@nanook @bajax

> Friendica is primarily a linear system but it does thread replies.

I haven't used it, but my experience interacting with Friendica/Hubzilla people is that they have trouble reading long enough threads. Pleroma's FE, as well as bloat and FediBBS, all present thread structure.

@nanook @bajax If the entire pipe on the dedi has been saturated and normal operations have been not just impeded but effectively DDoS'd, and if I blocked 500k uniques in a month (not counting the blanket ban on all PRC IPs a few days ago) and I'm still blocking at least one per second, 24 hours a day, I'd say it's extremely aggressive. A high number of reqs/second isn't necesarily aggressive, a low number isn't necessarily innocuous.

They are way more difficult to stop than BoardReader was ( blog.freespeechextremist.com/b… ) but they are somewhat less aggressively DDoSing.

@nanook @bajax

> Consider some UFW rules

I don't use UFW, just iptables; ufw is like a set of barely working filters in front of iptables. But it's not a matter of correctly blocking, like...why would I be blocking 8443 if I'm not exposing that port to the world?

I mean, have a look at that blog post, I'm a little ahead of the game on this one.

@nanook @bajax @mkultra @DiamondMind

> mariadb-repair database

Guaranteed not to work, because I'm not using that database (it's Postgres, and it doesn't have the same issues that MyISAM/InnoDB have), and if it were a database error, probably someone else would be seeing it by now. Error rate for the last 100k requests is 0.82%.

What is happening is probably some bullshit issue we have on occasion where a malformed post causes the frontend to give up. Usually what fixes the issue is filling up the timeline to bump the problems out of the first page.

@nanook

> It's one more than we have.

Well, one's one. That's all I'm saying. If they can scale it up, that's good. Maybe they can't. Whatever. One's an experiment, a prototype, it's one.

@nanook @bajax

> Mastodon is a Microblogging forum, in the same vein as twitter, and is limited to 500 characters per post. Friendica and Hubzilla do not have this limit

The number of characters per post is not related to the data and UI correctly modeling the structure of the thread in which those posts appear.

@nanook @bajax UFW rules turn into iptables rules and I it has some sort of fail2ban integration and I have never seen UFW work properly on any machine where it's set up, so I don't use it.

As noted, though, it's not really got to do with "too many connections on 443" but differentiating between legitimate traffic and the problem traffic.

@mkultra @bajax @j @nanook

Browser refresh usually fixes when I see those notifications.

@nanook

> The Chinese pattern has always been build one, prove it works, built a thousand so I am confident that they will.

That is how scale works, yes. That's not the Chinese method, that's everyone's method.

"It works" and "It works at scale" and "We can build the scale" are all different questions and they did the first one. If they do the rest, sure, that's great. I'm interested in cheap nuclear energy. As far as the likelihood that they do or do not do this, a nuclear power plant is not like building a cell phone. Maybe unforeseen problems occur. Maybe they don't.

Right now, though, no one has built thorium salt reactors at scale. That's it. I understand you would like them to be real and viable and I would like them to be real and viable but that has yet to be demonstrated so I am waiting. I am not building nuclear power plants at present so I have no influence on the outcome.

@nanook @bajax

It's rare, I don't sweat it. I can read all my obscure internet micro-celebrity mutuals, and reliably so.

And I can say al sorts of fun words. Perfection.

@nanook @bajax

> I'm at a bit of a loss as to understanding exactly what you are trying to communicate. I run a hubzilla and I run a friendica and both are capable of very large posts.

This is exactly my point. Pleroma can display arbitrarily large posts as well. But the difference between a warehouse and a junkyard is organization. Data is not different: it can be architected such that it is easy to follow a thread or not. Compare the way mutt/Thunderbird/sylpheed/etc. structure email threads versus the way GMail structures them.

Case in point, the thread seems to you like a stream of non-sequiturs rather than a branching conversation. And I say "Threads seem to confuse a lot of Friendica/Hubzilla people." and I think this is the UI.

@SilverDeth @bajax @j @mkultra @nanook How often do you see those?

At present, I basically never run into errors using bloat.fse, but that's bloat.

@nanook @menherahair @bajax I know what nftables is.

I am old as shit.

@bajax @j @mkultra @nanook

Once or twice a month, it's not a big deal. My assumption is always you're tinkering with something below the hood, and I don't sweat it.

Using Brave for now, but I swap browsers a lot.

@nanook @SilverDeth @bajax

> The only thing a refresh is going to do is resend the same request

It sends one request, then that request triggers several others. One of the things is that the UI is reinitialized because Pleroma does most (nearly all) of its UI client-side, so if something made the UI crash, refreshing fixes it.

> so if not dropping packets, something else is intermittent

Yes, something else. As previously noted, the backend accepts data that sometimes crashes the frontend.

> It does suggest a resource exhaustion of some sort.

Nope. Sometimes. That is not the case here. FSE is mostly idling this time of night. Weather conditions have caused FSE to fall over a few times.

> Perhaps run dmesg after a failure

I appreciate you trying to help debug software that you have never run but I don't think you're going to be able to help much.
idle.png

@nanook Sure, but electronics manufacturing is a much more well-understood problem than building nuclear power plants.

I'm not saying it won't happen or that they can't do it. I hope they *can* do it. I'm just saying that they haven't done it yet.

@SilverDeth @bajax @j @mkultra @nanook

> My assumption is always you're tinkering with something below the hood, and I don't sweat it.

Sometimes it's that, sometimes it's weirdness in the hops between here and the datacenter or the datacenter to you.

@nanook I have heard you say that you think it is going to be easy. If it were easy, they'd already have a thousand plants. Unless Kirk Sorensen stood up a thousand plants, it's all unproven.

Making one of something is very different from making a hundred of it. I will believe that they can make a hundred when they make a hundred. Right now, it seems possible, it may even seem plausible, but it is not *done*. Telling me that they can is not going to affect my belief in whether or not they will do it, and there is no reason to convince me, since neither of us can affect the outcome.

There is also no reason: if you are reasonably convinced, then I will agree with you in a couple of years. There's no reason to hurry, is there? Do I need to have a positive belief in the practicality of scaling up thorium reactors *before* the thorium reactors are scaled up?

I'm fairly convinced that the reason we don't have cheap power is more or less entirely a political matter.

If you try to build one in a 1st world country, there's so much regulation that it's just not going to happen.

If you try to build one in a non-1st world country, you're gonna get bombed because "muh nuclear proliferation".

China is working on it, but they're probably facing quiet international backlash because once the cat's out of the bag, everyone is going to want one...

It's basically like Free Energy suppression, except it actually happens.

@cjd @nanook
>I'm fairly convinced that the reason we don't have cheap power is more or less entirely a political matter.

it's a matter of finance.
as an intern, i ran numbers for an investment firm my first year of grad school.

it's really just not profitable to build nukes. they take a long time to build and they cost a lot.

if it was possible to make cheap power, you could just do it, nobody is stopping you from putting down a power plant on a strip of land, i worked with solar farms a few years ago there is basically no barrier to entry for those guys you just pay the money and wire it into the grid.

@nanook Okay, here, like, this is the clearest I am able to be on the topic: I am not going to agree with you that a thing that I do not believe to be certain is certain. I hope that thing works out but it has not worked out yet. The only thing you can say to change my mind on the topic is "They have just built the 100th thorium salt reactor". So far, what we have is a viable prototype: that's good news. If it's trivial like you keep insisting, then we'll have a hundred in no time flat. I will believe it when I see it. I hope I see it.

> So to be clear it's already been done but not scaled up to commercial power levels, the Chinese will be the first to do that.

Maybe. It doesn't look like anyone else will do it before they do. The first energy-positive fusion reaction just happened a the big fusion reactor in Europe: that's encouraging, too. A lot of interesting things are happening and if any of them pans out, it'll be really cool. None of them have panned out yet.

I do not see the urgency. My belief does not influence the outcome. So there are no stakes, there's no urgency. I have no reason to form a belief in either direction. It is definitely possible based on what I currently know, which is why I say "I will believe it when I see it" and not "That's bullshit and will never happen."

> if it was possible to make cheap power, you could just do it

And then men with guns come and take away your house.

You were doing the math on a pressurized water reactor, and all of the safety equipment that is expected when you have hot radioactive stuff under high pressure.

If you use molten fuel (not even a thorium breeder, just plain old boring uranium), you have no pressure to deal with, you could use ceramic pipes, a ceramic Archimedes pump, so basically you need beryllium and lithium fluoride, ceramic clay, u233, high purity graphite, a boiler & steam turbine, and lots and lots of concrete.

None of those things are that costly. They're not *cheap*, but they're not expensive in comparison to being able to crank out like 30kw of power all day and all night.

If it weren't for regulation, there'd be youtubers doing this, I'm sure of it.

@BowsacNoodle @cjd @nanook

>smaller
you'll always have fixed costs, such as staff on hand. nukes actually benefit the most from being bigger

en.wikipedia.org/wiki/Small_mo…

Predictably, only Russia and China have them, because they're the only ones who care about having cheap power AND are tolerated to have nuclear weapons.

Small modular reactor - Wikipedia

^{Contributors to Wikimedia projects (Wikimedia Foundation, Inc.)}

@cjd @nanook
>if it weren't for regulation
>guys with guns
i can tell you what it looks like on our side: you model it the way you model something like an earthquake, it's just a risk priced into the operational cost.

from what i recall, this risk-adjusted cost was not substantial. this directly contradicts your thesis that "guys with guns" regulation is the barrier.

>if you use thorium
i've long been aware of internet guys talking about thorium reactors. it wasn't something we had data for. the tech is interesting, and i hope it takes off and is everything people promise. i also really like the idea of a fusion reactor.

one thing i will remark is, if thorium is as good as people are saying, why is nobody building more of them? you can just build power plants: it's not any different from building an apartment complex or running a machine shop, anyone can do it.
regulation or not, if it was as incredible as people bill it, people would be building them en masse, you couldn't stop me from building ten thorium reactors, it's literally free money. but that's not what we see happening.