rodlette

If you just want to play with IPv6 on a VM, most VM providers will offer a /64, which is enough to have an address per service on your machine. If you wanted to play with IPv6 on multiple subnets, you'll need something larger than a /56, since subnets should be /64.

I rely on my home's v6 /56, so I don't have experience with using VMs for this, but I know of a few providers that offer /56 (and above):

* Mythic Beasts and Linode offer a /56 on request. They're not cheap VM providers though.

* https://ifog.ch/en/vps offer /48.

* https://tunnelbroker.net/ offer /48, which can be used via any VPS/home.

https://reddit.com/r/ipv6 for more info.

* IPv6. A genuinely useful tool, in particular for homelabs: multiple globally routable addresses per machine. One address per service. No need for Host/SNI vhosting. Works well with containers. To get v6 support, either find ISPs/SIMs that do v6, or wireguard to a VM that providss a /56.

* SSH ForcedCommand. Lots of usecases here, for backups, file storage, git, etc.

* Verilog as a tool for software developers to learn digital electronics. VCS/code/simulation/unit tests are all a lot more familiar and expected for developers.

* Writing tools yourself. There's often decent stable libraries that do 90% of what you want, and the remaining 10% is less effort than dealing with awkward integration with off-the-shelf tools. This relies on having low overhead packaging/deployment, e.g. Nix/Guix/Bazel.

OMEMO has replaced OTR. Conversations.im supports OMEMO.

Conversations.im and its developer produce great things for the community:

* Conversations.im itself

* https://snikket.org/open-source/ and https://quicksy.im/ are built on the conversations app

* Hosted XMPP: https://account.conversations.im/

* Compliance testing: https://compliance.conversations.im/

> Next I played with a bunch of hardware projects specifically targeting inexpensive Arty-A7 board to get comfortable with FPGA tooling.

What sorts of hardware projects?

Ask and ye shall receive: https://www.freedesktop.org/software/systemd/man/latest/syst...

I use it for the usecase you describe.

Great, and unexpected for Guix whose community and development seems heavily oriented to Emacs.

One day I hope to return to Guix, I've never had such high introspection ability as with Guix.

I am not an AI doomer, nor overly concerned with AI risk, but this post isn't a good rebuttal to the AI doomers. It mostly says "Nuh-uh, you're all being stupid".

The usecase of providing ethernet to devices without wireless makes sense.

I also hear about travel usecases for these, even for devices with wireless which I don't understand.

Some cite security, which I don't understand. Which device's security depends on the network to provide security? Are they weak on 4G?

Some cite avoiding captive portals. Are they that annoying that it's worth maintaining and carrying a travel router?

I trust this is just my lack of imagination. Please enlighten. :)

This is covered in "Information Theory, Inference, and Learning Algorithms" by David MacKay ( https://www.inference.org.uk/itprnn/book.pdf ):

> Why unify information theory and machine learning? Because they are two sides of the same coin. In the 1960s, a single field, cybernetics, was populated by information theorists, computer scientists, and neuroscientists, all studying common problems. Information theory and machine learning still belong together. Brains are the ultimate compression and communication systems. And the state-of-the-art algorithms for both data compression and error-correcting codes use the same tools as machine learning.

* In compression, gzip is predicting the next character. The model's prior is "contiguous characters will likely recur". This prior holds well for English text, but not for h264 data.

* In ML, learning a model is compressing the training data into a model + parameters.

It's not a damning indictment that current AI is just compression. What's damning is our belief that compression is a simpler/weaker problem.

I've ran greenfield at FAANG, and smaller companies.

Greenfield at smaller companies has more technical decisions to be made, since as you say, FAANG already has strongly preferred frameworks.

It's a shame, since making technical decisions is an important source of professional growth.

Though it's enabled by default, I consider the burden of proof to be to advocate for its utility. Would I opt into this if it was disabled by default?

It has some utility, and some costs: traffic noise, some privacy leak, system complexity, centralization.

On balance it's not for me so I disabled it.

See https://blog.cryptographyengineering.com/2019/10/13/dear-app... for a critique of the privacy of Safe Browsing.

Tangentially... I try to not use any third parties that say what I can/cannot see. That includes SafeBrowsing, which I disable via...

  browser.safebrowsing.blockedURIs.enabled
  browser.safebrowsing.downloads.enabled
  browser.safebrowsing.downloads.remote.enabled
  browser.safebrowsing.malware.enabled
  browser.safebrowsing.passwords.enabled
  browser.safebrowsing.phishing.enabled

I tried it. It looks nice, but audio/video calls didn't work at all between two FluffyChat clients.

I trust the developer too, but the developer isn't the only person you must trust when you accept wide ambient permissions. You must also trust the entire supply chain.

I see it as good hygiene to remove wide ambient permissions.

+1. I prefer to use uBlock Origin Lite, even on Firefox.

The ambient permissions that uBlock Origin requires on all sites is too risky.

This feels like more like a politically charged rant, rather than the type of output you'd hope from the Department of Computer Science at Cambridge.

> “Godfather of AI” Geoff Hinton, in recent public talks, explains that one of the greatest risks is not that chatbots will become super-intelligent, but that they will generate text that is super-persuasive without being intelligent, in the manner of Donald Trump or Boris Johnson.

> Perhaps this explains why the latest leader of the same government is so impressed by AI, and by billionaires promoting automated bullshit generators?

> Graeber observes that aspects of university education prepare young people to expect little more from life, training them to submit to bureaucratic processes, while writing reams of text that few will ever read. In the kind of education that produces a Boris Johnson

> AI systems like ChatGPT are trained with text from Twitter, Facebook, Reddit, and other huge archives of bullshit, alongside plenty of actual facts (including Wikipedia and text ripped off from professional writers).

Replying to myself... looks like some Matrix homeservers support OpenID-style login. That's probably fine for family.

I just checked https://matrix.org/ecosystem/clients/

I'll try FluffyChat, thanks!

I remembered another issue with Matrix.

Signal and XMPP (via Quickly) have a simple phone number based signup workflow that my family have grown used to.

My family are not happy on having to remember/use passwords/keys. That's a shame, but is ultimately a constraint I have to deal with when persuading them to install/use an IM app.

I've tried a few times. It always felt... clunky?

I tried Element. UI felt slow, I was unable to find notifications in scrollback. Clicking the notification button would take me to random messages.

They have devs and support engineers earning 700k, more than the CTO?

I'd prefer a federated solution, but XMPP doesn't yet have decent support for group chat that doesn't depend on being connected. https://xmpp.org/extensions/attic/xep-0369-0.1.html is still experimental.

Bravo to Signal for being easy enough for my family to use!

I just use plain DNS with plain Wireguard.

It's not as magic, but DNS can be made easy/automated with https://github.com/octodns/octodns or https://github.com/StackExchange/dnscontrol .

> Tailscale is the commercial tool to help setup and manage wireguard networks, and it had a big security incident earlier this year (though they were prompt to rectify).

> I don't use tailscale but I almost did. One of the things that caught me was not wanting to give a third party any control.

Yes, https://tailscale.com/security-bulletins/#ts-2022-004 was scary. Their response was great, but I'm reminded of Tim Cook's response to a question over what he'd do if it was in Zuckerberg's position (re a FB leak): "I wouldn't be in that situation".

Tailscale is a technical marvel, but it does a bit too much for my liking, for something that has the access it does... even more if you enable its SSH key support. I use plain Wireguard, but would like LAN discovery and STUN.

https://www.wirehub.org/ looks nice, as does https://git.deuxfleurs.fr/Deuxfleurs/wgautomesh which does LAN discovery too (as Tailscale does).

Agreed.

I don't think the PR reviewer is calling the author a liar. The reviewer said FreeBSD is not willing to lie to its users about version, and instead provided useful info, pointing out the relevant example.

Still, there's a lesson for reviewers that tone is important, especially over text and where a power differential exists.

Also, the author notes:

> The very first comment in the thread from a reviewer was simply only one word: “Why?” - That’s it, just one word. Why did I do this? Why work with such old hardware? Why is the sky blue? How the hell am I to know the context of what they’re referring to when asking a one-worded question?

Its clear from https://github.com/freebsd/freebsd-ports/pull/189#discussion... that the author is asking why is this line needed, something a PR author should always be able to justify.

Alternative: use the iperf CLI tool with one of the servers listed on https://github.com/R0GGER/public-iperf3-servers .

I prefer this, since I already use iperf to measure speed between hosts I manage.

I agree, it's not ready yet. What we see is a community forming, developing the software/protocols/skillsets needed to make Linux Mobile viable. It's expected that it's messy in the beginning. :)

Nice comprehensive and prompt response!

https://www.devever.net/~hl/xmpp-incident advises "Add support for enforcing the presence of CT proofs (known as Signed Certificate Timestamps (SCTs)) in TLS certificates, and enable this support by default.".

Does Snikket do this?

They've had a resurgence in the Linux Mobile community, now that they've caught up with their shipping backlog.

Librem 5 is the most suitable/polished/stable daily driver for Linux Mobile at the moment.

Purism are funding a lot of the development (Phosh, in particular) that other phones benefit from.

Purism's other products don't seem to have as strong an advantage over their competitors though.