No, the prompt won’t respond to non keyboard input devices. You have to click with a mouse or keyboard.

There should really be a timeout confirmation, you know, like most prompts of this sort. It’s a pretty big design flaw imo.

Yeah that’s definitely annoying… Maybe it will be a delayed release?

It probably doesn’t help that there’s so many simultaneous Rumiko shows going on lately. It feels like we went from 5mph to 100mph in terms of output of her stuff lately.

If you think about AI systems as effectively complex DSP problems and equations, then logically any system that takes inputs that are potentially the outputs can cause system feedback or recursive (destructive) loops. What scares AI companies is that, while most recursive loops are easy to detect immediately, “content loops” will be much harder to detect as the delay time between inputs is much larger compared to, say, audio or programming loops where feedback is obvious immediately.

This is effectively the theory behind the practice of data poising, and it’s hard to say there’s no validity to it as most AI companies are terrified of data poisoning. If it didn’t work, companies wouldn’t be so adamantly vocal about their distaste for model poisoning conceptually. Also, a lot of time and money is spent trying to “detect” AI content for a reason – that reason is actually to help aid the detection of AI output which must be “valuable” to the companies to spend the resources on it.

Conversely, AI makers have learned of ways to avoid this by simply having human semantic “grading” of the content done by third parties. This is why there are so many deals going on in Africa / SE Asia where these AI companies are hiring English speakers to effectively “wash” the input by giving it contextual “extra information” and rough validation scoring. This is an expensive solution, though, so they’re very much dependent on AI being the bees-knees of lucrative investment for this process to continue. I’d also argue, with the rate at which AI development has slowed down, the semantic grading of content being fed into the system also has diminishing returns. However, this is effectively a “survival of the fittest” style evolutionary simulation, where the AI is only interested in training off information it happens to find is “right” or “close enough” or whatever metric the grader finds. The feedback is less of a problem if the validity of the input can be assured or “cleaned up” to prevent unintended loops, basically.

Now, “are the programs that claim to poison the datasets effective?” Hmm, that’s a difficult one to answer. Personally, I have some skepticism around these models as their origins are vague and most are not adopting an “open data” approach or even an open binary approach (freeware) for distribution. I understand that the concern from the makers is that publicly talking about how the sausage is made makes the software less effective, but it’s hard to validate that the people behind these models are providing the service as intended and that they aren’t doing anything with the data being sent to them for “protection.” There’s no assurances that they aren’t training models off the data artists send in themselves, for example, or any guarantees to how that data will be used for training. So it’s kind of a “miss” for me, unless there’s a project someone is aware of that is both open-source and open-data (I find that ‘open-source’ in the AI field is a hugely misleading moniker, as AI follows a “data is king” philosophy and the program that trains the models is inherently less important as a result.)

I do think it’s bizarre we live in a society where smoking tobacco or drinking alcohol is “normal” yet weed is somehow a taboo subject.

It’s especially funny since Japanese policy on weed was basically forced on Japan (and the rest of the world) by the United States with racially motivated trade policies that conveniently helped the alcohol business thrive in a post prohibition world.

Wow I never knew that Great Ace Attorney was used for Anti-weed propaganda. That’s crazy (and probably a little bit niche?)

I really love the art parallel’s this manga wears on it’s sleeve. The fact that Coco’s solution to being inadequate compared to others is to embrace the parts of her that make her “unique” and do what she knows in a new and creative way is so apt to the artistic process.

It will show up, eventually. I wouldn’t be surprised if they move it to the start of the next episode as a gag.

Damn,

Someone really wants to know how much K-ON is consumed and by who. It’s the ultimate kompromat – you’ll be too embarrassed to ever admit the truth.

Really love the manga so I was excited for the show.

I’m a bit concerned that the pacing of the TV show is a bit slow and would have preferred that the Mother/Daughter interaction was handled a little bit more abridged compared to the manga. Comic pacing can generally be a bit more punctual despite covering a lot more due to pacing being “flexible” based on reader engagement, if that makes sense.

Overall though a very faithful adaptation. But I do worry that some people might bounce off the show if it doesn’t “get to the point” fast enough, I’ve seen people seemingly thinking this will be a magical girl slice of life show and it definitely isn’t that…

How will this be achieved? I’d be curious to see if a really rudimentary estimation is used based on GPU benchmarks internally (like an estimation kernel trained of Valve’s internal benchmarking of machines against the steam machine, for example.)

Nice, might like to check it out myself now.

How have people felt about this show, long term?

I’ve been putting off watching it because I’m super into the old show and not sure this can live up – but maybe I’m willing to give it a shot as the old show definitely had it’s fair share of problems compared to the original manga release.

Good shit to him. He’s 100% right here.

6 year old is a bit of a tough territory here for anime recommendations, stuff like Rumiko Takahashi works (Ranma 1/2 and Inuyasha) almost seem like they’d be fine but they probably tackle slightly too mature subject matter and would be better in early teens age.

This is a bit of an old school recommendation, but you might enjoy checking out the original Astroboy, old Tatsunoko shows and stuff like that. The older anime tends to skew a little less “mature”. However, sensibilities of what is ok for kids in America vs Japan always makes this difficult. For example, Detective Conan is a great kids anime by Japanese standards, but talking about murders and kidnappings might be a little much by American standards. So YMMV.

Obviously Pokemon is a good suggestion here. The original Digimon Movies are also pretty good and are done by a great director (Mamoru Hosoda) so you might like to check it out.

Well, I mean, good. Because the old policy fucking sucked for artists.

That is interesting, I’ll make a note of it.

but that kind of separate intranet system is difficult to search the way the Internet is.

Yes, but node operators would probably be best encouraged to run simple indexer searches, especially in a system where agreements could be made about how hard-hitting said searches should be (and a mutual respect for robots.txt and the like). I’ll be honest and say that the Internet has been barely searchable for at least half a decade, if not a whole decade if I were critical.

I think that technology like Tailscale has sold me on the concept of on-internet intranets, as in subnets with extreme firewall policies that doesn’t prevent you from accessing the broader net when necessary but gives network maintainers strict control on how their networks are bridged. I’ve been thinking about this to the degree that I’ve been trying to do more research into how this can be achieved with open source technologies like Headscale.

Ideally, you’d want to have a peer-to-peer relay server option for bridging multiple “trusted” networks which would then provide a broad DNS resolution to let you access services that are advertised for bridged networks. So it would be like if, via tailscale, I could connect to another person’s tailnet using specific domain names if those services were exposed via a “bridge node”, so to speak.

Tailscale themselves have no reason to implement this though; As a business, they would actually prefer you buy larger client counts. I don’t blame them for this, it’s the basis of their business. But I think, long term, multiple intranets will be really important for digital sovereignty for both smaller nation states and individuals. We can no longer trust the broad web as it was. The fediverse is the first step, the next is tighter meta-networks in tandem with federated internet services.

Can’t read the whole article because of paywall.

But the implication that AI is a “good bubble” is pretty rich, as I think that the degree at which they’re ripping people off left-right-and-center by stealing their information will become less palatable to the US government once it’s clear that they caused the market to tumble. They’re literally dependent on the idea that people think AI is “irreplaceable” to the daily lives of most people in order to withstand a crash in terms of political capital, which I think they’re very far from establishing at this point.

I like the idea that they think that educated jobs only belong to women. That’s an interesting thought.

The sad truth is that this shit isn’t going to replace “highly educated” jobs, and that the AI gravy train will end once people start to enforce basic intellectual property enforcement. Time is ticking, and the market taking a hit now is making them scramble.

“This is an Xbox” now replaced with “Nothing is an Xbox.” 😈