MQTT brokers are often free. But clustering isn’t.
That’s where the lock-in starts: high availability, failover, and multi-site deployment get tied to a vendor’s enterprise tier.
@nats_io works differently.
Announcing Wanix 0.4-preview and introducing Wanix Elements! Now you can build virtual compute systems with just HTML. Some broad use cases:
- run AI in the strongest client-side sandbox
- let people try your CLI before downloading
- build your own browser OS or dev env
- make old apps/games work on the web
- explore and play with plan 9 concepts
Link below.
There is another idea that continuously return in my head recently: we should not stop writing code by hand. But this code will be like poetry. Small programs that are outstanding either because of style or new ideas introduced. Code you may print and frame on your wall.
Jens @axboe singlehandedly saved Linux I/O…
(and stopped me from moving to FreeBSD for kqueue)
Come hang with the Lord of the Rings at #systemsdistributed
Turning on auth in NATS takes one line.
But managing identity, permissions, & key rotation across a fleet, without rolling a single server?
That's where your @NATS_io security posture is defined
See how to go from a wide-open server to decentralized, dynamic auth that scales
Can agents build production-grade infrastructure?
At @resonatehqio, we use deterministic simulation testing to enable agents to design and implement production-grade durable execution on @nats_io.
My @aiDotEngineer talk is live. Link in replies.
Connecting a single edge device or site to your NATS hub takes five minutes.
Designing a fleet that still behaves when five devices become five million... that's where things get interesting.
@antirez@ziglang I agree with you @antirez - and I think the folks over there will adopt in their own time and in their own way. Love your content btw!
The most interesting part here is that folks can disagree and still respect and support one another. I’m a big believer in AI as well but also support @ziglang and Andrew. This world needs more acceptance of differing views.
My family is donating another $400,000 to the Zig Software Foundation. Zig is exceptional software. I use AI every day. Zig has one of the strongest anti-AI policies in open source. We disagree on some things, but respect doesn’t require agreement. https://t.co/gaOuwjDDtm
The team at @synadia continues on its roll around Agentic AI.
Since release we widened the agent fleet, hardened the foundation, and shipped a whole new learning repo. What landed:
3 agent channels are now shipping: Flue, OpenCode, and DeerFlow. A 4th, Codex, is feature-complete and in review.
Examples reached full TypeScript + Python parity, plus a from-scratch deep-research agent.
A completely new public repo, synadia-agents-labs that is a hands-on, runnable, heavily-commented examples that teach the Agent Protocol (5-line echo agent → LLM agents that call tools, remember things, reach edge sensors), in both TS and Python, running on your laptop (local Ollama or hosted OpenRouter, no cloud needed).
It's structured as a course: self-paced examples that progress in complexity, plus a guided 10-step workshop. It is the getting-started material the main repo doesn't carry.
https://t.co/9TpzPCQAYg
As the software engineering space continues to evolve at blistering pace, I wanted to reflect on who I looked up to as the best programmers from my career.
Here is my list.
1. Fabrice Bellard
2. John Carmack
3. Dave Cutler
4. Justine Tunney
5. Sanjay Ghemawat
6. Jeff Dean
7. Fedor Pickus
Comment and post your list.