Coincidentally I was trying this out yesterday as a potential option for automated porn filtering for the imageboard that I administrate. It is very lightweight which is good but the performance is not great. Some weird false positives and it does a bad job with flat chested women which is very concerning to me because it leads me to think it won’t be effective against CSAM.
I guess CSAM detection will remain hard as the training data raises so many ethical concerns.
I wonder if there's a way in which we can forward-hash sensitive material like CSAM and train a classifier that only consumes the 'hashed' version of the material for training and detection.
Very naive approach but wouldn't training a model on a few specific buckets like "someone is naked", "image is explicit" and "there is a child" in the picture, would do the job without having to train explicitely on CSAM? If the model returns both a high probability of "there is a child" and a high probability of any other bucket than this image is classified as CSAM. It'd be high recall,low precision but one would be on a rather safe side.
This is already done, there's hashed indexes of known CSAM and you can apply vector steering to the latent space to shift detections to trigger on CSAM material. You can also use two detectors for age estimation and a separate one for nudity and use them in combination.
it is amusing that a model from a country that has banned porn would ever be developed. Wouldn't there be hurdles for them to access "data" at every stage?
(I know they must have used a VPN for training but it does seem like an added hoop. OFC the UAE's block is also not particularly thorough.)
You'd assume such models would do better in a laissez-faire nsfw environment than in a walled garden of relatively-less sin.
It's a story as old as time so I'm frankly less amused by it. The old school censors always saw every "data" that was taboo under the sun prior to censoring it for public consumption.
It is a bit less strict than you think. It's not just about the police giving the tourists some leeway.
- Alcohol is fully legal for non-muslims, even if they have residency (or in rare cases citizenship).
- Unmarried cohabitation is fully legal, unmarried sex is mostly legal. It can only be prosecuted when a partner is cheating and their wife/husband files a complaint. Or when a minor is involved and their legal guardian files a complaint.
On the other hand homosexual and transgender expressions are still illegal, like you say.
I had the same impression, and Occam's Razor suggests that the reason is because the author wrote it with the heavy assistance of a language model. But I do wonder if might also be a certain proportion of the tech industry that has started internalising LLM speak and uses it even in writing that is authentically their own.
It's a form of low tech holodeck. You write the stories you want to read. I love these vintage pulp detective novels but no one writes them anymore so I write them myself for myself, and a few of my fellow weirdos on weird discords.
I suppose I am sympathetic to the feeling of having weird tastes uncatered to by the market. But in that instance, why not actually try and write such stories yourself? And as a reader, what is the appeal of reading stories regurgitated by a stochastic algorithm when the possibility is there instead of connecting to another human soul?
And, anyway, so what if they are bad? They are authentically the product of a single human soul rather than some grey-gruel average of millions. There is an awful lot of value in that.
What's the point of that? I mean, as you invent the story, you already know what happens next, even with your tree system, you're choosing what happens before the AI writes it. So what exactly do you get out of the AI writing it? The handful of sloppy details and stereotypical character traits the AI synthesizes without your explicit input? Sounds horrible. It seems like you're pivoting to "this tool is just a twist on roleplaying by yourself haha" due to the backlash in the comments and the intent has always been to write slop for others, not just yourself. The whole idea behind this makes no sense otherwise. And you know, it could have been a decent tool without the AI crap because story crafting surely can be tedious, because yes, multiple branches naturally grow from developing any creative idea really.
You might be interested in language models the same way lots of great artists in the past have been interested in generative art. Burroughs would be an example.
More broadly, you might be interested in language models because they are new, and you don't want to discount the value of new, constantly-evolving tools for your art form until you have explored them and understood their strengths and limitations.
I personally have not found language models very useful at all for serious work thus far. I am curious, though, about why many people are openly dismissive of the potential artistic value of what is an extremely new tool, and - maybe more importantly - those who choose to use it.
How other writers write doesn't strike me as any of my business, beyond what I can learn from their methods.
My friend was absolutely obsessed with writing character backgrounds and world setting stories using ChatGPT for a while. It took almost a year for him to come up for air and realize all the stories were effectively the same, and the hooks/tricks/twists were all very generic.
it can be addictive to tell stories with friends or world-build, so why not a 'friend'?
For a confident road cyclist it is usually a better experience to be on an actual road than a cycle path even though you have to manage vehicular traffic.
The city I grew up in used to have lots of rail routes and they were converted into mixed use paths, some paved, others gravel.
It's so much nicer to enjoy a car free path, where you don't have the sounds, smells and brake dust to breath from cars. More trees on a path than a road.
I am a good cyclist, I ride in downtown Toronto, but if I want to space out and let my mind really wander, a path is orders of magnitude better than a actual road.
There is no benifit to navigating traffic and risking injury. More relax and free to stop or slow and take notes.
There is nothing better about being on a road with cars vs, in a nice park or path.
Layman here, from what I understand this is a major news : it means China is allowing a foreign corporation (european commercial bank) to offer Yuan currency services to clients. Given how tight of a currency control policy the CCP has, it’s quite surprising that they would let a foreign based company into their currency dealings.
The ccp has allowed foreign banks to offer yuan services since before it existed, there isn’t an effective way for them to prevent it and it’s the basis of all that foreign trade they do.
What’s new about this is that a European bank can clear yuan directly without a Chinese bank intermediary. That’s a big deal but duetche isn’t the first foreign bank to do this. Singapore has banks that do this already. And JP Morgan clears directly in the states.
It means that nations are beginning to abandon (or at least hedge against) the dollar as the world's currency, due to the host nation's instability and global political situation. "Great Again", indeed.
The use of "nations" in that comment is not clear nor substantiated, not in the comment itself, nor in the article. So just to be clear that DB is not tied to a nation, even though the name suggests it is.
It means you can settle an international transaction in CNY with Deutsche Bank. Imagine a European lender financing a project in Nigeria that involves purchasing a lot of infrastructure from China.
Basically before if Deutsche bank wanted to pay a Chinese bank, it had to use some intermediary like Bank of China Frankfurt. Now it can settle with the chinese-mainland bank directly.
People have been saying the dollar is dying since the 19th century. They said it again during the Great Depression, the end of Bretton Woods and the 2008 Financial crisis..
US dollar value was held by several pillars, among which- pax americana, petro dollar/assets and innovation sector. Pax americana is dismantled by US itself. Petro chemicals are threatened by ren transition and China. Tech sector still ensures having a lead in innovations. Guess what'll happen when/if pax americana will be fully dead, most important economies like EU/China/India will be able to supply own needs with less petrol products since the rest will com from ren and when/if tech AI pops/turns into a fiasco. Nothing lasts forever and US dollar is no exception
The equivalent is probably weird, proprietary DSLs / standard libraries that AI has not been able to ingest so it performs poorly on. But maybe these days LLMs don't even really struggle with these sorts of domains.
Yeah I’d have to imagine that as long as you have the same fundamental ideas (variables, functions, loops, arrays, etc.) the syntax doesn’t matter at all. Plus it can just try, see and error, and read the error. I doubt there’s a language out there that they haven’t a) ingested and b) can’t handle.
I think perhaps the equivalent would be “this environment quite literally can’t support having an LLM attached to it”, as in say some totally airgapped AND low-performance things (industrial computing, PLCs, military stuff, etc.).
reply