I tried replicating this myself, and got no similar results. It took enough coaxing just to get the model to not specify existing tariffs, then to make it talk about entire nations instead of tariffs on specific sectors, then after that it mostly just did 10, 12, and 25% for most of the answers.
I have no doubt this is possible, but until I see some actual amount of proof, this is entirely hearsay.
These folks include presenting a false person as being of age, then switching to underage at the time of meetup when the target shows up.
I’ve never seen even a single instance in my own viewership of numerous channels that engage in pedophile hunting where the person is presented as being above the legal age of consent, then only switching to underage at the time of the meeting. They’re presented as underage from the get-go.
Then the group tries to kill the person
Again, this doesn’t seem to be a widespread thing compared to the number of them that simply lure them to a location then ask them questions (and directly state that they are free to leave at any time since they’re not law enforcement and can’t arrest them) The people you’re talking about are a small minority of both the actual number of pedo hunters, and the number of overall views received.
And the perpetrators think this is justice.
I doubt the people that are explicitly lying to farm content think it’s justice. I do believe the people actually catching people who voluntarily contacted someone presented as underage from the start do.
It depends on how these channels are going about finding their victims for it to be considered similar.
Remember, entrapment is based around luring someone to do something they otherwise would not have done had the operation to entrap them not occurred. If they created an account posing as a minor, then directly DM’d a person asking if they wanted to do x/y/z with a minor, that would be entrapment.
But if they made an account claiming to be a minor on social media, and the person contacted them voluntarily, asked their age, was told it was under 18 and still continued messaging, then sent explicit photos, that’s not entrapment.
However, if they were then the people who initiated the conversation about wanting the person to come to their house / visit them somewhere, that could be considered entrapment, and the only evidence against the person that could be eligible for use in court would be the explicit material they sent without being prompted.
It varies case-by-case, but from what I’ve seen, most of the larger operations tend to try and avoid entrapment-like tactics in most cases, where they only allow the other person to initiate unlawful behaviors, rather than prompting anything themselves.
couldn’t they just run down the registered sex offender list
The point of their channels is usually to find new predators that haven’t been caught, so they can then face legal consequences, (or at least be pushed to stop acting on their desires) rather than to punish people for being pedophiles in general, so it wouldn’t really make sense to go after those who were already convicted.
Exactly. Most things need to optimize for the lowest common denominator of understanding, and buttons with words and fields that have explicit purposes and positioning are a much easier starting point than “use command -help and figure out the syntax yourself,” even if someone who learns the syntax could then possibly be more efficient at using it.
Good job.
We cant even handle humans going psycho. Last thing I want is an AI losing its shit due from being overworked producing goblin tentacle porn and going full skynet judgement day.
That is simply not how “AI” models today are structured, and that is entirely a fabrication based on science fiction related media.
The series of matrix multiplication problems that an LLM is, and runs the tokens from a query through does not have the capability to be overworked, to know if it’s been used before (outside of its context window, which itself is just previous stored tokens added to the math problem), to change itself, or to arbitrarily access any system resources.
Here’s the key distinction:
This only makes AI models unreliable if they ignore “don’t scrape my site” requests. If they respect the requests of the sites they’re profiting from using the data from, then there’s no issue.
People want AI models to not be unreliable, but they also want them to operate with integrity in the first place, and not profit from people’s work who explicitly opt-out their work from training.
Here’s the same Obsidian Canvas document open in Obsidian, and Hi-Canvas: (*just realized the last connection is missing, that was user error while taking the screenshot, disregard)
They’re not fully cross compatible, but as another user mentioned, the open source spec being worked on is picking up steam as the Open Canvas Working Group (OCWG) and even larger industry canvas platforms are trying to make the format something they can easily import and export in that open format.
So hopefully you won’t have to worry about migration much longer :)
While that’s technically possible, it’s very difficult, and in my opinion, highly unlikely.
You know, offline features like seeing how long your device will stay on to tell you how long your device will stay on. Everything you need!
The only beneficial use I’ve had for “AI” (LLMs) has just been rewriting text, whether that be to re-explain a topic based on a source, or, for instance, sort and shorten/condense a list.
Everything other than that has been completely incorrect, unreadably long, context-lacking slop.
Yep, that’s what I have to use any time I need to look at a Reddit thread. Although the “fuck spez” comments that people replaced their old comments with definitely make the experience worse… as intended. Good for them.
Reddit has said it is working to convince logged-out users to create accounts as logged-in users, which are more lucrative for its business.
By “convince,” they mean:
Yes, very convincing, heavily coercive even!
It’s whole deal is “declarative” system configuration. Essentially, it means if your config file is identical to someone else’s, your systems will have identical software/dependencies, and everything should, in theory, run the same, generally speaking.
So for instance, to install a package, instead of running sudo apt install nameofpackage, I would just edit my NixOS config file to have the line pkgs.nameofpackage in it, run a command to “rebuild” my system using sudo nixos-rebuild switch, and it would automatically be installed.
That’s not the whole of it of course, but that’s just a general overview. It’s really good for if you’re running multiple systems that need the same software, because all you need to do is copy the config file over, run sudo nixos-rebuild switch, and the systems now have identical software.
Oh yeah, and you can also easily rollback. If you break anything, you can, while starting to boot NixOS, just select the previous configuration, start your system, and any changes you’d made to software/settings will just be undone. It’s great for troubleshooting.
AFAIK NixOS also has the largest number of supported packages out of any distro.
True, but I’m of the belief that we’ll probably see a continuation of the existing trend of building and improving upon existing models, rather than always starting entirely from scratch. For instance, you’ll almost always see nearly any newly released model talk about the performance of their Llama version, because it just produces better results when you combine it with the existing quality of Llama.
I think we’ll see a similar trend now, just with R1 variants instead of Llama variants being the primary new type used. It’s just fundamentally inefficient to start over from scratch every time, so it makes sense that newer iterations would be built directly on previous ones.
So are these techiques so novel and breaktrough?
The general concept, no. (it’s reinforcement learning, something that’s existed for ages)
The actual implementation, yes. (training a model to think using a separate XML section, reinforcing with the highest quality results from previous iterations using reinforcement learning that naturally pushes responses to the highest rewarded outputs) Most other companies just didn’t assume this would work as well as throwing more data at the problem.
This is actually how people believe some of OpenAI’s newest models were developed, but the difference is that OpenAI was under the impression that more data would be necessary for the improvements, and thus had to continue training the entire model with additional new information, and they also assumed that directly training in thinking times was the best route, instead of doing so via reinforcement learning. DeepSeek decided to simply scrap that part altogether and go solely for reinforcement learning.
Will we now have a burst of deepseek like models everywhere?
Probably, yes. Companies and researchers are already beginning to use this same methodology. Here’s a writeup about S1, a model that performs up to 27% better than OpenAI’s best model. S1 used Supervised Fine Tuning, and did something so basic, that people hadn’t previously thought to try it: Just making the model think longer by modifying terminating XML tags.
This was released days after R1, based on R1’s initial premise, and creates better quality responses. Oh, and of course, it cost $6 to train.
So yes, I think it’s highly probable that we see a burst of new models, or at least improvements to existing ones. (Nobody has a very good reason to make a whole new model of a different name/type when they can simply improve the one they’re already using and have implemented)
Nix is definitely where I’d bet on good GUI development, if I had to choose. I’ve had my own issues with NixOS in the past, primarily due to me simply not being able to fully grasp some of the language for configs, but I’d definitely assume that GUI editing would be much easier due to the more predictable nature of the config files.
This is what I think would be most probable too. It’s still too socially stigmatized to outright say something that effectively just means “I’m a Nazi,” so they’d have to play it off as some kind of joke.
I do, however, think a lot of them would suddenly, for no reason in particular, begin saying things like “Well, Hitler was a bad guy, but he did have some good ideas…” while going on to describe some of the worst ideas known to man.
I’ve seen no evidence that these kinds of traits are inherently biological.
Regardless of the fact that we have significant evidence that these more “new” forms of masculinity that incorporate less domineering and aggressive mannerisms are beneficial to men, I simply haven’t seen any evidence that these traits are biological.
In the same way that when you don’t socialize a child to prefer certain clothes or toys, (or stigmatize against them) they generally just go with what they prefer in the moment along lines that don’t match the gender binary, from what I’ve seen, the same is generally true for behaviors. We’re heavily influenced by our cultures and by extension, our upbringing, to a degree that explains why these mannerisms are commonly expressed along gendered lines.