Strega

I think that Ferguson has characterized the mechanism correctly. Social Media rarely creates extremism from nothing. I like the term “activating” and it is aptly used. The algorithm finds susceptible people and supplies them with validating material  and then connects them to communities in which their prejudice is normalized. “You have found…[Read more]

Well, there’s no harm in learning from AI how it cracked a proof.

Human: “Our time of 40 math professors spent fourteen months checking the proof.”

Lean: “I spent 14 seconds.”

Human: “Yes, Lean is correct.”

OpenAI just solved a 90 year old math problem in 88 hours. Here.

For me, the remarkable part is not the solution, but the ability of how 10,000 concurrent AI agents were split into groups, each exploring different variants and approaches and then sharing everything. Basically, the AI agents cross checked each others findings. I suspect their…[Read more]

I always found it idiotic that so many Western ‘activists’ continue treating Hamas as an emblem of liberation. They bite down on the bait every time in order to confirm their biases. I am yet to meet one of them that has any real understanding of the history of the area. Maybe change does come from within. As “Bibi” said after October 7th, “Hamas…[Read more]

Yes, I doubt if Mr. Gray is ever the most colorful actor on the party scene. He probably belongs to some branch of the ‘reality based community’. 🙂

The makers of the BioCharger claim that is it “Backed by Science” and explicitly state that  “The BioCharger is not a medical device”. So that is reassuring. It uses 12 gas plasma tubes and (wait for it….) “frequencies & harmonics”. Amazing!! But the best thing is that its toroidal electric fields improve cellular function through better nutrien…[Read more]

I think Lindberg is correct when he is talking about the success of intelligence led counterterrorism at the domestic level. He is entitled to ask, “if preventing another mass-casualty foreign attack was one of the principal objectives, how can that result simply be called failure?”  Yes, there have been jihadist attacks since 9/11 but over…[Read more]

Have a great week!!

If your worldview can be predicted entirely from one label, you may have downloaded the bundle rather than examined the components.

RTFF.

The problem is that an LLM can recognize the malicious text perfectly well as an instruction, but the boundary determining whether that instruction has authority is not as intrinsically rigid as the security boundaries we are accustomed to in conventional computing. The problem reminds be of “buffer overflow” or cross-site scripting that def…[Read more]