“This is not the first time Musk has gone after the site. In December, he posted on X, “Stop donating to Wokepedia.” And that wasn’t even his first bad Wikipedia pun. “I will give them a billion dollars if they change their name to Dickipedia,” he wrote, in an October, 2023, post.

You can’t buy it, but you can go create a conservative fork called whatever you want, and you can fund its operation. If people think that it’s better, they can choose to use it.

There’s a Conservapedia, which I would call pretty off-the-rails; people have tried this before.

https://en.wikipedia.org/wiki/Conservapedia

Conservapedia (/kənˌsɜː®vəˈpiːdiə/; kən-SU®-və-PEE-di-ə) is an English-language, wiki-based, online encyclopedia written from a self-described American conservative[2] and fundamentalist Christian[3] point of view. The website was established in 2006 by American attorney and activist Andrew Schlafly, son of Phyllis Schlafly,[4][5] to counter what he perceived as a liberal bias on Wikipedia.[6][7] It uses editorials and a wiki-based system for content generation.

Examples of Conservapedia’s ideology include its accusations against and strong criticism of former US president Barack Obama—including advocacy of Barack Obama citizenship conspiracy theories[8]—along with criticisms of atheism, feminism, homosexuality, the Democratic Party, and evolution. Conservapedia views Albert Einstein’s theory of relativity as promoting moral relativism,[9] claims that abortion increases risk of breast cancer, praises Republican politicians, supports celebrities and artistic works it believes represent moral standards in line with Christian family values, and espouses fundamentalist Christian doctrines such as Young Earth creationism.[10][11] Conservapedia’s “Conservative Bible Project” is a crowd-sourced retranslation of the English-language Bible which the site says to be “free of corruption by liberal untruths.”[12]

I’ve no fundamental objection to doing so. I don’t think that existing aircraft are considered ready for that, though.

It sounds like the issue is the bytes that come before the data it wants, rather than after.

Could probably search for it, if one could determine from a snippet of the content whether-or-not one is too far.

I suppose that more automated testing could help catch some of these before releases.

Hmm. Do they accept an HTTP Range-Request?

It looks like it’s more-commonly called “coffee concentrate”.

https://www.amazon.com/s?k=coffee+concentrate

I do also see it sold as “coffee syrup”, but that gets mixed up with syrups that are added to coffee.

Facts are not copyrightable, just their presentation. So I don’t think that it’s possible to say that it’s impossible to summarize material. A court is going to say that some form of summary is legal.

On the other hand, simply taking material and passing it through an AI and producing the same material as the source — which would be an extreme case — is definitely copyright infringement. So there’s no way that a court is going to just say that any output from an AI is legal.

We already have criteria for what’s infringing, whether a work is “derivative” or not.

My bet is that a court is going to tell Brave “no”, and that it’s up to Brave to make sure that any given work it produces isn’t derivative, using existing case law. Like, that’s a pain for AI summary generators, but it kind of comes with the field.

Maybe it’s possible to ask a court for clearer and harder criteria for what makes a work derivative or not, if we expect to be bumping up against the line, but my guess is that summary generators aren’t very impacted by this compared to most AI and non-AI uses. If the criteria get shifted to be a little bit more permissive (“you can have six consecutive words identical to the source material”, say) or less permissive (“you can have three consecutive words identical to the source material”), my guess is that it’s relatively easy for summary generators to update and change their behavior, since I doubt that people are keeping these summaries around.

If you make iced coffee with it its going to be fine.

If you make iced coffee, like, mixing stuff up with ice, I suspect that you’re better off just getting coffee syrup, which is available, and lets you select how dilute you want the coffee. Doesn’t make a lot of sense to ship large bottles of water around when you already have a sink.

Yeah, I was gonna say.

A typical US household circuit is 15 amps. At 120 volts, that’s 1800 watts. You’ll also see some 20 amp circuits, but that’s still just 2400 watts. You’d have to go to the heavy appliance circuits to draw more than that.

I wouldn’t be surprised if the 240V circuits that some countries use can probably squeeze more on; might be aimed at that.

kagis

https://www.quora.com/How-many-amps-are-in-the-UK-household-current

In the UK, the standard household current is typically supplied at 230 volts with a frequency of 50 Hz. The circuit breakers in most homes are usually rated for 16 amps for general-purpose circuits, such as lighting and power outlets.

That’d be 3680W. So you could run it on a typical household circuit in the UK, though it’d eat up a lot of the circuit’s capacity.

EDIT: I mean, I assume that realistically, this is for datacenters trying to optimize compute cost, buy fewer PCs to run more compute cards in parallel, so they probably aren’t going have as much of a problem with it.

Also, if they keep ramping up how much heat they’re putting out per unit of space, I don’t know whether existing datacenters are going to be able to handle the cooling.

I mean, it’s pretty safe to say that they’ve got good incentive to do it to the best of their ability. Right now, the cards are selling for above MSRP because there isn’t enough supply. They aren’t making that money — it’s going to scalpers and such.

If they wanted to constrain supply to benefit, they’d increase the prices they’re charging.

That’s a common problem with these local models that lack company-provided guardrails. They could expose people to any manner of things.

Generated locally with ComfyUI and a Flux-based model:

A red ball balancing on a white ball with a blue ball balancing on top.

That didn’t sound like the best idea ever for someone using jets so often.

I am sure that he doesn’t fly on commercial public transport.

since 2015

Honestly, I’d say that a lot of Trumpism’s stuff is more-or-less in line with the stuff that the John Birch Society has promoted, and that goes waaaaay back. I mean, Trump talking about annexing Canada/Panama/whatever, no — in fact, that’s one of the few cases that I think that they’d take a dead-opposite position on, since they’ve a horror of the North American Union. But there’s a lot of overlap outside that.

https://en.wikipedia.org/wiki/John_Birch_Society

The John Birch Society from its start opposed collectivism as a “cancer” and by extension communism and big government.[29][30] JBS publications referred to the fight against Communism as a spiritual war against the devil.[25]: iv, 156–157  Allegations that so-called “Insiders” have conspired to control the United States through communism and world government are a recurring theme of JBS publications.[31] The organization and its founder, Robert W. Welch Jr., promoted Americanism as “the philosophical antithesis of Communism.”[32] It contended that the United States is a republic, not a democracy, and argued that states’ rights should supersede those of the federal government.[33] Welch infused constitutionalist and classical liberal principles, in addition to his conspiracy theories, into the JBS’s ideology and rhetoric.[34] In 1983, Congressman Larry McDonald, then the society’s newly appointed chairman, characterized the JBS as belonging to the Old Right rather than the New Right.[35] The society opposes “one world government”, the United Nations (UN),[36] the North American Free Trade Agreement (NAFTA), the Central America Free Trade Agreement (CAFTA), the Free Trade Area of the Americas (FTAA), and other free trade agreements. It argues the U.S. Constitution has been devalued in favor of political and economic globalization. It has cited the existence of the former Security and Prosperity Partnership as evidence of a push towards a North American Union.[37][38] The JBS has sought immigration reduction.

The JBS opposed the civil rights movement of the 1960s and the Equal Rights Amendment in the 1970s.[16][39][40] It has campaigned for state nullification.[41][42] It opposes efforts to call an Article V convention to amend the U.S. Constitution,[43][44] and it has been influential at promoting opposition to it among Republican legislators.[45] The JBS also supports auditing and eventually dismantling the Federal Reserve System.[46][non-primary source needed] The JBS holds that the United States Constitution gives only Congress the ability to coin money, and does not permit it to delegate this power, or to transform the dollar into a fiat currency not backed by gold or silver.[non-primary source needed]

If you look at the article, it was only ever possible to do local processing with certain devices and only in English. I assume that those are the ones with enough compute capacity to do local processing, which probably made them cost more, and that the hardware probably isn’t capable of running whatever models Amazon’s running remotely.

I think that there’s a broader problem than Amazon and voice recognition for people who want self-hosted stuff. That is, throwing loads of parallel hardware at something isn’t cheap. It’s worse if you stick it on every device. Companies — even aside from not wanting someone to pirate their model running on the device — are going to have a hard time selling devices with big, costly, power-hungry parallel compute processors.

What they can take advantage of is that for a lot of tasks, the compute demand is only intermittent. So if you buy a parallel compute card, the cost can be spread over many users.

I have a fancy GPU that I got to run LLM stuff that ran about $1000. Say I’m doing AI image generation with it 3% of the time. It’d be possible to do that compute on a shared system off in the Internet, and my actual hardware costs would be about $33. That’s a heckofa big improvement.

And the situation that they’re dealing with is even larger, since there might be multiple devices in a household that want to do parallel-compute-requiring tasks. So now you’re talking about maybe $1k in hardware for each of them, not to mention the supporting hardware like a beefy power supply.

This isn’t specific to Amazon. Like, this is true of all devices that want to take advantage of heavyweight parallel compute.

I think that one thing that it might be worth considering for the self-hosted world is the creation of a hardened network parallel compute node that exposes its services over the network. So, in a scenario like that, you would have one (well, or more, but could just have one) device that provides generic parallel compute services. Then your smaller, weaker, lower-power devices — phones, Alexa-type speakers, whatever — make use of it over your network, using a generic API. There are some issues that come with this. It needs to be hardened, can’t leak information from one device to another. Some tasks require storing a lot of state — like, AI image generation requires uploading a large model, and you want to cache that. If you have, say, two parallel compute cards/servers, you want to use them intelligently, keep the model loaded on one of them insofar as is reasonable, to avoid needing to reload it. Some devices are very latency-sensitive — like voice recognition — and some, like image generation, are amenable to batch use, so some kind of priority system is probably warranted. So there are some technical problems to solve.

But otherwise, the only real option for heavy parallel compute is going to be sending your data out to the cloud. And even if you don’t care about the privacy implications or the possibility of a company going under, as I saw some home automation person once point out, you don’t want your light switches to stop working just because your Internet connection is out.

Having per-household self-hosted parallel compute on one node is still probably more-costly than sharing parallel compute among users. But it’s cheaper than putting parallel compute on every device.

Linux has some highly-isolated computing environments like seccomp that might be appropriate for implementing the compute portion of such a server, though I don’t know whether it’s too-restrictive to permit running parallel compute tasks.

In such a scenario, you’d have a “household parallel compute server”, in much the way that one might have a “household music player” hooked up to a house-wide speaker system running something like mpd or a “household media server” providing storage of media, or suchlike.

Everything you say to your Echo…

I don’t have an Echo.

Mark Carney was sworn in as Canada’s 24th prime minister on March 14, declaring “We will never, in any shape or form, be part of the US,” rejecting Donald Trump’s annexation threats.

Carney won the Liberal leadership with 85.9% of the vote despite having no elected experience.

In recent weeks, the Liberals have reversed a political freefall, sharply rebounding to such a degree that a previously expected Conservative majority in the next general election looks increasingly unlikely. The shift in the polls has been so dramatic that pollsters have struggled to find any historical precedent.

A newly released poll from Abacus Data showed the Conservative support had shrunk to 38%, with 34% going to the incumbent Liberals.

I don’t know what impact the Trump administration is having on the likelihood of conservatives having political power in the US in the future, but it sure isn’t having a positive effect on conservatives in Canada.

Robert F. Kennedy, Jr. has a broad range of unorthodox views on medicine.

A survey of over 100,000 Germans revealed that 94% won’t buy a Tesla vehicle.

Ehhh…

So, normally, you want a random sample in polls, which is very unlikely to not be representative of the population as a whole. If they have 100k people, it very probably isn’t a random sample, because you only normally take something like 1k to 2k people for randomly-sampled polls; there’s a rapidly-declining value above that. If the sample set is self-selected rather than randomly-selected, you can get results that are pretty different from the population as a whole.

fires up Google Translate

While I can’t seem to get the survey page to load, the domain it’s on is apparently t-online.de; it sounds like it’s a reader survey, which won’t be random.