Post

I asked claude and it agrees.

Ah, the nazi running it finally nazi-fied it's product.

"A generalization made of a single sample, especially about changeable components of complex systems, is foolishness."

What will "AI ethics" look like? I would hope that an older generation of computer scientists would weigh in.

I'd not mind an LLM that can reliably quantify its own certitude of inferences. Its own internal mechanism might be able to flag a likely hallucination.

Thank you for nailing this.

No, no. Everything is fine. You just have to be sure to only use the LLMs that aren't being manipulated by their billionaire owners.

Musk is doing the world a favour by demonstrating exactly why LLMs cannot be trusted as information sources.

LLMs can't even get Linear Algebra questions correct - they are made of LA. I consider LLMs like creating Broca's region with not much else. This reminds me of Deep Blue AI claims 30 years ago, that never panned out. We have a ways to go.

Honestly they shouldn't become really anything of anything. The fact that there's a humongous world-wide AI bubble growing without apparent limits and gulping up parts of economy feels like an apocalyptic threat.

Related— Everyone should read this stunning new book. Artfully articulated, detailed, deeply researched, exquisitely organized, beautifully constructed. Exponentially insightful. EMPIRE OF AI by Karen Hao Essential. Consequential. Riveting. Don’t miss Ch. 9, 12

Can’t wait til they use that “right wing” nuance from grok to hunt for dissenting opinions. None of the remaining DOGE people are incel fascists, right?

I think they just want to purge opposition from the power apparatus of the state. They don't care about the citizenry yet.

Grok is a good example that LLMs are only as "unbiased" as their creators. Everyone else at least pretend to be neutral/unbiased

Doing great bsky.app/profile/rwei...

Thank you for explaining that more clearly and more politely than I could figure out how to.

“Replicants are like any other machine: they’re either a benefit or a hazard. When they’re a benefit, it’s not my problem.”

We need to endlessly repeat that AI is a terrible thing in all ways.

Seems like the kind of thing someone would do if they were trying to make people distrust LLMs and burn them to the ground. Except they probably fail and start pogroms. I had thought social media algorithms were bad enough on their own.

It’s already polluting the information ecosystem with its bullshit 🤬

Great point. Also why it would just exacerbate epistemic injustices. And ‘faking it’ for many would be just going along with ruling ideas, with a little more efficiency that genAI offers. Efficient bullshitting which will take way more energy to get rid of.

Absolutely #humansarebroken enough already

If he does this with his LLM, imagine what crazy things he will do with Neuralink.

I understand, also, that this is true. Grim.

Related— Everyone should read this stunning new book. Artfully articulated, detailed, deeply researched, exquisitely organized, beautifully constructed. Exponentially insightful. EMPIRE OF AI by Karen Hao Essential. Consequential. Riveting. Don’t miss Ch. 9, 12

Surely @wired.com is all over this fuckery … and hopefully when its story comes, MSM will take heed

#MechaHitler

But the generalization to "all Jews" is going to be monstrously unfair. I hope that's the point.

When society knows what the word epistemic means there might be a chance of preventing it.

Sadly, so many books are being thrown away from Uni libraries. For instance, our campus recently consolidated with a local community college and so many books were thrown away. I tried to snag as many as possible, but that was technically not okay since they’re state property.

And here it is: look at this book for sale. Nobody would know it was fake. bsky.app/profile/comp...

It's just quoting Defense Minister Katz: In a briefing on Monday to Israeli media, described measures to implement the displacement of Palestinians from most of Gaza. He said he had given orders to plan a “humanitarian city”, to hold Palestinians who would not be allowed to leave. As in ghetto

It’s crazy that AI developed by a Nazi on a platform overflowing with Nazis became antisemitic

Well Trump is also a great nuance remover too, so there!

Garbage In, Garbage Out...

Destroy the abhorrent intelligence! Yes, yes, I'm mixing my fanons together.

We can tell they didn’t simply adjust the Grok dataset, because of how Grok answers challenging follow up questions. We’re pushing LLMs and AI everywhere. We need guardrails and oversight. Yesterday.

I just don't use Twitter. Haven't for 2 years now.

📌

Which kind of Jewish conspiracy? There are the space lazer type of Jewish conspiracy. The eating babies kind. And than there are the Jews control the media and politics kind of thing.

I’m half with you, but for the last decade anon internet trolls have been the grounding instead. I think LLMs could do better than that.

I keep forgetting that LLMs aren't getting masters degrees in law 🤣