I’m way more worried about where the energy is coming from and what the true cost of storage is, rather than where I get it from. Every conversion/storage has an energy and materials cost. As bad as petrol burning is, I have to imagine coal burning + transfer loss has to be about as bad. Not to mention the nature of lithium cells.
We don’t need more charging stations to make EV viable, we need more nuclear power plants and cleaner battery tech first.
“[an] integrated vehicle system that uses, at minimum, the GPS location of the vehicle compared with a database of posted speed limits, to determine the speed limit, and utilizes a brief, one-time visual and audio signal to alert the driver each time they exceed the speed limit by more than 10 miles per hour.”
Honestly the only part of this that is unreasonable is that it isn’t immediately followed with “the database updates will be maintained and provided in an open, unencrypted format for free for the life of the vehicle, and the tracking data cannot be used for any other purpose”. GPS is a one-way, triangulation-based signal. It doesn’t inherently track or leak anything. I think we would be a lot safer if we all could agree what speed to go.
I am not joking lol but I do sometimes forget most people don’t live in this space the same way I do. I think people use these names because the programs themselves are forked often and the software names are very unspecific otherwise. I meant to imply that I was using the main branches of these softwares.
I have this running at home on a used r630 (CPU only). oobabooga/automatic1111 for LLM/SD backends, vosk + mimic3 for tts/stt. A little bit of custom python to tie it all together. I certainly don’t have latency as low as theirs, but it’s definitely conversational when my sentences are short enough.
Yeah I’m really curious what his take is going to be on this one lol. Technically it doesn’t have a layer-2 capable bridge mode like other VPN solutions like openvpn, but that’s about all I can think of. It’s still objectively a virtual network, made private by a keypair exchange.
Probably just blindly paroting something someone told him. Awkward way to learn that one lmao.
unless the bill has changed since the last time I read it, there were fines for hosting the service in US datacenters, and fines for companies allowing US data to exist in non-us datacenters. I don’t think you could interpret the bill as imposing a civil penalty to a user using a vpn and accessing it.
costs only an email address and a promise to sign up for a 37% APY credit card.
You would expose a single port to multiple vlans, and then bind multiple addresses to that single physical connected interface. Each service would then bind itself to the appropriate address, rather than “*”
I thought tiktok came out of music.ly? I didn’t think it had roots in vine.
Always happy to try and productively add to someone’s learning.
I guess I just misunderstood what you were arguing then. For posterity: I believe datasets containing children is fine, datasets containing csam is not, and the legality of generating csam should be left up to psychologists on whether or not it is a societal net benefit. Whichever way is better for children that exist is my vote.
I’m not arguing whether or not it should be legal, I was just offering my first hand experience in regards to the capabilities of these local models since people seem to be confused as to how this actually works.
That would be true, it’d be pretty difficult to build a model without any pictures of children at all, and then try and describe to the model how to alter an adult to make a child. Is anyone asking for that though? To make it illegal to have regular pictures of children in these datasets?
I’m not going to say that csam in training sets isn’t a problem. However, even if you remove it, the model remains largely the same, and its capabilities remain functionally identical.
Movie theaters are a great analogy to what streaming services should be
Yeah, turns out when the monopolies are eliminated, people get more competition and a better deal on the consumer end. It’s why I’ll never understand people who say streaming services became as bad as cable.
I’d argue that streaming is in such a bad place right now because each streaming service has a monopoly on their own content. Sure, you could argue that studios “compete” with each other on the content they produce, but I’d argue that cable companies were a different layer of the stack entirely. Cable companies all offered the same channels and the same content, and in areas where they did overlap, competition to offer the best delivery of those channels was great. What made cable bad was that there was little incentive for companies to geographically compete. In the era of streaming, companies have little incentive to allow their content to compete across platforms.
If you ask me, every streaming platform should be broken up from their production parents, so that streaming companies can compete on what they offer, and how they deliver it. There is no incentive for the platforms themselves to compete with each other. It’s all about how hard the services can enshittify before people stop watching the content they have a monopoly on.
You should consider reversing the roles. There’s no reason your homelab cannot be the client, and have your vps be the server. Once the wireguard virtual network exists, network traffic doesn’t really care which was the client and which was the server. Saves you from opening a port to attackers on your home network.
It doesn’t need csam data for training, it just needs to know what a boob looks like, and what a child looks like. I run some sdxl-based models at home and I’ve observed it can be difficult to avoid more often than you’d think. There are keywords in porn that blend the lines across datasets (“teen”, “petite”, “young”, “small” etc). The word “girl” in particular I’ve found that if you add that to basically any porn prompt gives you a small chance of inadvertently creating the undesirable. You have to be really careful and use words like “woman”, “adult”, etc instead to convince your image model not to make things that look like children. If you’ve ever wondered why internet-based porn generators are on super heavy guardrails, this is why.
There is also the argument that it’s more complicated under the hood and harder to troubleshoot, particularly because of it’s inherent parallelism and dependency-tree design, whereas initv was inherently serial. It was much more straightforward to pick the order in which services started and shut down on an initv system.
For example, say I write a service and I want it to always be the first service stopped during a shutdown, and I want all other services to wait for it to stop before shutting down. That was trivial to do on an initv system, it’s basically impossible on systemd.
For those wondering, yes I did run into this situation. My solution was clobbering the shutdown, poweroff, and restart binaries with scripts earlier in path search that stop my service, verify that they’re stopped, and then hook back to systemd to do the power event.
Okay, I’m hooked, I have to know the non-clickbait story