tal

most AI generated content that tries to look real ends up quite uncanny

I think that a lot of people who say this have looked at a combination of material produced by early models and operated by humans who haven't spent time adapting to any limitations that can't be addressed on the software side. And, yeah, they had limitations ("generative AI can't do fingers!") but those have rapidly been getting ironed out.

I remember posting one of the first images I generated with Flux to a community here, a jaguar lying next to a white cat. This was me just playing around. I wouldn't have been able to tell you that it wasn't a photograph. And that was some time back, and I'm not a full-time user, professionally-aimed at trying to make use of the stuff.

kagis

Yeah, here we are.

https://sh.itjust.works/post/27441182

"Cats"

https://lemmy.today/pictrs/image/b97e6455-2c37-4343-bdc4-5907e26b1b5d.png

I could not distinguish between that and a photograph. It doesn't have the kind of artifacts that I could identify. At the time, I was shocked, because I hadn't realized that the Flux people had been doing the kind of computer vision processing on their images as part of the training process required to do that kind of lighting work at generation time. That's using a model that's over a year old

forever, at the rate things are changing

from a non-expert on just local hardware, and was just a first-pass, not a "generate 100 and pick the best", or something that had any tweaking involved.

Flux was not especially amenable, as diffusion models go, to the generation of pornography last I looked, but I am quite certain that there will be photography-oriented and real-video oriented models that will be very much aimed at pornography.

And that was done with the limited resources available in the past. There is now a lot of capital going towards advancing the field, and a lot of scale coming.

As for the ‘corporate control’ aspect, all this stuff is racing towards locally run anyway (since it’s free).

I am not at all sure about that. I use an XT 7900 XTX and a Framework Desktop with an AI Max 395+, both of which I got to run LLMs and diffusion models locally, so I've no certainly no personal aversion to local compute.

But there are a number of factors pulling in different directions. I am very far from certain that the end game here is local compute.

In favor of local

• Privacy.

• Information security. It's not that there aren't attacks that can be performed using just distribution of static models (If Anyone Builds It, We All Die has some interesting theoretical attacks along those lines), but if you're running important things at an institution that depend on some big, outside service, you're creating creating attack vectors into your company's systems. Not to mention that even if you trust the AI provider and whatever government has access to their servers, you may not trust them to be able to keep attackers out of their infrastructure. True, this also applies to many other cloud-based services, but there are a number of places that run services internally for exactly this reason.

• No network dependency for operation, in terms of uptime. Especially for things like, say, voice recognition for places with intermittent connection, this is important.

• Good latency. And no bandwidth restrictions. Though a lot of uses today really are not very sensitive to either.

• For some locales, regulatory restrictions. Let's say that one is generating erotica with generative AI stuff, which is a popular application. The Brits just made portraying strangulation in pornography illegal. I suspect that if random cloud service is permitting for generation of erotic material involving strangulation, they're probably open to trouble. Random Brit person who is running a model locally may well not be in compliance with the law (I don't recall if it's just commercial provision or not) but in practical terms, it's probably not particularly enforceable. That may be a very substantial factor based on where someone lives. And the Brits are far from the most-severe. Iranian law, for example, permits execution for producing pornography involving homosexuality.

In favor of cloud

• Power usage. This is, in 2025, very substantial. A lot of people have phones or laptops that run off batteries of limited size. Current parallel compute hardware to run powerful models at a useful rate can be pretty power hungry. My XT 7900 XTX can pull 355 watts. That's wildly outside the power budget of portable devices. An Nvidia H100 is 700W, and there are systems that use a bunch of those. Even if you need to spend some power to transfer data, it's massively outweighed by getting the parallel compute off the battery. My guess is that even if people shift some compute to be local (e.g. offline speech recognition) it may be very common for people with smartphones to use a lot of software that talks to remote servers for a lot of heavy-duty parallel compute.

• Cooling. Even if you have a laptop plugged into wall power, you need to dissipate the heat. You can maybe use eGPU accelerators for laptops

I kind of suspect that eGPUs might see some degree of resurgence for this specific market, if they haven't already

but even then, it's noisy.

• Proprietary models. If proprietary models wind up dominating, which I think is a very real possibility, AI service providers have a very strong incentive to keep their models private, and one way to do that is to not distribute the model.

• Expensive hardware. Right now, a lot of the hardware is really expensive. It looks like an H100 runs maybe $30k at the moment, maybe $45k. A lot of the applications are "bursty"

you need to have access to an H100, but you don't need sustained access that will keep that expensive hardware active. As long as the costs and applications look like that, there's a very strong incentive to time-share hardware, to buy a pool of them and share them among users. If I'm using my hardware 1% of the time, I only need to pay something like 1% as much if I'm willing to use shared hardware. We used to do this back when all computers were expensive, had dumb terminal and teletypes that connected to "real" computers that ran with multiple users sharing access to hardware. That could very much again become the norm. It's true that I expect that hardware capable of a given level of parallel compute will probably tend to come down (though there's a lot of unfilled demand to meet). And it's true that the software can probably be made more hardware-efficient than it is today. Those argue for costs coming down. But it's also true that the software guys probably can produce better output and more-interesting applications if they get more-powerful hardware to play with, and that argues for upwards pressure.

• National security restrictions. One possible world we wind up in is where large parallel compute systems are restricted, because it's too dangerous to permit people to be running around with artificial superintelligences. In the Yudkowsky book I link to above, for example, the authors want international law to entirely prohibit beefy parallel-compute capability to be available to pretty much anyone, due to the risks of artificial superintelligence, and I'm pretty sure that there are also people who just want physical access to parallel compute restricted, which would be a lot easier if the only people who could get the hardware were regulated datacenters. I am not at all sure that this will actually happen, but there are people who have real security concerns here, and it might be that that position will become a consensus one in the future. Note that I think that we may already be "across the line" here with existing hardware if parallel compute can be sharded to a sufficient degree, across many smaller systems

your Bitcoin mining datacenter running racks of Nvidia 3090s might already be enough, if you can design a superintelligence that can run on it.

https://www.investopedia.com/warren-buffett-s-massive-war-chest-11826399

Warren Buffett's Berkshire Hathaway Inc. (BRK.A, BRK.B) has amassed the largest pile of cash ever held by a public company. At $344 billion, Berkshire Hathaway's war chest is more than the combined cash reserves of Apple Inc. (AAPL), Microsoft Corp. (MSFT), Alphabet Inc. (GOOG), Amazon.com Inc. (AMZN), and NVIDIA Corp (NVDA)—despite them being collectively 14 times Berkshire's market value.12 Also striking is that the record-breaking stockpile has doubled in just over a year.

So what gives? As in everyday life, companies save for three main reasons: to prepare to weather an economic storm, to make a major purchase, or because they think what's available isn't worth it—in market parlance, it's overvalued.

A key chart value investors like Buffett use could help us narrow down the options: the S&P 500 index's historic price-to-earnings ratio. That's because it now sits 67% above its historical norm and almost 50% above its early 2022 value. This remarkable deviation could be a major reason that the famed Oracle of Omaha could be storing cash.

Why Buffett's Cash Pile Keeps Growing

Buffett famously preaches a straightforward investing philosophy: Be fearful when others are greedy. Given Buffett's "pledge" to Berkshire shareholders to practice "extreme fiscal conservatism" and since market valuations have been well above historical norms, it's no surprise, perhaps, that Berkshire sold over $100 billion in stocks during the first nine months of 2024, including cutting its massive stake in Apple by two-thirds.

No problem. Yeah, it's an invaluable off-site resource that new users don't get informed about. Indexes all of the Threadiverse instances, whereas any given instance can only search the communities that are local or at least one local user has subscribed to.

https://lemmyverse.net/communities?query=networking&order=active_month

Actually, if I were Canada, I'd probably want to be selling lumber into the EU too, which would make their life a lot easier regarding the US.

The EU has gotten a lot of lumber from Russia.

Canada has had long-standing trade disputes with the US over lumber, where the US lumber industry has pushed for protectionist policy. Like, decades and decades.

https://en.wikipedia.org/wiki/Canada%E2%80%93United_States_softwood_lumber_dispute

The Canada–U.S. Softwood Lumber Dispute is one of the largest and most enduring trade disputes between both nations.[1] This conflict arose in 1982 and its effects are seen till today. British Columbia, the major Canadian exporter of softwood lumber to the United States, was most affected, reporting losses of 9,494 direct and indirect jobs between 2004 and 2009.[2]

If Canada had two very large markets, that'd probably make their life easier. And the EU does not have a lot of lumber, which is why masonry construction is more common than in the US and Canada in most of Europe.

kagis

https://www.theglobeandmail.com/business/article-lumber-industry-softwood-crisis-summit-bc-ottawa-sawmills/

American import taxes on softwood from Canada now total 45.16 per cent for most Canadian producers.

The U.S. Department of Commerce raised duty rates in the summer. Lumber supplies from Canada currently face an anti-dumping duty rate of 14.63 per cent and a countervailing duty rate of 20.53 per cent, equalling 35.16 per cent for most Canadian producers. That’s up sharply from duties totalling 14.4 per cent previously.

New 10-per-cent tariffs on shipments of softwood lumber from Canada and other countries took effect on Oct. 14. U.S. President Donald Trump announced the new levies in late September, citing Section 232 of the Trade Expansion Act, which allows him to impose tariffs on the basis of national-security concerns.

“This situation is the direct result of the protectionist economic policies advanced by President Donald Trump,” Mr. Parmar said.

https://globalnews.ca/news/11477871/bc-softwood-lumber-tariffs-us-higher-russia/

B.C.’s softwood lumber U.S. tariffs now higher than Russia’s: ‘Let that sink in’

https://forestmachinemagazine.com/russian-conflict-timber/

Due to take effect from the end of December 2025, the EUDR is the EU’s most meaningful effort yet to end its complicity in skyrocketing levels of forest destruction and affiliated human rights abuses. It does this by regulating the trade in timber, palm oil, soy, beef and other products driving that destruction. The law would also greatly aid efforts to stem the EU’s imports of conflict plywood from Russia and Belarus.

Canada’s Energy Minister Tim Hodgson said the Toronto meeting will aim to formally launch a new initiative designed to curb China’s market influence. The Critical Minerals Production Alliance will “secure transparent, democratic, and sustainable critical mineral supply chains across the G7,” he said. Under the alliance, the governments of Britain, Canada, France, Germany, Italy, Japan, and the United States would mobilize private investment to expand critical mineral production that bypasses China.

I'm going to guess that Canada has a bunch of said reserves and would be happy to sell them, particularly if the US is decoupling supply chains from China and the EU from Russia.

kagis

https://www.canada.ca/en/natural-resources-canada/news/2024/12/canada-to-unlock-critical-minerals-rare-earth-development-in-northern-quebec-and-labrador-with-new-funding.html

Canada to Unlock Critical Minerals Rare Earth Development in Northern Quebec and Labrador With New Funding

Investments in critical minerals infrastructure are essential for Canada to seize the enormous economic opportunity presented by the low-carbon economy and to capitalize on our rich mineral resources. Canada is well positioned to be a global leader and a first-class producer of a wide variety of critical minerals that are essential for powering the clean economy, strengthening national defense capabilities and ensuring national and economic security. By developing and expanding critical mineral value chains — from mining and processing to manufacturing and recycling — Canada can create good jobs, support economic opportunities and contribute to a resilient and secure future.

https://www.canada.ca/en/campaign/critical-minerals-in-canada/critical-minerals-an-opportunity-for-canada.html

Canada’s critical minerals

What makes them critical

To be considered a critical mineral in Canada, a mineral must meet both of the following criteria:

• the supply chain is threatened

• there is a reasonable chance of the mineral being produced by Canada

It must also meet one of the following criteria:

• be essential to Canada’s economic or national security

• be required for the national transition to a sustainable low-carbon and digital economy

• position Canada as a sustainable and strategic partner within global supply chains

Yeah.

Honestly, even limiting it to, say, the WiFi network, having a default admin login is not great.

Like, Android isolates apps from the rest of your Android system, but not from touching the rest of the network. If any random app I install on my phone can reflash my WAP's firmware or something like that, that's not great.

In all seriousness, after the COVID-19 thing, the JIT manufacturing people didn't start stockpiling this stuff or at least obtaining alternate sources?

The European Automobile Manufacturers’ Association (ACEA) issued an urgent warning on Wednesday saying its members, which include BMW, Fiat, Peugeot and Volkswagen, were now working on “reserve stocks but supplies are dwindling”.

“Assembly line stoppages might only be days away. We urge all involved to redouble their efforts to find a diplomatic way out of this critical situation,” said its director general, Sigrid de Vries.

Another ACEA member, Mercedes, is now searching globally for alternative sources of the crucial semiconductors, according to its chief executive, Ola Källenius.

We had to face this exact problem, automakers blowing up due to lack of reliable chip supply, like five years ago and automakers went running for state support and didn't get it hammered out then?

kagis

Looks like it's not just European auto manufacturers, either:

https://gmauthority.com/blog/2025/10/u-s-auto-plants-could-shut-down-in-a-few-weeks-due-to-chip-shortage/

U.S. auto plants could face major production slowdowns within weeks as a deepening conflict between China and Western nations over semiconductors threatens to disrupt global supply chains. The issue first came to light earlier this month, as GM Authority reported previously.

As reported by Bloomberg, the situation stems from Beijing’s decision to block chipmaker Nexperia from exporting components from its China-based facilities following the Dutch government’s move to seize control of the Chinese-owned company. According to the Motor & Equipment Manufacturers Association, or MEMA, the largest vehicle supplier group in the United States, manufacturers could begin feeling "significant impacts" in as little as two to four weeks if the issue persists.

EDIT:

https://en.wikipedia.org/wiki/2020%E2%80%932023_global_chip_shortage

Between 2020 and 2023, there was a worldwide chip shortage affecting more than 169 industries,[1] which led to major price increases, long queues, and reselling among consumers and manufacturers for automobiles, graphics cards, video game consoles, computers, household appliances, and other consumer electronics that require integrated circuits (commonly called "chips").[2][3][4]

I mean, come on.

My American take is that this is really on France. Like, I don't have the ability to see all the horse-trading that happens behind the scenes, so maybe France is making tons of concessions on other things that we just can't see, but France bailed on the Eurofighter, and got a lot of concessions on the FCAS that I probably wouldn't have made from the standpoint of the other parties involved.

France wants the FCAS to be a CATOBAR carrier-based fighter. No other European partners want a CATOBAR fighter, and there is almost no international market for this, since few countries plan to operate CATOBAR carriers, and most of those already have their own fighters. Maybe India. There are maybe two other routes for France I could have imagined if they are determined to do a carrier fighter: (a) accept STOBAR carriers for compatibility with other European parties operating carriers, (b) get a bunch of other EU members to fund a CATOBAR carrier fleet. But neither of those have happened. Otherwise, I'd say either drop the carrier fighters or use F-35Cs. CATOBAR just makes a bunch of tradeoffs that don't make sense for anyone else in order to stock the hanger on a single ship, and it makes the aircraft less-competitive from an export standpoint. France pushed on Germany to make the FCAS more-exportable, extracting a requirement that no one sole FCAS member could deny export to a customer (the concern being that an export customer might be worried about buying FCAS if Germany might block export of parts or aircraft). But making the thing for CATOBAR does no favors for export.

I also don't see a huge benefit here for France. France avoids F-35Cs, likes to not use US planes...okay, fine. But they're already relying on US hardware

the catapult that they're using is going to be a US one, because they don't have the funds to go develop their own. Is France going to do sea control with a sole CATOBAR carrier? That'd be a rationale for using a nuclear powerplant, having high, sustained speed. But it can't be operational all the time, which is a large drawback. It's a single point of failure for French naval aviation, unlike, say, the Brits, which decided to do two less-expensive carriers.

France's claimed alternative to the FCAS

at least what Dassault keeps saying they will do, though I have no idea if they really speak for France

is to just go do an updated Rafale. That's going to put France significantly behind the curve in terms of having a competitive manned fighter. My understanding, from past reading, is that the reason that France might do this is because France is just going to bet really heavily on large UAVs and will rely on such a UAV coming out, being technically-successful, and not slipping schedule. Then France would worry about doing a manned fighter sometime down the road, with fewer time constraints. That's got some considerable risk.

France wants Dassault to have lead on the project and a lot of control over it and work. One concession that France made was to give Germany lead on a (much smaller) Franco-German tank project, the MGCS. Spain, as far as I understand, hasn't gotten anything comparable. And France is only contributing a third of the capital for the project.

My understanding is that France wants to maintain an independent French capability to do a full aircraft in the future, without requiring any other EU partners. But if the argument is that FCAS is a European project and represents European defense-industrial integration, then that should be something that France is willing to put on the table.

My impression from past reading is that both Germany and Italy had had some interest in trying to get FCAS/Tempest efforts merged. That is, both have considered that to be an option.

https://www.aviacionline.com/fcas-and-tempest-programs-to-eventually-merge

FCAS and Tempest programs to eventually merge

At least that is what the head of the Italian Air Force, General Luca Gorett, thinks, as he expressed it last Tuesday before the parliamentary defense committees.

Goretti also noted that since the two programs are currently in their «conceptual phase,» it is normal that, in this initial period, each country evaluates options in terms of technology.

«But it is natural that these two realities merge into one, because investing huge financial resources in two equivalent programs is unthinkable,» Goretti told members of the parliamentary defense committees.

https://www.defensenews.com/industry/techwatch/2021/07/12/will-the-fcas-and-tempest-jet-programs-merge-germanys-top-air-force-officer-hopes-so/

Both are years away from a flight-ready demonstrator aircraft. In the meantime, the German Air Force chief of staff said he has spoken to his Italian and British counterparts about possibly combining efforts.

“It can be that we go on different tracks. Hopefully we will merge eventually,” Lt. Gen. Ingo Gerhartz said in an exclusive interview with Defense news en route to Berlin, Germany, from Mihail Kogălniceanu International Airport outside of Constanta, Romania.

I've read statements in the past out of France that they would certainly not consider such a merger of efforts acceptable.

If France actually lines up a European fighter project and then winds up being the odd man out after Germany and maybe Spain wind up with Italy, the UK, and Japan doing Tempest instead...that just seems bonkers to me.

From what I've read, one problem with Germany joining Tempest is that Japan has the short pole on timeline

they have some internal hard timeline constraints to have Tempest out. Germany's timeline is less-constrained, and may not want to push the aircraft out so soon. But in general, Germany's got more flexibility than France does. Playing hardball with Germany seems kinda nuts.

If France bails, then they've got an older airframe for naval aviation for some time. They're preserving their independent aircraft production capacity, but for what? Do they plan to need to operate independently from the rest of Europe's military-industrial capability? Like, let's say that France wants to do a seventh-gen fighter. Great, they can do all of the aircraft themselves. But they are unlikely to be in a position to do a competitive one, because they won't have a defense budget on par with the US or China or whoever for it.

And if they bail and withdraw their offer of Germany's leadership on the MGCS, then they've got another mess to fix there, since now they're out their next-gen tank.

Naval aviation lets you reach out and touch something a long ways away. But it sounds like The Italian-British-Japanese coalition is likely going to be focusing hard on range with Tempest:

The U.K. Royal Air Force officer in charge of defining requirements for the Tempest future fighter says the program’s top priority is a large payload — roughly twice that of the F-35A stealth fighter. The same officer says the service is eyeing “really extreme range” for the new aircraft, with potentially enough internal fuel to fly across the Atlantic without refueling.

Using a land base further away doesn't mean that you can have as-high a sortie frequency. But point is, that does at least offset some of the benefit of using naval aviation. If you have a lot of long-range, land-based fighters, that can provide some of the benefits of naval aviation.

I mean, I'm sure that everyone involved has their reasons. But to me, it seems like France gets surprisingly little out of what it's fighting for. An independent industry that is likely to have a hard time producing competitive fighters while hedging against risk of...what, influence from other countries in Europe? Those countries probably have other levers of influence. A non-US-originating fighter on a carrier that's US-originating hardware? A carrier aimed at sea control, but only a lone one, which makes for a very limited inventory to control the seas? None of those seem to buy a huge amount. And for that, trading off scale, export potential, utility in the probably-more-important terrestrial capacity?

It just seems like that France has a lot of FCAS asks, and that some of those asks are not buying France a lot relative to the tradeoffs that they'd probably have to make to get them.

My understanding is that America doesn't actually have 5G telco equipment manufacturers

it's a weak point in the US tech lineup

which is one reason that there was such a major kerfuffle several years back over it. The US does not want China in its sensitive telco infrastructure. There was some point I remember where some US senator said that if Europe wasn't going to support Nokia or Ericsson

Europe does have some entrants -- and just let Huawei take the 5G market, that the US would need to buy one of them. There was some real concern from the US that Europe might just accept ceding the 5G provider market, which would make the US dependent on China.

kagis

This is probably what I remember. I could have sworn that it was a senator, but maybe there were multiple statements or it was actually AG Barr.

https://www.reuters.com/article/technology/white-house-dismisses-idea-of-us-buying-nokia-ericsson-to-challenge-huawei-idUSKBN2012A5/

White House dismisses idea of U.S. buying Nokia, Ericsson to challenge Huawei

WASHINGTON (Reuters) - U.S. Vice President Mike Pence on Friday and the top White House economic adviser dismissed an unusual suggestion from U.S. Attorney General William Barr that the United States consider taking control of two major foreign rivals of China-based Huawei Technologies Co Ltd.

White House economic adviser Larry Kudlow added later on Friday that the United States was working closely with Nokia and Ericsson, saying the companies' equipment was essential to the buildout of 5G infrastructure. But he said the "U.S. government is not in the business of buying companies, whether they're domestic or foreign," adding that "there's nothing to prohibit American tech companies from acquiring" them.

"That's the plan the president has endorsed and will be carrying forward," Pence said, adding that the United States can expand 5G "by using the power of the free market and American companies."

In a remarkable statement underscoring how far the United States may be willing to go to counter Huawei, Barr on Thursday disclosed proposals "by the United States aligning itself with Nokia and/or Ericsson."

EDIT: IIRC, Cisco or some other US company that I can't remember made some 5G hardware, but they weren't on par with Nokia and Ericsson.

I think that, TP-Link aside, consumer broadband routers in general have been a security problem.

• They are, unlike most devices, directly Internet-connected. That means that they really do need to be maintained more stringently than a lot of devices, because everyone has some level of access to them.

• People buying them are very value-conscious. Your typical consumer does not want to pay much for their broadband router. Businesses are going to be a lot more willing to put money into their firewall and/or pay for ongoing support. I think that you are going to have a hard time finding a market with consumers willing to pay for ongoing support for their consumer broadband router.

• Partly because home users are very value-conscious, any such provider of router updates might try to make money by data-mining activity. If users are wary of this, they are going to be even more unlikely to want to accept updates.

• Home users probably don't have any sort of computer inventory management system, tracking support for and replacing devices that fall out of support.

• People buying them often are not incredibly able to assess or aware of security implications.

• They can trivially see all Internet traffic in-and-out. They don't need to ARP-poison caches or anything to try to see what devices on the network are doing.

My impression is that there has been some movement from ISPs away from bring-your-own-device service, just because those ISPs don't want to deal with compromised devices on their network.