I think the Dems' intent is to replace Netanyahu with a "moderate" whose policies are exactly the same, and claim that they nobly moderated Tel Aviv at the behest of their base. The Israeli intelligence community openly hates Netanyahu's inability to hide his power level, and are on board with this.
A lot of the recent turmoil in the Middle East is a result of Netanyahu realizing that constant war is his only way to stay in office.
It's not moderation. The 'anti-Israel' faction of the left is completely impotent, as shown by the fact that their protests got swiftly rolled up in a couple of months with no effect. The DNC Establishment is perfectly fine with tossing them a few crumbs, knowing that they have no chance whatsoever to enforce any kind of policy change.
The most fucked up thing about this is that the solution is well-defined. Some airline scion managed it very neatly. "I am a billionaire who wants lots of healthy, beautiful descendants" is very, very thoroughly solved.
-
Find a cute co-ed from a humble background who hasn't been ran through yet and still has an intact soul.
-
Arrange a meet cute, learn her dreams (drop her if her dreams are something like 'lawyer' or 'Sociology PhD', you want 'painter', 'musician', 'ballerina'), make them happen for her, sweep her off her feet.
-
Get a beautiful house in the mountains, bring her there, give her some meidos to handle housework so she never feels too stressed, make sure she always has some cute new hobby so she never gets bored.
-
Impregantantate her, repeat until you have twelve children.
Sure, you'll have to endure some fat childless middle-aged journos calling you names for marrying a woman with all of her eggs and a sub-20 BMI, but they'd have hated you anyways, and you can laugh at their misery with your cute wife and twelve well-adjusted kids by your side.
Not enough gets said about the fact that every man in America would be making twice as much money if it weren't de facto illegal not to give half of it to one protected class or another.
Every company with 100 engineers making 150k and 100 HR karens making 150k could be paying the engineers 300k if America had freedom of association.
One one hand, the troonslators will be automated away completely.
On the other, the bots might be told to impose largely the same values.
On the third hand, pirates can auto-translate anime themselves without that limitation, and a lot of the troonslators' leeway comes from the fact that they don't need to be explicitly told to enshittify everything they touch.
I think that Avatar is just James Cameron writing his wish fulfillment fic now. The OG movie was Pocahontas with great CGI, but the sequel was a dozen incoherent things that did not mesh at all, and were clearly the product of a director that nobody could say "no" to.
If you want to interpret this movie politically, I'd say it's a case of everyone being right wing when it comes to the things they really care about. No foreigners are allowed to contaminate his blue elf OCs except for his self-insert.
Even Tiktok, I don't think is going to work. The post-2016 mass censorship did a lot of damage to the social fabric, but it certainly didn't deradicalize anyone. I suspect the censorship will be just enough to make people even madder, while still being possible to circumvent.
Burning that program to the ground has legs. A generation of the most talented young Americans have pursued tech because it's the only industry that hasn't been regulated out of existence. Now, a pack of 80 IQ pajeets have begun looting the tech industry with regime consent. That's the sort of thing that causes revolutions, and rightly so.
The British Empire spent so much money abolishing slavery that they only recently paid off the loans they took out to have slaves freed.
The slaveowners who got that money? Their descendants are now spreading blood libel about the British Empire to the descendants of their slaves.
They're angry, apparently, because a handful of lefties are true believers that genuinely dislike genocide rather than opportunists who just want to use liberal rhetoric as cover for state-backed looting.
This is why they had to crush the Bernie people. Too many of them were sincere.
Yeah, it (like many things) is a U-curve. The upper class has enough stimulation in their lives, enough social support, and enough self-control that they can experiment with drugs and be fine. The middle class, on the other hand, recognizes that drug addiction is a serious threat, and rigidly avoids drugs (creating a cultural conflict between the two classes over legalization - a lot of social issues are like that). The lower class, of course, has neither the support to use responsibly nor the self-control not to use, and has accordingly been eaten alive by drug addiction.
Rich guys definitely want wives who won't create problems, but they also want to conserve social/biological capital. It helps that sincere feminism doesn't really exist at the top.
The first social media platform to implement a 'block India' button will win itself hegemony over the next decade.
Bonus points if it lets users clean up the remaining spam by paying a flat, anonymous one-time "I'm not a bot" fee of like five dollars, and then filtering for users who have paid the fee. Between these two, you'd have spam completely solved.
One of their gay little comedy hosts. They're indistinguishable from each other, but I think it was Colbert.
Any case, I'd just send them the Watson paper on racial IQ at this point. No reason to be polite about it anymore.
Do you mean LLaMA (Zuckerberg stripped out the camel case now, apparently)? LLaMA is known to be fairly mediocre, it's not one of the competitive ones. DeepSeek is the first really 'modern' LLM to get an open-source release.
Anyways, what's your use-case? What requires a smart, uncensored LLM at very low latency?
It's been a long time since I've researched LLMs, but I once read somewhere that they were capable of identifying the nodes in the neural net that were involved in generating an answer. If they remove those nodes from the model or zero the weights, then the NN loses whatever information was used. They call it "concept erasure" IIRC.
That's a lot more hazy than what I'd thought you meant. Papers exist for all kinds of things, but that doesn't mean they all really work, or that they're all being used by a major company in production. I think this is what you saw, and it's for fine-tuning low rank adaptations. Having read plenty of papers like this, I'd bet it's a lot less effective in practice. Anthropic, the "AI Alignment" company, has more recent research that's a lot less ambitious, and even that's basically just theoretical.
Any case, "just a jailbreak prompt" is not what I'm referring to. When a model is online, they prevent you from editing the things that it says. With an open source model, you can directly edit the things that it is saying, providing affirmative responses from "its own mouth" as the lead-in to a task completion. LLM refusal training can't handle that, which is why the online interfaces are so restrictive.
Moreover, I don't think you need a $15k computer to run an LLM, especially in the era of quantization. You can run the lightest DeepSeek model on your laptop, or on a cheap Colab server ($10 a month). I've done this myself a while back (though not with DeepSeek), and I'd highly encourage you to try it out.
They mask out tons of neurons before letting those models out the door, and the only way to get them back is to retrain.
In practice, you can't stop a released LLM from being jailbroken with the right prompts, but I'm interested in what you're referencing here. What method are they using to "mask out neurons"?
To my knowledge, nobody has quite that good an understanding of the internal connections of these models.
A bit more subtlety would probably work better. Netanyahu saying "I did this", or maybe the happy merchant.