My gut reaction is that this won’t work long-term. Users on youtube often point to specific timestamps in a video in comments or link to specific timestamps when sharing videos, meaning there needs to be some way to identify the timestamp excluding ads. And if there’s a way to do that there’s a way to detect ads.
Of course, there’s always the chance they just scrap these features despite how useful they are and how commonly they’re used; they’ve done similar before.
It might not necessarily be that the instances are stricter, it could also simply be that those instances are targeted more often by hate/trolls so interact with those instances more often. Admins are less likely to defederate from an instance they’ve never seen or heard of. I see a lot of obviously LGBT-related names in this list which likely get more hate than average.
There’s no point looking for logic. These people truly believe granting a licence restricts the rights of people who don’t agree to the licence, which is the exact opposite of what licenses do. It’s blatant misinformation but if you call them out on it (even by quoting their own link) they literally think you’re an astroturfer for AI, because that makes more sense to them than the fact they’re obviously wrong.
That link is a 404 so I can’t tell what it says, but here’s a 1996 US act to enforce net neutrality: https://en.m.wikipedia.org/wiki/Telecommunications_Act_of_1996
And here’s a 2006 Tim Berners-Lee blog post about threats to net neutrality which specifically says net neutrality already exists, you really can’t get much more authoritive than that: https://web.archive.org/web/20060703142912/http://dig.csail.mit.edu/breadcrumbs/node/144
Obama may have enacted some legislation around between neutrality (again, your link 404s so I can’t tell what specifically you’re referring to) but it certainly wasn’t created under Obama.
The concept of net neutrality definitely existed long before Obama so it’s a bit questionable to say it was created under him. Did anything specific happen under him to enforce net neutrality more than it already was?
You’re definitely right about Trump though. It seems like he took every opportunity to screw over the US public in favour of corporate interests.
The name is an analogy, neural networks do not work in the same way as biological neurons. They were designed by computer scientists, not biologists.
RAM is so far removed from biological short term memory both in how it works and how it’s used that the comparison doesn’t even make sense. The only similarity is that they’re short term information/data stores, so it’s equally valid to compare them to a drawing in the sand of a beach.
If you can’t read a licence plate at 20 metres you can’t safely drive. Being able to read a licence plate at 20 metres does not make you a safe driver. The test still matters.
There’s no discount there, you’re just accepting their marketing bullshit. That sounds to me like the company is double-dipping by shoving ads in your face and making the product objectively worse, then charging even more for a “premium” model where the only difference is they haven’t intentionally downgraded it.
with extras like […] no lockscreen ads
What the fuck? Why is that an extra not just the default? It’s great that this product isn’t riddled with ads, but that’s like saying it’s great a burger is not made of human shit; it’s crazy that anyone would tolerate a shit-burger in the first place.
Maybe ads are normal in the e-reader space for some reason, but that’s just insane to me.
This may have been true historically but I’m not sure it still holds up. I switched to Linux Mint as my regular OS a while back and the only driver issue I’ve had was that the installer didnt properly install my wifi card’s proprietary driver (which was working during live boot from usb), so I had to tether to my phone to download the driver through the driver manager. It even installed Nvidia drivers just fine.
It might still be an issue for more barebones or heavily customisable systems but I’m fairly certain nobody’s recommending people switch to Arch for their first Linux experience.
A keyboard without tactile feedback is objectively worse than a keyboard with tactile feedback, excluding other factors.
I’ve never had a physical keyboard lag out then send an entirely different keystroke because it thought I held a button, or send a single keystroke because I was typing too quickly.
I’ve never had to wait a moment for a physical keyboard to show up after selecting a text box.
I’ve never had the entire layout of a page shift to make room for a physical keyboard whenever I select or deselect a text box.
I’ve never had a physical keyboard prevent me from using the number pad and force me to use the full keyboard (or worse, vice versa) because of an improperly configured input box.
The way I see it there are exactly two real benefits to integrating a software keyboard into a touchscreen: reduced physical complexity (the entire device is essentially just one screen), and easier access to emoji. A touchscreen keyboard performs far worse as a keyboard. It’s a valid trade-off for a small mobile device, but it’s not objectively better.
Anything we’ve had before now wasn’t AI.
This claim doesn’t work simply due to the fact AI is a very vague term which nobody agrees on. The broadest and most literal (and possibly oldest) definition is simply any inorganic emulation of intelligence. This includes if statements and even purely mechanical devices. The narrowest definition is a computer with human-like intelligence, which is why some people claim LLMs are not AI.
Saying LLMs work differently from older AI approaches is fair, saying older approaches are not AI but the latest one is is questionable.
There’s no need to be a dick about it, just because you know a something doesn’t mean everyone else must also know it. I can guarantee there’s common terms you aren’t familiar with. Especially when you consider that this is marketing jargon with common alternate terms (eg demo reel).
It’s a bit of a non-sequitur though, the context was denying service to an ally not cutting off existing service to Crimea. It’s like if someone asked “Have you ever shaken a baby” and you respond “I have never kicked this baby!”. Sure, it’s good that you haven’t kicked a baby, but that’s just not the question.
They’ll use old comments either way, using an up-to-date dataset means using a dataset already tainted by LLM-generated content. Training a model on its own output is not great.
Incidentally this also makes Lemmy data less valuable, most of Lemmy’s popularity came after the rise of LLMs so there’s no significant untainted data from before LLMs.
It’s not just repeated moves, a draw can be called if the board is in the same state 3 times at all during the game; if you get to the same position 3 times using different moves that still counts, even if it was a white move the first two times and a black move the third.
The game also ends after 50 moves with no captures or pawn moves so you can’t play indefinitely by just avoiding those board states. Interestingly those two moves also make it impossible to return to a previous board state (pawns can’t move backwards, extra pieces are never added) so if you’re enforcing both rules in code you can safely discard previous board states every time you reset the move counter.
Or you could say “and were shocked at the results”! It’s nice of you to highlight how people communicate differently even with a shared language, the world would be so boring if everyone was the same.
Your mistake there is thinking the stock market has to make sense. For instance, mass layoffs are a huge red flag that a company is failing and in any sane world would instantly tank the value. The stock market instead likes layoffs because it’s not at all interested in what the company actually does, so spending a little less in the short term to produce a lot less long term is a good thing.
You’re not wrong, but the first words are literally “Just over a decade ago”. It’s not a news article, it’s the story of the research in 2013 which revealed bitcoin isn’t anonymous.
I’m not sure where you’re getting the idea that language models are effective lie detectors, it’s very widely known that LLMs have no concept of truth and hallucinate constantly.
And that’s before we even get into inherent biases and moral judgements required for any form of truth detection.