williamstein

This recent long interview with the founder of NATS has some new information and deeper insight into what actually happened: https://changelog.com/podcast/641

It ends with the host saying “Rug pulls are not OK” and the NATS founder responding “No pay, not okay.”

Synadia might not be very good at encouraging community involvement in development and has perhaps not made an effort. I've watched most of the videos they've posted on youtube, and they seem very proud of keeping total control of the project. I don't think there's anything fundamental to the NATS software that discourages community involvement; in fact, their code is extremely well written and easy to read in a developer friendly language (Go). However, there is something about how the project is actively run that seems to discourage community involvement. I've coincidentally been publicly documenting [1] my own surprising frustrations with trying to get involved during the last few months. Many other open source projects I'm involved with, e.g., Jupyter, SageMath, Cython, have much more welcoming governance structures, with regular welcoming developer meetings open to everybody, etc.

All that said, NATS is Apache licensed software, and any company is legally allowed to create and maintain a closed-source fork based on this. CNCF claims the trademark "NATS" and the domain name nats.io.

I have often wondered if the name "NATS" is an overly confusing name. People often think it means "Network Address Translation". I suspect changing the name to something else that involved PubSub or Queue or DB would be a net win in the longrun.

[1] https://github.com/williamstein/nats-bugs

UPDATE: Synadia has agreed with the CNCF requirements. See the new joint statement: https://www.cncf.io/announcements/2025/05/01/cncf-and-synadi...

For some applications especially those that benefit from cpu-bound parallelism, it is possible to write much faster code with Go than Node.js, e.g., consider the recent port of the Typescript compiler from Javascript to Go:

- https://news.ycombinator.com/item?id=43332830

Context:

- https://news.ycombinator.com/item?id=43829008

- https://github.com/cncf/toc/issues/1632

- https://github.com/nats-io/nats-server/issues/6832

I also built a lot on NATS Jetstream, and I'm questioning my decision in light of the threatened license change.

I evaluated a range of options and selected NATS for a project in Jan 2025, then spent every waking moment developing and launching this project over the last four months. Absolutely central to my choice of NATS was that it is a CNCF project. You see, I was badly burned a few years ago but what happened with RethinkDB, which I used for a similar purpose!

I read substantial chunks of the NATS source code and contributed many small PR’s and suggestions back. Sometimes the devs were very welcoming and sometimes not. In fact, a few weeks ago I started this repo to document some contributions frustrations: https://github.com/williamstein/nats-bugs

Anyway this rug pull is extremely disappointing to me. It is especially sad because NATS is excellent technology, and no matter what their just shot themselves in the foot, and I had really really had hope for them to grow a lot in the next few years.

Fortunately, this has significantly improved since dedup was rewritten as part of the new ZFS 2.3 release. Search for zfs “fast dedup”.

And even dedup was finally rewritten to be significantly more memory efficient, as of the new 2.3 release of ZFS: https://github.com/openzfs/zfs/discussions/15896

Totally. And yet rigorous proof is very difficult. Having done some mathematics involving nontrivial proofs, I respect even more how difficult rigor is.

Not open source.

What is “the BIG ONE” in PNW that could happen without warning any second?

Last I checked, I think with nextjs, if you want to build once and deploy with different base paths (without having to rebuild), this is impossible, since the base path is hardcoded all over the build. This makes some forms of deployment of nextjs sites much more difficult than with other frameworks.

They have open sourced a lot of valuable code.

I wrote one https://wstein.org/books/modform/README.html

There was also a very interesting and popular HN post a few days ago from the ischool about how to better incorporate AI in teaching.

I was 6 years ago but in math.

What do you guys think of the UW information school, where this thesis is from?

I’ve never used wasmcloud, but they claim the following advantage over k8s on this page: “Leverage WebAssembly components to extend Kubernetes and distribute applications across clouds, regions, clusters, and edges.”

I think they use NATS.io to include “edge” services, which is something that is challenging to do in a single k8s cluster.

I don’t think they are trying to replace k8s, but instead are solving a slightly different problem, which is attempting to build something somewhat like k8s, but with very different networking and security constraints.

Docs here? https://github.com/vasanthv/ahey

And yet Facebook has 3 billion monthly active users: https://www.demandsage.com/facebook-statistics/

plot3d did work for me (on Chrome). Impressive.

As an academic, I always thought of arxiv as where you put your papers first, before they are peer reviewed. Before that we used our webpages, but they kept breaking.

The GitHub Readme seems easier to understand: https://github.com/apache/kvrocks?tab=readme-ov-file

This particular Oxide and Friends podcast includes a fantastic discussion of AI agents, the problems with their definitions, and at least one promising use case: https://oxide-and-friends.transistor.fm/episodes/predictions...

One example - Google Cloud network egress charges aren’t known until up to ~2 days after they happen. Since they can be obscenely expensive (eg $0.23/GB), they can make budget computation difficult.

I agree. The command line IPython by Fernando Perez was very inspired by Mathematica. He used Mathematica as a grad student and wanted a similar environment. In 2006-2007 Tom Boothby, Alex Clemesha and I wrote the first general used interactive web notebook called "The Sage Notebook", which became very popular with SageMath users over the years; the first version of Jupyter looked very similar to the Sage notebook. The Sage notebook was heavily inspired by everything around Google Wave and Google Docs (at the time), but definitely also by Mathematica's notebook. In particular, Alex Clemesha had recently been a physics undergrad and was a heavy Mathematica (and "Web Mathematica") user, and wanted a similar environment in a browser.

Thanks for explaining that background!

The Jupyter community maintains a public spec of the notebook file format [1], the kernel protocol [2], etc. I have been involved with many alternative Jupyter clients, and having these specs combined with a friendly and welcoming community is incredibly helpful!!!

[1] https://github.com/jupyter/nbformat

[2] https://jupyter-client.readthedocs.io/en/latest/messaging.ht...

No.