seems hard to tackle, but dividing the problem into two parts makes it much easier to solve. Mathematically, an operation is executed Exactly-Once if: - It is executed At-Least-Once. - It is executed At-Most-Once. At-Least-Once Delivery An operation is executed at least once if, even in the presence of failures, the system retries it until it succeeds. How? 1. Retries: Implement a retry mechanism to ensure the message is retried until it is successfully processed. (Take into account that not all operations are retryable) 2. Persistence: Store the message and its state in a durable message queue to ensure it is not lost and can be retried. 3. Monitoring and Alerts: Set up monitoring to detect and alert if retries exceed a threshold, indicating potential issues needing attention. At-Most-Once Delivery An operation is executed at most once if the system prevents the same message from being processed more than once, even if it is retried. How? 1. Idempotency is your friend. An idempotent operation can be applied multiple times without changing the result beyond the initial application. 2. Generate a unique idempotency key for each message/operation. 3. Before processing the message, check if the idempotency key already exists in a persistent store: - If it exists, do not process the message again and return the previous result. - If it does not exist, process the message and store the key. Don't double-charge your customers; instead, combine At-Least-Once and At-Most-Once. ;)

@kozlovski What are the trade-offs between storing data in S3 and a database?

🚨 NEW: Diskless Topics were just accepted into Kafka! In case you've been living under a rock for the past 3 years, Diskless Topics are a new innovative topic type in Kafka where the broker writes the data directly to S3. It changes Kafka massively: • up to 90% lower costs vs classic Kafka due to no cross-zone replication. At 1 GB/s, we're literally talking ~$100k/year versus $1M/year 🔥 • stateless brokers. No disk to manage means no state on the broker, making them as easy to run as Nginx. 🏎️ • instant scalability & elasticity. Because this architecture is leaderless (every broker can be a leader), new brokers can be spun up, and traffic can be redirected effortlessly. Hot spots are no longer a thing. Just-in time scaling means you don't need to overprovision as much. 🙌 • network topology flexibility - you can scale per AZ (e.g more brokers in 1 AZ) to match your applications topology This all comes at one simple tradeoff - higher request latency (up to 2 seconds p99). I revisited the history of Diskless topics. Aiven was the first to do two very major things here, for which they deserve massive kudos: • 🏆 First to Open Source a solution, and commit to contributing it to mainline OSS Kafka • 🏆 First to have a product that supports both classic (fast, expensive) topics and diskless (slow, cheap) topics in the same cluster. The combination of these make open-source Apache Kafka VERY competitive to all the proprietary solutions that were first to market. Most importantly: • users can actually save 90%+ costs. Proprietary solutions ate up a lot of those cost savings as their own margins while still advertising to be "10x cheaper" 🤑 • users do not need to perform risky migrations to other clusters • users do not need to split their streaming estate across clusters for access to diskless topics • adoption will be a simple upgrade and setting topic.type=diskless The Diskless initiative by Aiven, frankly said, really raised the bar. Just two days after the KIP announcement in 2025, AutoMQ changed their open-source license to Apache and began donating proprietary features. Diskless topics promise to open up the gate for better data lake (Parquet/Iceberg) integration as they offer an S3-first write path in Kafka. It's one of the few exciting things left in Kafka nowadays, especially post Confluent being acquired by IBM.. In this week where I bet Aiven is popping champagne, let us congratulate them for keeping open-source alive & well. 🥂

YES! Someone reverse-engineered Apple's Neural Engine and trained a neural network on it. Apple never allowed this. ANE is inference-only. No public API, no docs. They cracked it open anyway. Why it matters: • M4 ANE = 6.6 TFLOPS/W vs 0.08 for an A100 (80× more efficient) • "38 TOPS" is a lie - real throughput is 19 TFLOPS FP16 • Your Mac mini has this chip sitting mostly idle Translation: local AI inference that's faster AND uses almost no power. Still early research but the door is now open. → github.com/maderix/ANE #AI #MachineLearning #AppleSilicon #LocalAI #OpenSource #ANE #CoreML #AppleSilicon #NPU #KCORES

@ctatedev @turborepo should also enable https for local development

portless v0.4.0 is out. Local dev servers now get HTTPS + HTTP/2 automatically: 𝚙𝚘𝚛𝚝𝚕𝚎𝚜𝚜 𝚖𝚢𝚊𝚙𝚙 𝚗𝚎𝚡𝚝 𝚍𝚎𝚟 -> 𝚑𝚝𝚝𝚙𝚜://𝚖𝚢𝚊𝚙𝚙.𝚕𝚘𝚌𝚊𝚕𝚑𝚘𝚜𝚝 No cert setup. No port numbers. No config. github.com/vercel-labs/po…