Oxide goes to great lengths to allow you to own your servers and operate them in an airgapped environment. Could a tractor be built to operate airgapped even with onboard tech? Or to be able to connect to a local base station over e.g. LoRaWAN instead of the cloud?

If you're responding to an implication that people should go have their own nice place elsewhere and not bother me in my nice place... yeah, I can understand why that's problematic. But I don't think the position "we should build more nice places" is itself a wrong position to take.

There will of course be some places that are uniquely popular e.g. due to their geography or natural beauty. Humans tend to congregate around centres of economic activity, which means some places become popular. But creating sprawls of dormitory suburbs and efficiently piping everybody into the few places that are actually nice sounds terrible.

* my perspective from Australia

Self promo: I wrote a tiny post about an interesting technical detail of Thief's game engine - the world is actually solid, and gameplay areas are carved out of it like caves.

https://crabmusket.net/2025/the-solid-universe-of-thief-the-...

More here:

https://nothings.org/gamedev/thief_rendering.html#csg

> The client provider requires an API token to fetch flag values. This token is not scoped to a single app, so anyone with the token can evaluate flags across all apps in your account. Use the client provider with caution in public-facing applications.

https://developers.cloudflare.com/flagship/sdk/client-provid...

Can anyone clarify... why the client SDK, designed to be deployed to browsers, requires caution? Does this mean that any client could send requests with a new targetingKey and observe other users' flags?

While flags probably shouldn't be critical information, this seems like an interesting design choice.

I'm fairly AI-skeptical not on grounds of "do they work" but "are they good for the world". I feel that getting AIs to do this kind of review work is a rare case that doesn't outsource thinking and deskill workers. It doesn't trigger the same alarm bells as having the AI write the code (including having the AI fix the issues it discovers). That's setting aside environmental and other ethical concerns, which are still significant to me.

I have been impressed by the recent quality of AI code reviews*, but the experience of interacting with 3 separate AI reviewers via GitHub PRs is pretty terrible. Having more local-oriented and jj/rebase-aware review rounds would be great.

*context: fairly large PHP/Laravel backend and Vue frontend

[1]: https://milvus.io/blog/ai-code-review-gets-better-when-model...

[2]: https://github.com/liliu-z/magpie

[3]: https://blog.cloudflare.com/ai-code-review/

Can you clarify how you're spending tokens on waiting? My understanding is that the LLM isn't actually necessarily doing anything while a build runs. The whole process end to end may take longer for sure (ignoring things like the compiler catching more errors, that's really hard to factor in) but how does that correlate to more tokens?

I'm trying to work out why it feels bad to trust a private company with this kind of information, whereas "we" are happy to trust AWS with our servers, Hashicorp with our Vaults, etc.

But these businesses seem to rely on some amount of scale for their trustworthiness. Password managers seem like a cottage industry in comparison, especially as lots of their users will just be "normies" and even ones on a free tier, because ~nobody thinks they should pay for a password manager?

> I don't trust Private Equity or the Harvard MBA mindset to be allowed to hold on to my passwords.

I agree, but you have credible exit. As annoying as it is, it seems quite feasible to continuously migrate to the next provider who is currently in their "don't be evil" phase.

Someone on lobste.rs suggested there should be a worker-owned co-op for password managers. This fits my personal bias, but I wonder if it would be any more resistant to this failure mode? Co-ops can be bought out also, and depend on strong leadership to prevent this.

Maybe a customer-owned co-op instead of a worker-owned one could make it more impractical to buy out. Or a foundation model like Signal, Wikipedia etc.

EDIT: I'm reminded of https://fleetdm.com/ business model, which is heavily open source yet paid. That seems like essentially what Bitwarden was? And presumably Fleet is not protected from the same outcome, no matter how inspiring their example is right now.

When introducing new people to it, I like the topics to email threads, as they also have a subject line and replies.

It also gels with the "inbox" part of Zulip's UI.

As a huge fan of Zulip the app and the team behind it, I have intensely mixed feelings about the AI-ness of it all. But this does seem to be the most responsible way forward.

Zulip needed to be able to outlast its founders to be a truly sustainable project. The way they've focused on building up their contribution pipeline, the effort they spend on mentoring new developers, it has all built towards that being possible.

https://blog.zulip.com/2021/12/17/why-zulip-will-stand-the-t...

It seems like just yesterday that the core team started experimenting with using Claude to work on Zulip, which maybe adds to the surprise of this announcement. But I don't begrudge those individuals their choices. Ten+ years is a long time to work on any project.

https://blog.zulip.com/2025/11/24/zulip-ten-years/

Here's to the Zulip project continuing to maintain its engineering excellence and its community principles for the next ten years.

The examples of benefits, I don't disagree with, and the author has chosen examples that do align with cultural socially left-wing concerns (disability, class). But saying "code switching to appear PMC works" is the problem, not the solution. If you don't think institutions can adapt to that reality to protect themselves from peasants wielding LLMs, I think your analysis of power is missing.

The examples in the piece are just more examples of how AI can be used to paper over societal issues instead of addressing any of the root causes.

(And I should say that on the flipside, AI is not the cause of existing social problems, but a symptom of them.)