I use that one on iOS. In Firefox I use the native functionality (the cookiebanners.service.mode flag). See https://community.mozilla.org/en/campaigns/firefox-cookie-banner-handling/. I also set cookiebanners.ui.desktop.enabled to true to make this setting appear in the settings menu.
That’s not as effective, since it can’t block anything that’s hosted from a hostname that also serves regular content without also blocking the regular content. It also can’t trick websites into thinking that nothing is blocked and it can’t apply cosmetic rules. I use it for my devices, but in browsers I supplement it with uBlock Origin (or whatever is available in that browser).
Sure, but I’m just playing around with small quantized models on my laptop with integrated graphics and the RAM was insanely cheap. It just interests me what LLMs are capable of that can be run on such hardware. For example, llama 3.2 3B only needs about 3.5 GB of RAM, runs at about 10 tokens per second and while it’s in no way comparable to the LLMs that I use for my day to day tasks, it doesn’t seem to be that bad. Llama 3.1 8B runs at about half that speed, which is a bit slow, but still bearable. Anything bigger than that is too slow to be useful, but still interesting to try for comparison.
I’ve got an old desktop with a pretty decent GPU in it with 24 GB of VRAM, but it’s collecting dust. It’s noisy and power hungry (older generation dual socket Intel Xeon) and still incapable of running large LLMs without additional GPUs. Even if it were capable, I wouldn’t want it to be turned on all the time due to the noise and heat in my home office, so I’ve not even tried running anything on it yet.
The only time I can remember 16 GB not being sufficient for me is when I tried to run an LLM that required a tad more than 11 GB and I had just under 11 GB of memory available due to the other applications that were running.
I guess my usage is relatively lightweight. A browser with a maximum of about 100 open tabs, a terminal, a couple of other applications (some of them electron based) and sometimes a VM that I allocate maybe 4 GB to or something. And the occasional Age of Empires II DE, which even runs fine on my other laptop from 2016 with 16 GB of RAM in it. I still ordered 32 GB so I can play around with local LLMs a bit more.
I’m not going to defend Apple’s profit maximization strategy here, but I disagree. Most people won’t end up buying a cable and adaptare because they already have one, and in contrast to those pieces made of plastic and metal, the packaging is mostly made of paper. I’m pretty confident that the reduction in plastic and metal makes up for the extra packaging that’s produced for the minority that does buy a cable and/or adapter.
I would look into how Matrix handles this, for example. It involves unique device keys, device verification from a trusted device, and cross-signing. It’s not just some private key that’s spread around to random new devices where you lose track of.
They’ve implemented it in such a way that you only have access to an encrypted chat on a single device, so no syncing between devices. Syncing E2EE chats across devices is more difficult to pull off, but it’s definitely possible and other services do that by default.
“Uncensored”: https://x.com/KarlMaxxer/status/1823753493783699901. I don’t know if this is really true, but if it is, it’s something that they should’ve called out in their article.
Right, I see what you mean now. I misread your comment as explaining something that was already clear.
A false positive is when it incorrectly determines that a human written text is written by AI. While a detection rate of 99.9% sounds impressive, it’s not very reliable if it comes with a false positive rate of 20%.
“Our analysis shows that the two voices are similar but likely not identical,” Berisha said.
They also point out the main differences between the two voices in the paragraphs below this quote. I do believe that they hired a voice actress and that they didn’t train on SJ’s voice, or at least not entirely. However, I wouldn’t be surprised if there was big push for finding a voice similar to SJ’s voice in Her, no matter how much they deny this.
For me they only work in relatively quiet environments, or with earplugs. As soon as a car drives by it completely drowns out the sound. With music that might not be an issue, but with podcasts or calls it’s very annoying. I’ve bought earplugs especially for this, as my other earbuds have issues with wind while running, but it does feel like it’s defeating the purpose a bit. I guess turning them all the way up would also work, but that doesn’t feel healthy. Other than that I like them and the mic quality is also good according to people I’ve spoken with over the phone.
WSL 1 is a compatibility layer that lets Linux programs run on the Windows kernel by translating Linux system calls to Windows system calls, so in that sense I understand the name: it’s a Windows subsystem for Linux [compatibility]. It doesn’t use the Linux kernel at all. With WSL 2 they’re using a real Linux kernel in a virtual machine, so there the name doesn’t make much sense anymore.
And it does proper split DNS by default, using the search domains of each interface. That way you can configure a global DNS resolver while still being able to resolve local hostnames and without leaking other queries. I just hope they’ll also add DoH support, which is less likely to be blocked on a corporate network.
Not only that, but there are also two arrows on this map going into the UK. It’s up to the UK to take care of how well these arrows connect to important locations within their borders. There really doesn’t appear to be an issue whatsoever.
It’s off by default, but activated when you end your search query with a question mark. That option can be turned off.