Randall Hunt

A week ago on #TheContext, Toby Padilla from GitHub talked about his view that opening up MCP publication on the GitHub registry will spur adoption. Very good episode.

9

This is genuinely good. Anthropic is cooking. Introducing Cowork: Claude Code for the rest of your work. Cowork lets you complete non-technical tasks much like how developers use Claude Code. In Cowork, you give Claude access to a folder on your computer. Claude can then read, edit, or create files in that folder. Try it to create a spreadsheet from a pile of screenshots, or produce a first draft from scattered notes. Once you've set a task, Claude makes a plan and steadily completes it, looping you in along the way. Claude will ask before taking any significant actions so you can course-correct as needed. Claude can use your existing connectors, which link Claude to external information. You can also pair Cowork with Claude in Chrome for tasks that need browser access. Cowork is available as a research preview for Claude Max subscribers in the macOS app. Click on “Cowork” in the sidebar: claude.com/download

11

3 Comments

Lambda Durable Functions were developed to address the need for pausing a Lambda execution while allowing external work, such as an LLM generating a response, to continue without incurring idle VM costs or losing execution state. This challenge is not new. It’s not a “which is better” post it walks through the same real use case implemented both ways and looks at where complexity and cost actually live. https://lnkd.in/dBDdk94z #AWS #Serverless #CloudArchitecture #StepFunctions #AI

28

There's been some talk around storage unification and zero-copy lakehouse integrations recently, and I wanted to better define these terms as well as look at ways we should evaluate different design decisions in this space. So I’ve published a new post: A Conceptual Model for Storage Unification Storage unification isn't about “one storage format to rule them all,” but virtualizing them into a unified logical model. In this post, I lay out a conceptual framework for thinking about storage unification (aka data virtualization). I explore: 🔹 Internal vs. shared tiering (and why shared tiering is so tricky). 🔹 When materialization makes more sense. 🔹 Client vs. server-side stitching. 🔹 Lifecycle and schema management across evolving formats. 🔹 Seven key considerations for evaluating different approaches. This is the groundwork for a set of posts where I’ll dig into how some real-world systems deliver storage unification, primarily about unifying real-time and historical data. https://lnkd.in/dBiS2kxJ

45

Building reliable workflows across microservices, functions, and stream processors is one of the hardest problems in modern software architecture. Over the past few months, I’ve been thinking deeply about how systems make progress reliably in the face of partial failures, service boundaries, retries, and complex dependencies. I kept coming back to the same question: 👉 Can we construct a simple but powerful mental model to reason about coordination, durability, and reliability across different architectural styles? The result is “Coordinated Progress”, a 4-part series that breaks down: 🔹 Part 1 – The Graph: A systems view of modern architectures as graphs of computation, with workflows as subgraphs that must coordinate across time and failure domains. 🔹 Part 2 – Making Progress Reliable: A model built around two core ideas: Reliable Triggers and Progressable Work. If you get these right, you have a shot at resilient distributed execution. 🔹 Part 3 – Coupling, Synchrony, and Complexity: Design-time and runtime coupling, synchronous vs asynchronous edges, and how orchestration and choreography shape the behavior and complexity of your systems. 🔹 Part 4 – A Loose Decision Framework: Brings the series together with a decision-making lens. Applying the model to help make more reasoned and effective architectural decisions. Durability is the common thread, not just of data, but of progress itself. If you’ve ever struggled to explain why coordination matters, or how orchestration, event-driven architecture, queues, topics, RPC, and stream processing all fit together then this series might help. 🧩 Start here: Part 1: https://lnkd.in/d6PCVvhV Would love to hear what resonates (or doesn’t) with your experience.

115

10 Comments

Maintaining an open source project is hard. Most maintainers are managing globally distributed contributors who are largely working for free, while companies build products and profit on top of their work. At the same time, those same companies expect bug fixes, features, and security updates on timelines that work for them, not for us. When something goes wrong, the blame usually lands on the maintainers. If your business relies on open source, helping maintain it isn’t optional. It isn’t charity. It’s part of a secure software supply chain. The health of these projects doesn’t just affect maintainers, it affects every company and product that depends on them. I spend a lot of my time at work contributing upstream as a Kubernetes maintainer and supporting the broader ecosystem. I’m very lucky to work at Minimus, a company that recognizes the importance of that work and considers it a large part of my job as a Developer Advocate. It’s so rare to see that in a smaller company. I wrote more about this here: https://lnkd.in/geZBAVXV

211

7 Comments

Open Responses is an open-source spec based off OpenAI's Responses API, with backing so far from OpenRouter, Vercel, HuggingFace, LM Studio, Ollama, and vLLM. https://openresponses.org Will more providers get on board? It'd be so nice to have a true standard for LLM APIs.

69

5 Comments

It turns out that it is very easy to connect aider to ScalarLM to have a code-cli experience on your own LLMs. Someone is going to figure out how to RL train on their own code.

5

PostgreSQL’s upcoming out‑of‑cycle release (https://lnkd.in/gEu3jsSJ) is a reminder of something every engineer and engineering leader knows but rarely says out loud: operational maturity isn’t defined by the absence of issues, but by how quickly and transparently you respond to them. When a core piece of infrastructure like Postgres schedules an unscheduled release, it’s not just a patch—it’s a signal. * A signal to review your dependency surface area. * A signal to validate your upgrade pathways. * A signal to check whether your teams can respond to change without chaos. If your organization treats these moments as routine rather than disruptive, you’re already ahead. Oh by the way, if you have PgSQL self-managed (on prem or any public cloud), make sure you check this patch out on 26th Feb. Substring() may not be massive, but there's Secondary related fix which is important. #engineering #culture #postgresql

4

𝗧𝗵𝗲 𝗵𝗶𝗱𝗱𝗲𝗻 𝗽𝗿𝗼𝗱𝘂𝗰𝘁𝗶𝘃𝗶𝘁𝘆 𝗸𝗶𝗹𝗹𝗲𝗿 𝗻𝗼 𝗼𝗻𝗲 𝘁𝗮𝗹𝗸𝘀 𝗮𝗯𝗼𝘂𝘁: 𝘀𝗹𝗼𝘄 𝗮𝗿𝘁𝗶𝗳𝗮𝗰𝘁 𝗱𝗼𝘄𝗻𝗹𝗼𝗮𝗱𝘀 I've been talking to development teams across APAC, and the same frustration keeps coming up. Developers are spending way too much time waiting for Docker images, npm packages, and dependencies to download. What should take seconds is taking minutes, or worse. Here's what's really happening: - CI/CD pipelines crawling at a snail's pace - Cloud egress fees quietly eating budgets - Teams managing multiple solutions that don't talk to each other - Developers losing momentum every time they hit "build" The thing is, this isn't just a technical problem, it's 𝗮 𝗯𝘂𝘀𝗶𝗻𝗲𝘀𝘀 𝗽𝗿𝗼𝗯𝗹𝗲𝗺. When your best engineers are staring at loading screens instead of solving real challenges, that's innovation time lost forever. At Varnish Software we have seen customers cut their build times by 80% and slash cloud costs with intelligent caching. The solution isn't replacing everything you have, it's adding a smart layer that makes everything work more efficiently. Want to see how? My colleagues Guillaume Quintard and Adrian Herrera are hosting a webinar on universal artifact acceleration. Real strategies, real results. Because life's too short to watch progress bars.

8

1 Comment

Lydian has been, and continues to be, an Anthropic shop. Claude model variants and Claude Code are what we trust with our internal development efforts. That's even more true today than it was yesterday. We will continue to lean into that ecosystem - not only because it has consistently performed the best but, in addition, because Anthropic appears to be the only company treating the race to AGI like a functional adult.

22

1 Comment

Yasmin COO at StormForge discusses how teams often copy and paste resource configurations from existing applications without considering if they match their needs. She explains how resource management becomes an afterthought in the deployment process, with teams prioritizing speed over the proper configuration of CPU and Memory requests and limits in Kubernetes. Watch the full interview: https://ku.bz/Bs49B8sYG This interview is a reaction to Alexandre's episode https://ku.bz/z2Vj9PBYh

15

As LLMs continue to show up across the federal space, one thing matters more than model size or hype: control. AWS quietly made an important move by adding response streaming support to Amazon API Gateway, and it directly addresses a challenge many of us hit when deploying AI in the federal arena. Streaming responses mean APIs no longer have to wait for a full LLM output before responding. Users can see results as they are generated, which improves perceived performance and reduces impatience waiting for a response. With streaming, teams gain: - More control over how quickly responses are delivered, including intentionally slower or paced outputs when overuse is detected. - Better user experience without bypassing security, logging, or gateway controls that come from workarounds. - Support teams that want to stay serveless and keep their architecture simple. The ability to shape how an AI responds is just as important as what it responds with. Serveless infrastructure is finally catching up to how LLMs are actually being used in production. https://lnkd.in/gxe_tkrp

9

4 Comments

AI is no longer just for developers. Tools like Cowork are automating everyday, non-technical work. Developers get the edge by focusing on architecture, logic, and real problem-solving. https://lnkd.in/g4aNhDDe

11

Availity's build and release team plans to use a new integration between GitLab Duo and #AmazonQDeveloper #AIagents to help with code modernization and risk mitigation. #GitLabDuo #AWS Many thanks to Elizabeth Dobelstein for a great interview this week! https://lnkd.in/eg7bqVMC

6

1 Comment

Keeping your #software supply chain safe starts with what you bring in. JFrog Curation helps teams automatically vet open-source packages before they enter your environment—catching malicious, vulnerable, or non-compliant components early. The result? Cleaner builds, fewer surprises, and more confidence in every release. #OpenSource

2