That’s delusional. The EU won’t even deploy economic countermeasures against the US, let alone military resistance. The US could take Greenland by force tomorrow and the EU response would be a poetry recital.
> I don’t think these things work the way you think they work.
Perhaps, but nobody reasonable would have forecast this situation in the first place, what Trump is doing here already wildly outside of any recent precedents for the USA's behaviour.
> Nobody really cares about Greenland.
Nobody should, the inhabitable parts are tiny, the rest is a massive ice sheet, the population wouldn't even half-fill the largest single stadium.
> Nobody is willing to allow there to be real consequences, or even real inconveniences, as a result of anything to do with Greenland.
Tell that to Trump, he's the one threatening military force to get an island he's already allowed to build whatever bases he wants on. There's no good reason for him to have burned his bridges like this. Even if he doesn't invade, he's already severely weakening relations with people who thought they were American allies, who have already come to American aid when asked.
> The EU is primarily an economic and regulatory structure, not a military alliance.
Primarily, yes, but it does also have a mutual defence clause. Never been tested, of course. Why would anyone be dumb enough to threaten an EU member state with military conquest?
And yet, here we are.
> NATO is a paper tiger anyway, people will invent some justifications and keep doing business as usual.
Article 5 has been invoked exactly once, to aid the USA. NATO-minus-USA is going to be wild, almost certainly forces a lot of other members to rapidly develop nukes of their own even if this all goes "peacefully".
I’ll put myself in the minority here by saying that I think Trump is probably right. Greenland can’t be credibly defended by Denmark, the EU or even NATO. Article 5 is an untested foundation myth. Greenland is far away. Political will matters. We might be heading towards an independent Greenland if we continue following the status quo, which would be influenced strongly by adversaries and would be a US security nightmare.
I’d say that I prefer him to go about it a different way, except that I can’t see what that different way looks like when you want territory from another country that doesn’t want to give it to you.
And I say this as a European. Europe is not credible from a defense perspective and lacks the will to do very much of anything quickly or effectively. The best you can expect is a series of talking shops and some policy documents to be drawn up while the ice continues to melt.
If trump is actually deranged enough to use military force against Greenland we'll see how capable the EU is of defending it - and I suspect they'll put on a good show.
What the EU wouldn't be able to handle, I suspect, is would be a full ground invasion by China, not that China would/needs to do that.
If the US was genuinely concerned about the security of Greenland they should have discussed this with the EU and encouraged them to reinforce the island, and/or offered a joint base.
> If the US was genuinely concerned about the security of Greenland they should have discussed this with the EU and encouraged them to reinforce the island, and/or offered a joint base.
This is where it gets stupid... well, stupider.
The US already has a base on Greenland, namely the Pituffik Space Base / Thule Air Base [1].
The US used to have a larger military presence in Greenland, including other bases, but choose to downscale their presence following the end of the cold war [2].
This presence was predicated on the 1951 Defense of Greenland agreement between Denmark (and later the autonomous government of Greenland) and the US, which allowed the US great freedom in establishing their military presence in Greenland [3].
If Trump had just wanted a stronger military presence in Greenland, then all he would have had to do was ask, and Denmark and Greenland would most likely have agreed. Denmark, in particular, has done its best to align itself with the US, and Greenland, prior to Trump, was also interested in a closer relationship with the US as part of their move towards greater independence from Denmark.
That requires Denmark and the EU to be reliable defense partners. They had decades to invest in Greenland and its defense and what we end up with is a 12 member dogsled patrol armed with bolt action rifles, so they can defend themselves against polar bears.
The EU response to the rhetoric from Trump is to send 30 men and put out a press release telling everyone how harmless they are. The action “Poses no threat to anyone”. Their military show of force, poses no threat to anyone.
I get the feeling behind the anti-AI sentiment, I just disagree with the conclusions.
There’s a lot of fear around what will happen with AI, not so much of extinction but rather of two things: fear of losing income, and arguably more importantly, fear of losing identity.
People often are invested in what they do to the point that it’s who they are. That being replaced or eliminated might be a bigger psychological threat than lack of income, at least to those of us fortunate enough to be well off right now.
However, these threats are outweighed by the benefits that AI can eventually bring. Medical advances, power generation, manufacturing capability. Our systems for running society have a lot of problems, economically, politically, epistemologically. These can also be improved with AI assistance.
The real problem is the transition, it’s such a huge shift, and it will happen all at once to everyone, uprooting our idea of the world and our place in it.
What we need is to embrace AI and find a way to make sure that the transition and benefits of AI are distributed instead of concentrated.
For me this looks like the following: companies must commit to retaining some minimum number of employees in every currently existing function, to be determined proportional to their profit taking. This sets a floor on the job losses that can come later when AI really comes on stream.
The justification for this is three fold: firstly, it’s a safety mechanism, it ensures that regardless of the capabilities of an AI system, there are multiple humans working with it to verify its results. If they aren’t verifying diligently, then they’re not doing their jobs.
Secondly, jobs aren’t just a way of making income, they’re wrapped up in identity and meaning for at least some people, and this helps to maintain that existing identity structure across a meaningful cross section of society.
Third, it keeps the economy running, money circulating. You can’t have a market economy without consumers. UBI is one component of this too, but this is both more direct, more useful and more meaningful.
> However, these threats are outweighed by the benefits that AI can eventually bring. Medical advances, power generation, manufacturing capability. Our systems for running society have a lot of problems, economically, politically, epistemologically. These can also be improved with AI assistance.
Benefits come to those who have the means to access it, and wealth is a measure of the ability to direct and influence human effort and society.
How exactly do you propose that AI will serve the wellbeing of the worker/middle classes after they've been made obsolete by it?
Goodwill of the corporations working on them? Of their shareholders, well-known to always put welfare first and profit second? Government action increasingly directed by their lobbying?
> What we need is to embrace AI and find a way to make sure that the transition and benefits of AI are distributed instead of concentrated.
Sure. How? We've not done it with any other technological advances so far, and I don't see how shifting the power balance further away from the worker/middle class will help matters.
There's a reason why the era of techno-optimism has already faded as quickly as it's begun.
Of course, and distribution and ownership of benefits is the real issue here, but I think I’ve addressed that.
Let me be clearer: I said “companies must commit to” where the stronger phrasing is “companies are forced to by legislation”. But to begin with this might be voluntarily done by some number of companies.
Also, in this vision of society the AI companies (OpenAI, Anthropic, google etc) are taxed heavily. The taxation is redistributed, there is UBI for some fraction of the population, maybe the majority. Others still work in companies mandated to keep employees as I outlined above.
Importantly, we as a society specifically aim to bring about these benefits of AI by using the redistributed funds in part to invest in them.
Part of this is the free market, part is planned government investment. If one fails, maybe the other succeeds. Either way, we try to spread the benefits and importantly to ensure the benefits are actually there in the first place.
To me it is less the fear of losing identity and income.
These can be figured out if the system allows for it.
But if the system will not allow for it, for example if we loose our Democratic system along the line, all bets are off.
I think this is probably a trend that will erode with time, even now it’s probably just moved underground. How many human artists are using AI for concepts then laundering the results? Even if it’s just idea generation, that’s a part of the process. If it speeds up throughput, then maybe that’s fewer jobs in the long run.
And if AI assisted products are cheaper, and are actually good, then people will have to vote with their wallets. I think we’ve learned that people aren’t very good at doing that with causes they claim to care about once they have to actually part with their money.
A huge issue with voting with your wallet is fraud. It's easy to lie about having no AI in your process. Especially if the final product is laundered by a real artist.
Because voting with your wallet is nonsense, we can decide what society we want to live in we don't have to accept one in which human artists can't make a living. Capitalism isn't a force of nature we discovered like gravity, it's deliberate choices we made.
Which I assume is why you pay someone to hand-paint scenes from your holidays instead of taking photographs? And why you employ someone to wash your clothes on a scrubbing board instead of using a machine?
Or would you prefer these things be outlawed to increase employment?
You have to always devolve to individual responsibility and freedom to make your case. But games are a 250+ billion dollar industry employing hundreds of thousands of artists who's jobs are all threatened by generative ai in the future, that's systemic, structural. We can all look at that future and decide to make a different choice, that is actual freedom, what you describe is collective helplessness.
Everyone is talking about moving to Linux lately, it’s a bit of a trend. I wish they’d stop, for one simple reason: I’ve been using Linux exclusively (when I’m not forced to use macOS by work) for several years now, and I rather enjoy the lack of malware, spyware and other bullshit on the platform.
If the general public comes over this situation might end. Desktop linux isn’t a target right now because its niche, I’d prefer that didn’t stop.
While I sympathize with this angle, there's another side to this coin: if more people do the switch, maybe some applications will finally get linux versions.
I'm a Sunday photographer and quite like Lightroom and Photoshop (I know about the drama, but to me, I get enough value from them compared to Darktable and the GIMP to not switch just yet). It's the only reason I still have a windows pc hanging around the house.
I am in a similar boat; my media editing machine ruined windows 10 so that I can use Lightroom. But I would dearly love to ditch windows so I'm currently looking to try out running Lightroom under Winapps to see if it is usable. There's no way of passing the GPU through without something like SR-IOV so I'll have to see how it goes.
I was thinking of doing that, but since that would require me to switch the monitor and whatnot, it would be just like using two PCs. And since I only use my desktop for LR and not much else, jumping through the hoops with emulation doesn't make much sense.
How so? Winapps lets you run windows applications as if they were native to Linux, you interact with them the same way you would anything installed by apt/pacman/dnf etc. Unless I'm very much misunderstanding things (which I don't believe I am)
In the general case, I think you're right. WinApps seems to use RemoteApp functionality on windows to export just the window you're interested in from the virtualized guest vm to the host, which should behave mostly as a "native" app.
But you were talking about sr-iov, which is a whole different matter. Presumably, the goal is to have LR use that GPU for some of its functions. But LR doesn't support multiple GPUs: it does its computation on the same GPU that handles the output. For that, you need to connect the display to the passed-through GPU. Now, aside from intel, I don't think any mainstream GPU actually supports sr-iov, so you need to pass through the entire gpu to the guest VM (the host wouldn't see it anymore at all). This isn't how RemoteApp works, and I doubt WinApps handles this case.
I remember a project (Looking Glass?) that tried to somehow "bring back" the output to the host machine, but it didn't seem too robust at the time. I haven't followed it, so I have no idea if it's any better now, if it's still alive. If it does, this could possibly work if you had two GPUs (which I happen to have, since my CPU has an integrated GPU). But you'd still get the whole Windows desktop of the VM, not an RDP connection.
There's a lot of servers running Linux that are regularly targeted by malware.
There is a big difference in what software a desktop user runs versus what runs on a server, but the great thing about Linux is that you can keep just as much variation between your install and the average desktop user.
Your best bet for security is probably running OpenBSD, but within Linux, if you avoid common optional applications and services like Gnome, KDE, pulseaudio, systemd, etc., you'll have a significantly different attack vector. Avoiding Python and Node package managers and sticking to your distribution's package manager would be great, too.
Better spread the Linux word because with enough users more developers will be attracted and the race good vs bad hackers in OSS will be won be the former. "Nothing is hidden under the sun". Closed source is made to push malware secretly.
That is factually incorrect flamebait. Closed source is made primarily due to a desire to retain control. While one can use control for malicious reasons, the predominant use is to make money.
You overestimate how influential HN is. Everyone on HN is talking about moving to Linux. Which means, uh, nothing really changed for the general public.
Folks on reddit and hackernews aren't normal people. Outside of this bubble few people have heard of linux. Hell so few people I know use firefox which makes me mad. You are safe from that fear.
After a bit of digging - it looks like it's done to sharpen features as one of the standard steps in producing these images. Where there are rotational symmetries in the things they're looking at, they focus on the smallest unit, and then rotate accordingly. If you had a trilateral symmetry, or hexagonal structure, they'd rotate 3 or 6 times around the center.
You're not getting a real image of the thing, but apparently it's got data from those other segments mixed in with the rotations, so you're getting a kind of idealized structure, to make the details being studied pop out, but if you have some sort of significant deviation, damage, or non symmetric feature it'll show up as well.
So kind of like taking a picture of a human, and then taking each half, flipping along the midline, and blending to get an idealized Symmetrical Human?
Perhaps you assumed a "radially" which wasn't part of my analogy? :p
Land animals have a pretty consistent trend of exterior bilateral symmetry which is very noticeable. (Naturally, a completely normal Hunam such as myself cannot speak for how it may work in places other than my home planet Dirt.)
I understood you meant bilateral symmetry. And yes, there are similarities, but we are not bilaterally symmetric. At least not to the extent where you can flip an image and have that look normal.
Even faces look weird when flipped that way (there have been studies on this effect too). And that’s before you get into the issue that it’s common to have differently shaped breasts, different sized hands or feet. Ears shaped differently. Non-uniform teeth. And so on and so forth.
I think that might just be the original and it simply is symmetrical to that degree. I found a few more examples of "cryo-em center slices" and I've yet to find one that doesn't have really strong symmetry down to the small dot patterns.
Sure, but the headline wasn't "Google CEO says ‘vibe coding’ made software development ‘so much less like a video game.’"
In fact since many people think video games are enjoyable, making software development less gamelike might make it less enjoyable.
(But would further gamification make it more enjoyable? No, IMO. So maybe all we learn here is that people don't like change in any direction.)
I think we’re mixing our metaphors here, what I mean is at the end of the day you write code to get some result you actually care about, or that matters for some material reason. Work is labor at the end of the day. If you don’t care about that outcome or optimizing for it, then you may as well play a video game or code golf or something. What you now want is a hobby.
It depends on how you use it. I was running 15 agents at once, 12 hours a day for a month straight because it was more optimal to add more, and that wasn't very enjoyable. Now I'm back to writing code the enjoyable way, with minor LLM assistance here and there.
reply