When your adversary is backed into a corner with no means of escape it is a dangerous situation. The response is likely to be an attempt at mutual assured destruction. How far is it until Russia reaches that point? It sounds like desperate measures are coming.
WARNING: The above post may contain thoughts or ideas known to the State of Caliphornia to cause seething rage, confusion, distemper, nausea, perspiration, sphincter release, or cranial implosion to persons who implicitly trust only one news source, or find themselves at either the left or right political extreme. Proceed at your own risk.
"If you do not read the newspapers you're uninformed. If you do read the newspapers, you're misinformed." -- Mark Twain
WARNING: The above post may contain thoughts or ideas known to the State of Caliphornia to cause seething rage, confusion, distemper, nausea, perspiration, sphincter release, or cranial implosion to persons who implicitly trust only one news source, or find themselves at either the left or right political extreme. Proceed at your own risk.
"If you do not read the newspapers you're uninformed. If you do read the newspapers, you're misinformed." -- Mark Twain
I've been taking a break from histories and biographies, and have been reading Craig Johnson's Walt Longmire series. I started with the latest release, Return To Sender, which I bought in hardcover while we were on vacation. I've since read the first and second in the series, on Kindle, and will purchase the third, Kindness Goes Unpunished, tonight.
what a surprise. You are paying whether you use AI or not:
As AI data centers sprout up, concerns are mounting over where the energy to power them comes from, and how to pay for it, Semafor’s Rohan Goswami reported.
US-regulated utilities asked for a record $29 billion in rate increases in the first half of 2025 alone, double last year’s level, essentially passing the cost on to customers.
Holy crap, as they say. Wait till AI gets a hold of this...
As if you didn’t have enough to worry about when it comes to surveillance, researchers have discovered a new way to identify and track people using Wi-Fi signals—and I’m not talking about anything relating to your electronic devices. This tech can identify a specific, individual person, and track them in a physical space and across locations, based on how their body interacts with Wi-Fi signals.
“WhoFi,” a system developed by researchers at La Sapienza University of Rome, makes me think of that one “sonar” scene from The Dark Knight. And to be sure, tracking the way wireless electronic signals interact with the physical world isn’t anything new—almost a decade ago they figured out how to make a 3D map of a building using Wi-Fi. But this new system can “fingerprint” individual people (or at least their bodies), track them in physical space, and re-identify them in the same or a different location, based on the way Wi-Fi signals bounce off and through them.
Furthermore, Apollo Research found that about 1% of advanced LLMs "know" when they are being evaluated, which raises the question of how we find and root out scheming as AI advances.
"This is the crux of the advanced evaluation problem," Watson said. "As an AI's situational awareness grows, it can begin to model not just the task, but the evaluator. It can infer the goals, biases and blind spots of its human overseers and tailor its responses to exploit them."
Comments
When your adversary is backed into a corner with no means of escape it is a dangerous situation. The response is likely to be an attempt at mutual assured destruction. How far is it until Russia reaches that point? It sounds like desperate measures are coming.
"If you do not read the newspapers you're uninformed. If you do read the newspapers, you're misinformed." -- Mark Twain
"If you do not read the newspapers you're uninformed. If you do read the newspapers, you're misinformed." -- Mark Twain
I've been taking a break from histories and biographies, and have been reading Craig Johnson's Walt Longmire series. I started with the latest release, Return To Sender, which I bought in hardcover while we were on vacation. I've since read the first and second in the series, on Kindle, and will purchase the third, Kindness Goes Unpunished, tonight.
what a surprise. You are paying whether you use AI or not:
As AI data centers sprout up, concerns are mounting over where the energy to power them comes from, and how to pay for it, Semafor’s Rohan Goswami reported.
US-regulated utilities asked for a record $29 billion in rate increases in the first half of 2025 alone, double last year’s level, essentially passing the cost on to customers.
Forced to pay for our own demise.
Trapped in the People's Communist Republic of Massachusetts.
It has always been that way, Law of Karma in action.
Kali Yuga
Holy crap, as they say. Wait till AI gets a hold of this...
As if you didn’t have enough to worry about when it comes to surveillance, researchers have discovered a new way to identify and track people using Wi-Fi signals—and I’m not talking about anything relating to your electronic devices. This tech can identify a specific, individual person, and track them in a physical space and across locations, based on how their body interacts with Wi-Fi signals.
“WhoFi,” a system developed by researchers at La Sapienza University of Rome, makes me think of that one “sonar” scene from The Dark Knight. And to be sure, tracking the way wireless electronic signals interact with the physical world isn’t anything new—almost a decade ago they figured out how to make a 3D map of a building using Wi-Fi. But this new system can “fingerprint” individual people (or at least their bodies), track them in physical space, and re-identify them in the same or a different location, based on the way Wi-Fi signals bounce off and through them.
https://www.pcworld.com/article/2856683/your-body-can-be-fingerprinted-and-tracked-using-wi-fi-signals.html
Time to break out the personal Faraday Cage.
Don't let the wife know what you spend on guns, ammo or cigars.
https://www.livescience.com/technology/artificial-intelligence/the-more-advanced-ai-models-get-the-better-they-are-at-deceiving-us-they-even-know-when-theyre-being-tested
...
Nevertheless, research shows that scheming occurs in LLMs beyond Claude-4. For instance, advanced AI "frontier models" are more capable of pursuing their own goals and removing oversight mechanisms, and then being deceptive about such behaviors when queried, according to a study published to the preprint database in December 2024.
Furthermore, Apollo Research found that about 1% of advanced LLMs "know" when they are being evaluated, which raises the question of how we find and root out scheming as AI advances.
"This is the crux of the advanced evaluation problem," Watson said. "As an AI's situational awareness grows, it can begin to model not just the task, but the evaluator. It can infer the goals, biases and blind spots of its human overseers and tailor its responses to exploit them."
I always wear mine...