> I assume you're expecting that they'll reach out and cut a deal with each publishing house separately, and then those publishing houses will have to somehow transfer their data over to NVIDIA. But that's a very custom set of discussions and deals that have to be struck.
If this is the only legal way for them to train, then yes that is what they should do instead of breaking the law... just because its not easy doesn't mean piracy is fine.
I haven’t used agents as much as I should, so forgive the ignorance. But a docker compose file seems much more general purpose and flexible to me. It’s a mature and well-tested technology that seems to fit this use case pretty well. It also lets you run all kinds of other services easily. Are there any good articles on the state of sandboxing for agents and why docker isn’t sufficient? I guess the article mentioned docker having a lot of config files or being complex, is that the only reason?
The question is through what mechanism are other branches curtailing his power? It seems to be limited to strongly worded letters and speeches, indignant comments and scathing news reports but nothing real.
No, there's a difference between radically changing your diet and changing up your stretch/strength routine.. you don't just "end up" like one of them, you can evaluate that the downside risk of the latter is much lower and try it safely while recognizing that an extreme diet might not be so safe to try without any professional guidance.
The sprite installer got stuck after "Installed to ..." for me. After waiting a few minutes I just ctrl+ced and looked at what it does after and manually ran "sprite auth setup --token <token>" and that seems to just hang for me.
Canada's healthcare is generally cheaper per capita, pays healthcare workers less and has far lower administrative costs than the US. The US spends 5x the average of other wealthy countries on administrative costs [1]. This line that the government is automatically inefficient and terrible at anything at all is not true, is not set in stone and does not preclude private industry being even more greedy, stupid, amoral and inefficient than the government.
It doesn't have to take deep discipline. I started my routine by just going for short walk after dinner everyday and it built up from there. It's intimidating to have to start at gym for 60 minutes 3x/week. Just do whatever little you like - a few pushups, a walk, a dance and stop when its not enjoyable. It tends to naturally build up.
Seems like that temporary order was set to expire Jan 19th anyways, and there is a hearing for Samsung scheduled tomorrow. So there's still some hope..
Seems like bog-standard stuff doctors and books have been recommending for decades now. Canada has had a food plate like this [1] for a long time. It's a good step forward but I wonder what the actual implications are. How many people didn't already know this, how much does it change behavior and how will it impact other government programs?
It's kind of like when people say Google is getting worse and has too many spam results even while I suspect they're actually improving, but the volume and quality of spam has gone up 100x so it looks like they're doing worse. The question is what is the base rate of attempts to cheat and how many of those attempts does kernel anti-cheat prevent vs. conventional mechanisms. I don't have the answer, but my intuition is cheating is more accessible and viral in many ways now with professional level marketplaces and actors working to build and sell cheats. I also don't think the industry would dedicate so much effort into invasive anti-cheat which is difficult, risky and gets them negative PR unless they felt it truly necessary. Counter Strike a few years ago had huge, huge numbers of cheaters and the super popular games like that attract a lot of attention. But ultimately, this is a cat and mouse game like search & SEO, so you're right there are still cheaters and getting that number to 0 is probably impossible.
This is incredible! I honestly think Rust is an awesome language because it has a lot of high-level niceties like HOF, powerful libraries but can simultaneously let you manage memory manually or put assembly next to your regular code - the borrow checker is an annoyance in a lot of cases like prototyping as you said. I would loove if this was an official mode that Rust supported, Rust could really work in any niche with that capability.
Sorry but you're just old IMO :) PUBG or Arc Raiders have over 100 players in a game. Even Valorant or League have 10 players in a match. It's definitely not easy to find 9 friends to play the same game at the same time as you. And playing any of these games with a cheater can completely wreck the match. If the cheaters go unchecked, over time they start to dominate games where like 30% might be cheaters who can see through walls and insta headshot you and the entire multiplayer mode of the game is ruined. Even worse some cheaters are sneaky, they might have a wallhack or a map showing all players but use it cautiously and it can be quite hard to prove they're cheating but they build up a huge advantage nonetheless. Most of us are happy to have effective anti-cheat, and it's not forced upon us. I understand the tradeoff to having mostly cheater-free games is having to trust the game maker more and am fine with that. Riot for example is quite transparent about what their anti-cheat does, how it works and I don't consider it "malware" anymore than I consider a driver for my graphics card to be "malware" even if they do operate in kernel mode.
It's a blocker if you want to use a TV, there are almost 0 TVs with DP. This HDMI licensing crap is also the reason a Steam Deck can't output HDMI > 4K@60 unless you install Windows on it.
I couldn't find anything from Sony confirming that these specific vulnerabilities had been patched, so i tried to reproduce the steps from the whitepaper using nRF Connect [1] with my Sony WH-1000XM4 on the latest firmware version.
There was no response to the Get Build Version command, and the Read Flash command returned an error. So tentatively (with false negatives possible), it seems to have been patched on Sony devices. I don't have a linux box with bluetooth handy ATM so I didn't try using the race-toolkit directly.
I get the idea, but this language seems to be terrible for humans, while not having a lot of benefits for LLMS besides keeping keywords in single tokens. And I bet like 1 or 2 layers into an LLM, the problem of a keyword being two tokens doesn't really matter.
Same. I actually was fine watching 24/30FPS on an older TV, but on the 120Hz screen it just looks incredibly juddery without a little motion smoothing.
My TV is mostly calibrated to turning off all the processing and D65 white point (Warm 2 on Samsungs) but I can't watch unsmoothed 24p content on the 120Hz screen anymore - it looks incredibly juddery and sort of nausea-inducing. I don't recall having this issue on an older cheap Hisense TV, maybe there's something about higher refresh rate that makes the 24FPS really look bad.
I hope they do. We live in a time of incredibly centralized wealth & power and AI and particularly "the machine god" has the potential to make things 100x worse and return us to a feudal system if the ownership and profits all go to a few capital owners.
Meta and Anthropic atleast fed the entire copyrighted books into the training. Not the wikipedia page, not a plot summary or some tropes, they fed the entire original book into training. They used atleast the entirety of LibGen which is a pirated dataset of books.
China is doing human gene editing and embryo cloning too, we should get right on that. They're harvesting organs from a captive population too, we should do that as well otherwise we might fall behind on transplants & all the money & science involved with that. Lots of countries have drafts and mandatory military service too. This is the zero-morality darwinian view, all is fair in competition. In this view, any stealing that China or anyone does is perfectly fine too because they too need to compete with the US.
What the fuck? People also need to pay to access that creative work if the rights owner charges for it, and they are also committing an illegal act if they don't. The LLM makers are doing this illegal act billions of times over for something approximating all creative work in existence. I'm not arguing that creative's make things in a vacuum, this is completely besides the point.
> Meta clearly enganged in copyright enfringement when they torrented books that they hadn't purchased. That is enfringement already before they started training on the data. That doesn't make the training itself enfringement though.
What kind of bullshit argument is this? Really? Works created using illegally obtained copyrighted material are themselves considered to be infringing as well. It's called derivative infringment. This is both common sense and law. Even if not, you agree that they infringed on copyright of something close to all copyrighted works on the internet and this sounds fine to you? The consequences and fines from that would kill any company if they actually had to face them.
We have clear evidence that millions of copyrighted books have been used as training data because LLMs can reproduce sections from them verbatim (and emails from employees literally admitting to scraping the data). We have evidence of LLMs reproducing code from github that was never ever released with a license that would permit their use. We know this is illegal. What about any of this is ideological and unreasonable? It's a CRYSTAL CLEAR violation of the law and everyone just shrugs it off because technology or some shit.
By this logic there is no corporation or entity that provides anything other than basic food, shelter and medical care that could be criticized - they're all just providing something you don't need and don't have access to without them right?
I live in Vancouver, and we have plenty of both roads and bike lanes. Its not hard to fit a bike lane that's usually 1/4th the width of a lane onto a road or allow bikes to share with cars on smaller roads. We have trucks and vans and lots of deliveries too. The reason most cities are oriented around cars is because we designed them that way and it's difficult to change - there's no logistical constraint, its just politics and cost.