Dev Systems

How durable sessions unify human-to-human and human-to-agent messages

AI chats are often a rather solitary experience: just you and ChatGPT, sitting there together, solving a problem. But so many of the tasks that we perform day to day are ones that benefit from, or often even require, collaboration with other people such as colleagues, family members, or friends.So, if AI agents are helpful, and other people are helpful, then how can we provide a space for multiple people to collaborate with each other and with AI agents?This is a question to which the flagship A

Is AI making your teams better, or just busier?

AI adoption programs tend to end in the same place. Tools are accessible, usage is up, and there's a dedicated Slack channel for wins. Six months later, nothing about how the team works has fundamentally changed. People are doing the same things – just slightly faster. And it’s easy for programs to stall when you’re measuring the wrong thing. Adoption (whether people have access and whether they're using the tools) is visible and easy to report. It tells you nothing about whether the team i

One person, one vote: building live voting with message annotations

Live polls are a staple of conferences, streams, and all-hands: a question goes up on the big screen, everyone votes from their phone, and the bars race each other in realtime.There's a lot of different ways you could implement this.The most obvious way is a CRUD app backed by a server that votes are POSTed to, and the server keeps a running count. But this is the Ably blog, so we're obviously going to use Ably Pub/Sub to build something which does not need each vote to go via your server. That

Privacy-Aware Infrastructure in the AI-Native Era: An Asset Classification Case Study

 Privacy controls — systems that enforce retention, access, allowed-purpose, downstream-sharing, or anonymization policies — require a reliable understanding of data to function. Before such a control can operate effectively, it must know exactly what it is looking at. This can be complex, as demonstrated by a field simply named “age“: In one context, it might describe a person and require strict protections, while in another, it could be a cache time-to-live (TTL) numerical value in

Show HN: eBook to audiobook narration with realistic AI voices

For a while I've wanted to try out the new AI voices for long-form narration, but everything I found required a subscription that didn't justify my limited usage. I came across the open Kokoro model [0] and the voices are very good -- good enough to listen to for hours without the fatigue I got from legacy, robotic TTS voices. The model is 82m parameters and designed to run fast, but I still struggled to get reasonable times from CPU inference on my 12-core laptop. I thought a cloud-ba

Show HN: TLA+ Process Studio

Disclaimer: This was made with LLMs.I made this tool to help understand large business processes that can be modelled as a single state machine.The core loop of this is to enable to walk stakeholders through discussing each step, adding comments, and reiterating with an LLM of their choice to generate the TLA+ syntax on the left.Users can click through the green state nodes to see how things work visually.You can see some sample state machines in the dropdown in the top left.The power would come

A 30 Year OG Application Developer Available

https://www.youtube.com/watch?v=DACtpW9Q-hcThat link is the #1 Architectural Interior Design software used by the top firms in the world. What you are looking at is the result of pure architectural discipline. To handle massive global budgets, complex subcontractor workflows, and real-time synchronization with massive enterprise CAD suites, I engineered a closed-loop, self-aware data object model. The data objects carry their own application logic and database schemas. It is a zer

Ask HN: What's your go-to queue system?

Hello,I am building a new product and need to have a robust queue system but I'm not sure what product to choose.I've worked in Amazon before and usually AWS tools are the default go-to but in my own time for small project I've used things like: https://github.com/hibiken/asynq for basic tasks.My concern with the above is that it's still in "early" development; from their README:``` Status: The library relatively stable and is currently undergoin

Your Vercel AI SDK app is missing a session layer

If you have built an AI chat feature with the Vercel AI SDK, you have used its useChat hook. You give it your messages, and it streams the reply into your UI.You may have seen our post on the custom transport we built for the Vercel AI SDK. It swaps useChat's default transport for Ably AI Transport, adding resumable streams, cross-device and multi-user sync, conversation branching, history compaction, and stop-and-approve controls.This post tours a demo built on exactly

Stop vs disconnect - why canceling AI streaming is harder than it looks

You add a stop button to your AI chat app: a customer support agent, a coding assistant, a research tool the user can steer mid-task. A user clicks it mid-response. The frontend stops rendering. Then you check your backend logs and realize the underlying generation is still running, and you’re still paying for every token.This is not a bug. The Vercel AI SDK docs document it explicitly: in a resumable stream setup, calling stop() only closes the current HTTP connection and should not cancel the

Agentic apps that go beyond chat

You are planning a trip with an AI assistant on your laptop. You are chatting with the agent, and as you progress it is dropping pins on a map, building a day-by-day itinerary, adding up a budget, and streaming its reasoning as it goes. The state of your interactive session is a combination of the chat history, the synthetic UI constructed by the agent during that process, and structured state, the itinerary, arising from the decisions you each make.Building such an app has challenges beyond get

Show HN: Supaqueue – Node.js background job queue (no Redis needed)

Hi everyone,I have been using BullMq for most of my background job related work but lately I have been working on some smaller scale app where I having a full blown Redis setup with separate worker process would have been overkill.That is why I built a lightweight, in-memory Node.js background job queue. It comes with a Bull/BullMq-type API, concurrency control, schedulers, job retention and much more. It has zero dependencies and is fully typesafe. Use this when you need a simple, performa

Ask HN: Switching from backend development to graphics programming

I love computers. I wrote my first program in Borland C++ when I was 11.By chance, I managed at my 15 to get a job where I did some HTML pages and later some PHP programming. Making websites wasn't as fun as making games in C++. Overall PHP didn't seem as fun as C++. I made all of my lab projects at the university in C++ with Qt, or wxWidgets or bare WinAPI. In fact, I improved the university the internal testing system on our Math/Physics faculty. I got a privilege to do la

How to become an AI infrastructure engineer?

Hi,I currently work on a GenAI platform for one of the largest local industrial companies. My daily work mostly involves building inference infrastructure on top of a 48x H200 GPU, Kubernetes and vLLM. Hence, I'd say it's 80% SRE and 20% software engineering when it comes to building request routing and internal control planes.Although I have a background in backend engineering rather than ML research or low-level GPU programming, I am trying to understand what I need to learn to becom

Introducing AI Transport v0.3.0

Last week we introduced AI Transport v0.2.0 and made one idea the centre of the design: the session is the channel. Every input, output, and lifecycle event for an AI conversation is just a message published to an Ably channel, which is what makes a session durable, multi-party, and resumable. In v0.3.0, we added first-class support for presence and LiveObjects to AI sessions, allowing you and your agent to see who's online and update shared state in real time.We made the codec interface declara

How Meta Engineered Ultra-Narrow Batteries for AI Glasses

Smart glasses like the Ray-Ban Meta and Oakley Meta Vanguards need to pack enough energy to power features like cameras, speakers, AI workloads, and even a display. But it all has to fit into the glasses’ temple arms.So how do you place a battery with enough power to run a pair of smart glasses all day into a form factor narrower than an adult’s pinky finger? You have to rethink how batteries are made. In episode 86 of the Meta Tech Podcast, host Pascal Hartig sat down with Karthik and Myuran, t

Nobody trusted our internal dashboards. Now they live in code

How we used AI to fix a data trust problem, and built a governed reporting system the whole company can contribute to.We audited our skills library a few months ago and found twelve dashboards hiding in it.Not dashboards. Skills that built dashboards. Someone needed a view of some data, asked Claude to put it together, got a long HTML page out of it, and then wrapped the whole thing in a skill so others could run it again. Twelve times over, by different people, for different questions.This is w

See Anthropic Orchestrate the Narrative

tl;dr FOSS is the biggest threat to the largest new economic sector, so everything that economic sector does should be viewed through the lens of trying to kill it.I occasionally see articles and sentiments along the lines of, Anthropic is or is not, "scare mongering to boost the perceived cultural impact of their AI/ML tools; a sort of underhanded advertisement".If your job is to defend Anthropic online, it's a good angle to fight from. It's a viral subject with no prac

Toward More Controllable AI Video Editing: An Early Research Exploration at Netflix

By Zhuoning Yuan, Ta-Ying Cheng, Benjamin Klein, Bahareh AzarnoushIntroductionAt Netflix, we build technology to help storytellers bring their creative visions to life and to help members discover the stories they love.To connect stories with diverse audiences around the world, we produce promotional assets, including trailers, teasers, and social short‑form videos, that build on and elevate the original footage. Through close collaboration with the teams crafting these assets, we identified a r

How Netflix Simplified Batch Compute with Kueue

By Alvin Bao, Alex Petrov, Jennifer Lai, Aidan Sherr, and Samartha ChandrashekarAs a part of the journey to transition Netflix’s compute infrastructure to be more Kubernetes-native, we have leaned into incorporating components from the Kubernetes ecosystem into our container platform Titus. One example of this is our use of Kueue, a cloud-native job queueing system for batch workloads, which has largely replaced the custom queuing and scheduling logic in our homegrown managed batch solution Comp