The Russian plans don't stop at Ukraine, they extended to Moldova, Poland, and lots of other nations.

Is it viable for them to do that in the long term? You never know. Right now it seems clear that the answer is no, they can barely take Ukraine so how are they going to handle Poland and the other countries when they've already expended their economic tools that could have done so?

But it's generally in our better interest so make sure that they are as incapable as possible of expanding those borders.

We would be fools to allow them to expand those security borders because it means losing those allies and a huge amount of trade and resources that come with them.

They are either in our circle, or they are in Russia's. Russia treats these countries as former colonies, so as long as they are given freedom of choice they're going to ally themselves strongly with the west.

For the relatively low price of funding Ukraine we slow down Russia's advance at worst, and we totally halt it at best. If we manage to totally halt Russian advances and equip Ukraine to be competitors for Russia, we've now got a country that has just survived a war with Russia sitting on their borders armed to the teeth and angry at them.

They are literally fighting the war for us, without risking nuclear conflict. Why shouldn't we be interested in this tragedy?

> You always have to take responsibility for your actions, and Ukrainians pay for these now.

We aren't defending Ukraine because we are nice people, or defending them because we quite like the consequences of Ukraine's decision, and we're more than happy to support and ensure that those consequences are as minimal as possible, while ensuring the consequences of Russia's decision are as high as possible.

It weren't for that fact, the ability for countries like Germany to trade oil with Russia still would have totally overridden any possibility of sanctions.

And the West very frequently does not get involved in areas where terrible things are happening so long as those areas aren't really in their interests.

Which doesn't justify Russia in any form shape or way. Personally, because it's individuals we aren't involved in international politics and we are still perfectly valid in criticizing Russia for being hilariously evil.

And because in the world of might makes right, Russia is also losing.

So it doesn't really matter. Any Russian who tries to make this argument might be correct in calling out hypocrisy, but their weak and stagnant nation lacks the might to make the right.

We don't have to give them to Ukraine, but doing so generally furthers our interests.

https://www.youtube.com/playlist?list=PLWuyJLVUNtc0UszswD0oD...

freeCAD had a critical issue in topological naming (they are working on a fix as we speak and it's going to be out "soon"), and it's hideously clunky. Basically, when you resize a sketch and that causes more faces to be created it makes all your other stuff go crazy.

You can also use realthunders branch which fixes the naming issue as well.

Mangos videos teach you a bit how to think reasonably though freecads infuriating errors, too. The guy is worth a watch if you want to start using the best open source tool.

freeCAD sucks, until you learn it, and then it works alright. Learn to use data planes instead of sketch on face and the topo naming isn't even an issue anymore.

Edit:

The mods here apparently think responding to the Russian influence campaigns on their site is worthy of a shadow ban.

They should have improved their platform and lock useful features behind a paywall, like video and image hosting. They could have made money on their core appeal - hosting communities for people.

Instead, they've bloated their site into an adware riddled clusterfuck, ignored useful features that people actually wanted, and just turned into a hilariously useless company in general, making their site worse with every change instead of better.

We desperately need regulation that kills the network effect that feeds these companies, because it's clear their users aren't the priority.

Yes, it actually works. That's what matters at the end of the day.

But no, eventually you're going to have to fine tune on something that has a larger context or training a brand new model with the position embeddings semi randomized so that it learns to generalize in the first place instead of needing hacks like this to function.

But training a model costs millions and millions of dollars, so we're going to have to wait until some very generous group decides to do all that training and then release that model open source.

Or releases the model file for a paid charge fee, I'd pay like $200 for a good one.

This is so exciting because contacts length has been a problem with these models for so long and it's awesome to see open source finally cracking that egg.

I need another GPU.

edit:

I'm actually a little bit skeptical of this. Yes it's dynamically scaling which is great when you have a model that's not fine tuned, but I think it's not going to work out too well when you try fine tune a model on a target that's moving like that. I'd rather one that stay static so that perplexity is always increasing up to the max rather than doing much those graph shows where it gets worse over time.

That said I don't really know what I'm talking about so maybe it'll be better regardless.

Imagine loading an iron rod into your car, driving for a while, and then when you get into the gas station you dump a pile of rust off and buy a fresh iron bar.

They want the law and they want to promise safety while not being impacted by the overbearing regulation they've invited onto the rest of us.

HN is not wrong about the fact that there were trying to push regulation in the United States that would help their business. It just turns out the European laws will actually hurt their business.

A big truck is more deadly to smaller vehicles and we shouldn't make it unsafe to drive a regular sized car instead of one of these monstrously large trucks.

You could regulate things like requiring stop gates on trucks like semi trucks have though, which would at least help.

We forget that AI becomes more human as it gets more intelligent. These are not computers programmed with a hard limitation anymore, they're actually more prone to being manipulated than other humans are.

To be fair, I didn't say you have to have a good common cause. Just something to do in some capacity.

Like, you don't need help, it's fun to hang out and it's an excuse to get help.

My point is to say the relationships are different rather than transactional. Focused on a common cause and that sharing of it rather than relationships directly.

Or maybe it's just a me thing.

Like, at work for me I would be most rewarded by other people who are engaged in that common work cause and interact within that context. Not with stories about family life or opening up about whatever.

Hanging out with friends is kind of pointless without a common cause. To do play some game or sport or actually do something that's worth doing and is fairly collaborative.

To me a friend is like "hey I'm building a fence, wanna help?". And the ability to be building a fence yourself and asking the same. Or whatever project you can imagine.

The "you good" shit is mostly a consequence of lacking that sort of common cause, at least in my experience.

Or just watching other people do stuff is good too. "Hold my beer" silly stuff that you can watch and joke and one up each other about.

But if someone started that family huggy stuff (barring needing help or talking about their problems) I'd be distancing myself from that hard. In my experience those people are out to get you, smile on a snake.

I know that's less of a concern nowadays with modern http, but guess who still supports ie11

Standard library that isn't worthless - no more npm package spam.

Type knowledge in the runtime. It doesn't have to be typed, just make it so I don't have to do crazy checks on properties before casting to a class.

Imports without webpack.