johnklos

I'm not sure what kind of point you're trying to make here, but you're a bit off on a tangent.

You're handwaving. Anything can increase cost and complexity. Using Amazon S3 increases cost and complexity.

Set up two locations. Set up rsync. It's still cheaper than cloud storage by far, and that's even after paying yourself handsomely for the script that runs rsync.

I'm puzzled about why you'd suggest such an obviously silly thing.

Purchased, fully redundant storage is MUCH cheaper than anything in the cloud when talking about any time frame of a year or more.

Obviously you can "rent" storage for a month for less than the cost of purchasing it, but only idiot startup CTOs try to argue a comparison like that.

Two sets of storage are still cheaper, and we all have rsync.

Yes, I did, of course.

Storage is MUCH cheaper when you colo, and bandwidth requirements are a large part of why you colocate instead of just running servers out of an office building that has at least two upstream connections.

I'm really curious what you think they're using now. Certainly you read the article... It says they're using bare metal servers. That's basically colo where the provider owns, but doesn't control, the hardware.

VMs introduce security issues that bare metal don't have. Those security issues are mostly academic for most people and many projects, but not for software where a supply chain compromise could severely impact all users of that software.

Imagine if Wireguard were backdoored because someone working for the ISP that runs the VMs compromised their VMs through the hypervisor. How would a project audit an ISP? How could anything be trusted? Bottom line: it can't. ISPs don't give that kind of information to customers unless you're special (government, spend crazy money).

While it's still possible to compromise a machine through physical access, it's MUCH more difficult. How do you bring it in to single user mode to introduce a privileged user without people noticing that it's down, even momentarily, or that the uptime is now zero? Compromise like this is possible, but worlds more difficult to pull off than compromise through hypervisor.

Is colocation knowledge lost now? Do people no longer know how to configure a server or three, bring them to colo and run them? I don't understand how this is a story worthy of an Ars Technica article. Where's the issue?

If the issue is cost, slightly older Epyc hardware is quite affordable, and colo deals can be found for extremely reasonable costs. If it's expertise, then all they have to do is ask.

I think all these things are absolutely ridiculous.

I use alpine (the email client, not the Linux distro). Before that, I used pine.

Every single thing that gets loaded from anywhere on the Internet has to be the result of an action that I take. Nothing ever gets loaded automatically. I get to choose if I load the thing using the server that I'm connected to, or if I load it directly on my local machine. I know the implications of each.

The fact that programs, particularly ones that are supposed to be for the security minded like Signal, load anything by default, automatically, is just, well, naive.

I can't be the only person who thinks that people who don't think these things through shouldn't be working on apps and email clients. Sure, people would have a cow if their email client didn't load every frigging thing and run remote Javascript and so on, but in Signal? Really?

(end rant)

I see that this can be turned off. I will now tell everyone I know that uses Signal that this should, in fact, be turned off.

This is brilliant and elegant. It's quite satisfying to have ways to have older computers communicate with the rest of the world. Thanks, Cameron, for preserving and extending this work from and history of John Iannetta.

This seems like little more than a sales pitch. For instance:

> Second, companies like Cloudflare benefit from a fast, secure internet. If the internet is fast and reliable, more people will want to use it.

The author doesn't seem to have anything to say with any more substance than this gem.

Not everyone wants to participate in the x86 monoculture and its myriad bugs. A PowerPC machine has many benefits, particularly if someone wants to test that x86 assumptions aren't in their code or infrastructure tools.

Like running NetBSD on the Nintendo Wii, it also has a bit more personality ;)

The depth of this article is wonderful. The PowerPC line did have lots of good things going for it, and the Mac mini G4 is a good example of how much you can get done with modest space, power and heat.

I'm still using Mac mini G4s in several places, both for Mac OS X (legacy Final Chop) and as small, low power servers running NetBSD.

There are really only two drawbacks to the Mac mini G4, in my opinion:

Gigabit ethernet would've been a dollar or two more? Being stuck with 100 Mbps, or around 300 Mbps if one uses a gigabit USB adapter, isn't fun.

If the DIMM slot could take 2 gig DIMMs, this'd be a perfect machine. Other PowerPC Macs could take 2 gigs - heck, even the older PowerPC 604e Power Macs 9500 and 9600 could take 1.5 gigs - so being limited to 1 gig is a bit unfortunate, especially considering that 2 gig DDR DIMMs are a thing (later Xserve G5 units could take 2 gig DDR DIMMs).

Still, the Mac mini G4 is the only computer aside from SBCs that I've bought brand new, and I have always been very happy with my decision.

This is actually not bad at all. Whenever I want to get a feel for a search engine, I do the following searches:

* 1U vax

* 1U amiga

* vax 5g netbsd

The first tells me whether commercial interests are prioritized over other results. The second tells me whether titles and keywords are prioritized over content. The third, well, one can guess, tests whether the overwhelming number of results about "vax" and/or "5g" means the search engine doesn't even take "netbsd" in to account.

For reference, I have what I think is the only 1U VAX in the world (VAXstation 4000/30 in a 1U case) and I have a 1U Amiga, but it's not the only one. For the last result, I posted on Twitter (before it turned evil) about a VAXstation 4000/90a connecting to the Internet using a 5G hotspot over USB.

So lexy.uno did very well for the first search, did moderately OK for the second, and didn't fail spectacularly for the third. Considering how much other search engines suck, this is quite good!

I was anxious the whole way through. I bet it's both wonderful and nerve-racking to own a piece of history like this.

Does "But if you're going out of your way to look suspicious" advance the conversation?

That's oversimplifying a bit and missing some critical information.

Cloudflare wants to "protect" people from exposing even their general region. This has the side effect of making CDNs that aren't Cloudflare work worse. Cloudflare are being dicks because they do to others what they wouldn't want to be done to themselves, or what they themselves don't do to themselves.

It's not even that people are choosing to opt in to Cloudflare's bullshit. If you use Firefox in the US (and many other areas, but the US for sure) and you haven't manually configured Firefox or set up a canary domain, all your DNS lookups are going to Cloudflare, and they're using that to make other CDNs work less well. That's definitely shady and definitely bad on Cloudflare's end.

I'm glad some people are taking a stand.

Is this Hertz? Somehow Hertz in Mexico decided to add my email address to their mailing list, and I tried to complain to every level of Hertz to get them to stop. Their hosters didn't care, their upstream didn't care.

I decided to download larger files from their web site a few tens of millions of times, which I think cost them a few hundred dollars. Unethical? Perhaps, but I'm not the kind of person who just accepts that companies are too large to have humans that can communicate and that I should just accept their harassment.

It worked, though. I finally got a response from Hertz saying they were going to "get to the bottom of it", and I finally stopped getting their spam.

Cloudflare puts challenges on their abuse contact page and rate limits it to much slower than human speed. It's also still broken after years in that you can't report abusers who register domains through Cloudflare and/or host their DNS using Cloudflare.

They really don't want feedback from people who don't pay them.

> The society works best when people don't capture all the fruits of their labor for themselves.

Sure, but it sounds like you think people shouldn't be upset about businesses trying to capture all the fruits of people's labor, too.

Capitalism is evil, and people thinking that normalizing exploitation is OK is either shortsighted or it's also evil. Are you simply unaware that this is what's happening and what people are upset about? Have you never thought about it? Or do you want businesses to succeed in exploiting people's work? It sounds like it, because you wrote, "that's how it supposed to work".

I truly wonder if you're self-aware, or if you just think that you'll one day be on the side of the exploiters.

The author makes some good points, and the overall message is good, but:

Why would we care so much about being a property owner and not renter if we don't care about whether a hosting company can just turn us off?

Why would we care so little about privacy that we'd willingly use services that negate privacy and introduce tracking? For instance, the author suggests using Cloudflare, Azure, AWS and more, and none of these aren't abusive. Funnily enough, the Cloudflare hosted images on the page didn't even load for me;)

Does the author not know that Mastodon is software and not a social media network? Sure, it's a minor thing, but when people use terms incorrectly, it makes me wonder if they really know what they're talking about (people who'd rather fight about whether it's a term "everyone uses" instead of whether it's correct seem to not get this point).

The author writes (or quotes) somewhat derisively:

> Well of course it’s better to host your own blog! Also, while you’re at it, put your Mastodon server in a DigitalOcean droplet, throw some Cloudflare CDN in front of it, run your own Raspberry Pi to monitor uptime, and you’re golden! Oh, and don’t forget to also make sure to log into the droplet every once in a while to update the container, do an occasional database migration, and ensure that you check the logs for intrusions.

This all-or-nothing attempt to make the idea of self-hosting seem ridiculous is by itself ridiculous. Nobody needs to do all that. On the other hand, the audience for what the author is advocating shouldn't have a problem setting up a simple machine or VM instance to host a web site, blog, perhaps even DNS server themselves. All or nothing is silly, so even making this example reduces credibility.

All in all, I'd love to see people take ownership of their own things on the Internet. Nobody needs to self-host, but people should if they could, and they should ignore people who say it's too complex because most of the too-complex argument is the suggestion that it needs to be much more than it does. A single VM, a single small computer, even a Pi, can host most things.

But whether people self host or use someone else, choosing where to host matters. Don't use companies that negate the benefits of owning your own things, whether by lock-in or by letting them do all the tracking that Facebook would normally do. Don't use companies that'll disable your account because some idiot wrote a letter. Don't use companies that're so big that you can't talk to a human!

This article makes for an odd juxtaposition between doing thoughtful things and doing things that negate some or much of that thoughtfulness. I'd have a hard time recommending it to others without qualifications.

I've heard this about various Linux distros, but every time I've tried it has never been as simple as people say it should be.

I assume your twenty lines has customizations. How do you rebuild the OS and kernel the simplest way possible?

According to the app store:

    This app may use your location even when it isn't open, which can decrease device battery life.

Why would an emulator running Linux need someone's location?

I'm not attacking you. I just see too many people say to not do something, and think that's a silly thing to say to a technical crowd.

Also, I don't agree that it's not obvious when something goes down. You don't know how I or my users check my email. If anything, I hear from people much more quickly when there's an issue with email than web!

There's a huge difference between training and running models.

The multimillion (or billion) dollar collections of hardware assemble the datasets that we (people who run LLMs locally) run. The non-open datasets that we-host-LLMs-for-money companies do the same, and their data isn't all that much fancier. Open LLMs are catching up to closed ones, and the competition means everyone wins except for the victims of Nvidia's price gouging.

This is a bit like confusing the resources that go in to making a game with the resources needed to run and play that game. They're worlds different.

I love the idea of small homes for everyone, and I'd want to help with similar projects in my areas. This is inspiring.

Separately, I absolutely love the "lite" aspect of this news web page. Clean, simple, quick, with an easy button to load an image? Yes, please! More of this!

This is exactly what I do, and for exactly the same reasons.

I don't do any kind of content based filtering in part because some of my users do anti-spam, anti-phishing work, so of course they need to be able to talk about spam and phishing, and they need to forward along spam / phishing, without worrying about filters.

Also, every email, accepted or not, has specific reasons for what happens to them, not a vague set of rules that nobody knows like Gmail has.

Any post about self-hosting email inevitably gets posts from people who think it's their place to tell others what not to do.

Replace "email" in what you wrote with "web". Is it just too easy for something to go wrong? Sure, for certain kinds of people. Everyone should just stop hosting servers altogether, if we're worried about things possibly going wrong.

People here, generally, aren't technically illiterate. We don't need you to tell us that because you're not comfortable doing something, we shouldn't.

One way that I explain cognitive load to people unfamiliar with the term is to imagine crossing a lawn that has both autumn leaves and dog poop, and picture how much more mental energy one expends when trying to not step on dog poop.

Lithium AA or AAA batteries won't, nor will good rechargeable.

I just took some lithium AA batteries out of a system that were 25 years old. Not even a hint of leaking.

It's good to see this getting some attention. People should know that removing batteries and replacing them with something modern, particularly non-leaking modern, is a good thing.

Most motherboards work with 3 to 4.5 volts. If a circuit doesn't charge, then it's a simple matter of using two or three AA or AAA batteries. If it does, we can use rechargeable batteries (I've a box of Ikea ones), or we can put a diode in line.

For instance, here's a replacement set for a Quadra 630 motherboard:

https://www.klos.com/~john/q630_batteries.png

Intel has come a long way since butchering the Atom brand. When Atom was new, it had absolutely pitiful performance - much worse than what we got from its contemporaries, particularly considering how much power they took.

It's interesting to consider that this was the best that Intel had in the low power world in 2008, and this was after seeing how much of a misstep they made when they declined to work with Apple on processors suitable for the iPhone.

Even their later Atoms, like the Atom x7-E3950, were bad in real world performance, often being beat by low power AMD CPUs that were years older, like the Athlon 5350.

I'm surprised that Intel kept the Atom name after delivering such poor performance that "Atom" became synonymous with "woefully underpowered", but it seems to make sense that Atom is now just a low power, node optimized version of a previous full performance architecture.