deleted by creator

Piracy is a loosely defined term that in modern times involves the illegal taking of copyrighted material. This situation is about whether people, by distilling models from the AI companies, are illegally taking copyrighted material when the AI companies themselves took and used, arguably illegally, copyrighted material from the original authors.

As an IP attorney and engineer that practices right in this field, I have a lot of complex thoughts. First, copyright term in the modern US is way too long (thanks to Disney lobbying Congress a bunch). Nonetheless, I think that the manner and size of the taking of copyrighted material to train current AI models is way outside the bounds of fair use, despite what most courts are saying. However, in my opinion, the models are not really derivative works in the manner they exist now. Rather, the models as a whole are not even copyrightable material under the current law because the companies lack sufficient control in the creation of the output of the models. So at best, the models are protected by trade secret law. And if people can distill those models, then that is fine because trade secrets don’t protect derivative works, only things that are confidential. It would be like KFC saying you can’t try to figure out how we made this food by eating it and taking the food apart and tasting it and testing it. And, again in my opinion, contract law (i.e., the terms of use for using the models) also does not provide the companies a viable claim because of a weird section of copyright law that basically doesn’t allow state law claims, such as contract claims, to be brought when they are in effect basically copyright claims.

If everyone followed my interpretation of the law, original authors would get royalties for copyright infringement by the AI companies, the AI companies would only be able to stop distillation of their models by engineering a way to stop distillation, and people distilling such models would have no liability at all unless they illegally took the model weights and otherwise from the company in violation of trade secret laws.

But that’s not how this will turn out. Courts will, and already have, called this fair use. And even if it wasn’t fair use, courts will take (what should be a rare step reserved for congress) consideration of public policy, and that includes national security, and say that national security mandates that we as a country just accept the status quo for the good of our country and laws. While I agree that this technology poses a uniquely dangerous national security risk, if we’re being real, its from the corporations, not China or other countries at this time.

But hey, I’m just a lawyer that manages a massive AI, robotics, and cyber security IP portfolio for one of the leading universities in the world in these fields. The judges and politicians that make these decisions generally do not listen to me.

Normal people think you are joking, but as hiker trash I’ve been doing the same for years and I love it. I usually add some cheese, an olive oil pouch, salt, pepper, and maybe some dehydrated veggies and a cut up summer sausage if I’m really hungry. It makes for an excellent dish.

I hope the study specifically publishes the data on the Model 3. I bought one before I became much more privacy conscious, but I’ve done my best to mitigate the leakage from that. In doing so, I read through all of Tesla’s terms and conditions upon opt out and sending any of my information or information that can be used to identify or profile me, should never be sold to advertisers. I write software licenses for a living so I’m pretty confident in my reading on this. It would be interesting to see what can currently be mitigated by changing some settings and what cannot.

I wouldn’t say it is too far out of date, having just gone through all of it in the past few months. That said, Bazzell does provide a lot of smaller updates in his Unredacted magazine articles. I believe he said he is no longer planning to update the book because he couldn’t afford to take the time since the book was being torrented a bit too much (I might be misrembering). Either way, his magazine is really good too. I became a big fan of his after that first book.

For anyone unfamiliar, I strongly recommend Michael Bazzell’s book Extreme Privacy for a fantastic entry into this world. It covers much more than your typical OPSEC stuff, but it is thorough, well written, and fun to learn about.

I just realized that the guy selling all of these surveillance cameras has the same name as the CIA headquarters. The writers of this reality are too damn lazy.

I had a good friend slip in the shower, hit his head, and die. He was 27. Life can change abruptly in ways you would never expect. Be careful out there please.

deleted by creator

deleted by creator

deleted by creator

Removed by mod

deleted by creator

deleted by creator

deleted by creator

Referral Code: 8DHE3HVV

I have Cape and they have been great so far. I’ll leave some info below. Skip the first paragraph if you want to avoid their advertising scheme, but it was the big reason I could afford to switch to them so I thought I might be helpful to others.

They are currently running an early adopter promotion until the end of the year where you get an unlimited plan that is $70 per month for life, and if you use someone else’s referral Code it drops you down to $50/month for as long as that person stays on Cape. What’s crazy is that it then drops to $30 for another referral, $10 after another, and $0 after that. For as long as those other people are on. I thought that was a pretty cool deal honestly given that I was previously paying $55 a month for T Mobile.

So a quick review from my experiences so far.

Sign up was easy and basically needed no identifying information. I use Cape on Graphene OS through the stock messaging app. It allows me to send videos, images, and so on. There are some issues with group chats still I think, but I haven’t really fiddled with that. The coverage and speeds have been solid.

The IMSI rotation is automatic and a great protocol. They have a bunch of other very smart architecture choices. I used to write patents for telecomm companies like Ericsson and Qualcomm and I’m really impressed with their engineering. My friend is an independent security pentester and educator, and their company recently audited Cape. They said they did excellently and had no reservations on recommending them to me. I also watched a lot of interviews about the company and their system. After that, I felt comfortable testing Cape out. My hope was that they would be way more reliable approach than my previous set up (I followed the Bazzell set up with Mint under a pseudonym and VoIP.ms but voip.ms has been really spotty lately and I have young kids so that’s no good).

I still have VoIP.ms for my old number, but I use my main Cape cell number for close family and friends who won’t get on Signal for whatever reason. They also give you two free burner numbers for texting only, so I use those for 2FA and sign ups if needed. You can only text to those numbers and you currently can’t port them out, but I believe they are middle to end encrypted (check me on that one though).

There are many other features and caveats (like they’re working on RCS, they have different available features for iPhone users, etc.), but I recommend their interviews with channels like the hated one or techlore or elsewhere. Those helped me a lot.

I’m happy to answer any other questions. It took me a while to decide to try them out because I’m pretty tedious about my privacy set up, but I’ve honestly been pretty impressed so far and I love the idea of having a mobile core telecomm company adopt some actual privacy architecture in this middle space.

Edit: I just reread my post and realized how much I sound like a shill and how much I repeated what OP said. Sorry, I’m just high in my bed.

deleted by creator

deleted by creator

deleted by creator