Comment Re:Juvenile charges? (Score 1) 56
Any school that charges lunch money would be a school in a third world country, if even that.
Comment Re:And there is the problem. (Score 1) 19
I see that there's a difference between writing a document and understanding what you wrote and the use of an illustration image in the document.
If I use AI to generate an illustration image and specify that it's AI generated but the text is purely my creation then I don't see any issue.
Submission + - Nintendo says your Switch 2 isn't really yours even if you paid for it (betanews.com)
BrianFagioli writes: The new Nintendo Switch 2 is almost here. Next month, eager fans will finally be able to get their hands on the highly anticipated follow-up to the wildly popular hybrid console. But before you line up (or frantically refresh your browser for a preorder), you might want to read the fine print, because Nintendo might be able to kill your console.
Yes, really. That’s not just speculation, folks. According to its newly updated user agreement, Nintendo has granted itself the right to make your Switch 2 “permanently unusable” if you break certain rules. Yes, the company might literally brick your device.
Buried in the legalese is a clause that says if you try to bypass system protections, modify software, or mess with the console in a way that’s not approved, Nintendo can take action. And that action could include completely disabling your system. The exact wording makes it crystal clear: Nintendo may “render the Nintendo Account Services and/or the applicable Nintendo device permanently unusable in whole or in part.”
Yes, really. That’s not just speculation, folks. According to its newly updated user agreement, Nintendo has granted itself the right to make your Switch 2 “permanently unusable” if you break certain rules. Yes, the company might literally brick your device.
Buried in the legalese is a clause that says if you try to bypass system protections, modify software, or mess with the console in a way that’s not approved, Nintendo can take action. And that action could include completely disabling your system. The exact wording makes it crystal clear: Nintendo may “render the Nintendo Account Services and/or the applicable Nintendo device permanently unusable in whole or in part.”
Submission + - Anthropic can now track the bizarre inner workings of a large language model
tomatoguy writes: Having psychology-adjacent interests (and perhaps because it's a Friday afternoon), I found this fascinating.
What the firm found challenges some basic assumptions about how this technology really works.
MIT Technology Review
https://ancillary-proxy.atarimworker.io?url=https%3A%2F%2Fwww.technologyreview.com%2F2025%2F03%2F27%2F1113916%2Fanthropic-can-now-track-the-bizarre-inner-workings-of-a-large-language-model%2F (paywalled)
https://ancillary-proxy.atarimworker.io?url=https%3A%2F%2Farchive.is%2F4mujU (free)
= This caught my eye first: studying something that claims to be brainy using brain-investigation tools and approaches.
Anthropic says it was inspired by brain-scan techniques used in neuroscience to build what the firm describes as a kind of microscope that can be pointed at different parts of a model while it runs. The technique highlights components that are active at different times. Researchers can then zoom in on different components and record when they are and are not active.
= Secondly, LLMs get "better" when they know when to shut up.
The latest generation of large language models, like Claude 3.5 and Gemini and GPT-4o, hallucinate far less than previous versions, thanks to extensive post-training (the steps that take an LLM trained on text scraped from most of the internet and turn it into a usable chatbot). But Batson’s team was surprised to find that this post-training seems to have made Claude refuse to speculate as a default behavior. When it did respond with false information, it was because some other component had overridden the “don’t speculate” component.
What the firm found challenges some basic assumptions about how this technology really works.
MIT Technology Review
https://ancillary-proxy.atarimworker.io?url=https%3A%2F%2Fwww.technologyreview.com%2F2025%2F03%2F27%2F1113916%2Fanthropic-can-now-track-the-bizarre-inner-workings-of-a-large-language-model%2F (paywalled)
https://ancillary-proxy.atarimworker.io?url=https%3A%2F%2Farchive.is%2F4mujU (free)
= This caught my eye first: studying something that claims to be brainy using brain-investigation tools and approaches.
Anthropic says it was inspired by brain-scan techniques used in neuroscience to build what the firm describes as a kind of microscope that can be pointed at different parts of a model while it runs. The technique highlights components that are active at different times. Researchers can then zoom in on different components and record when they are and are not active.
= Secondly, LLMs get "better" when they know when to shut up.
The latest generation of large language models, like Claude 3.5 and Gemini and GPT-4o, hallucinate far less than previous versions, thanks to extensive post-training (the steps that take an LLM trained on text scraped from most of the internet and turn it into a usable chatbot). But Batson’s team was surprised to find that this post-training seems to have made Claude refuse to speculate as a default behavior. When it did respond with false information, it was because some other component had overridden the “don’t speculate” component.
Submission + - DOGE software engineer's computer infected by info-stealing malware (arstechnica.com)
gkelley writes: Login credentials belonging to an employee at both the Cybersecurity and Infrastructure Security Agency and the Department of Government Efficiency have appeared in multiple public leaks from info-stealer malware, a strong indication that devices belonging to him have been hacked in recent years.
Kyle Schutt is a 30-something-year-old software engineer who, according to Dropsite News, gained access in February to a “core financial management system” belonging to the Federal Emergency Management Agency. As an employee of DOGE, Schutt accessed FEMA’s proprietary software for managing both disaster and non-disaster funding grants. Under his role at CISA, he likely is privy to sensitive information regarding the security of civilian federal government networks and critical infrastructure throughout the US.
Kyle Schutt is a 30-something-year-old software engineer who, according to Dropsite News, gained access in February to a “core financial management system” belonging to the Federal Emergency Management Agency. As an employee of DOGE, Schutt accessed FEMA’s proprietary software for managing both disaster and non-disaster funding grants. Under his role at CISA, he likely is privy to sensitive information regarding the security of civilian federal government networks and critical infrastructure throughout the US.
Submission + - Police Dismantles Botnet Selling Hacked Routers As Residential Proxies (bleepingcomputer.com)
An anonymous reader writes: Law enforcement authorities have dismantled a botnet that infected thousands of routers over the last 20 years to build two networks of residential proxies known as Anyproxy and 5socks. The U.S. Justice Department also indicted three Russian nationals (Alexey Viktorovich Chertkov, Kirill Vladimirovich Morozov, and Aleksandr Aleksandrovich Shishkin) and a Kazakhstani (Dmitriy Rubtsov) for their involvement in operating, maintaining, and profiting from these two illegal services.
During this joint action dubbed 'Operation Moonlander,' U.S. authorities worked with prosecutors and investigators from the Dutch National Police, the Netherlands Public Prosecution Service (Openbaar Ministerie), and the Royal Thai Police, as well as analysts with Lumen Technologies' Black Lotus Labs. Court documents show that the now-dismantled botnet infected older wireless internet routers worldwide with malware since at least 2004, allowing unauthorized access to compromised devices to be sold as proxy servers on Anyproxy.net and 5socks.net. The two domains were managed by a Virginia-based company and hosted on servers globally.
On Wednesday, the FBI also issued a flash advisory (PDF) and a public service announcement warning that this botnet was targeting patch end-of-life (EoL) routers with a variant of the TheMoon malware. The FBI warned that the attackers are installing proxies later used to evade detection during cybercrime-for-hire activities, cryptocurrency theft attacks, and other illegal operations. The list of devices commonly targeted by the botnet includes Linksys and Cisco router models, including:
— Linksys E1200, E2500, E1000, E4200, E1500, E300, E3200, E1550
— Linksys WRT320N, WRT310N, WRT610N
— Cisco M10 and Cradlepoint E100
During this joint action dubbed 'Operation Moonlander,' U.S. authorities worked with prosecutors and investigators from the Dutch National Police, the Netherlands Public Prosecution Service (Openbaar Ministerie), and the Royal Thai Police, as well as analysts with Lumen Technologies' Black Lotus Labs. Court documents show that the now-dismantled botnet infected older wireless internet routers worldwide with malware since at least 2004, allowing unauthorized access to compromised devices to be sold as proxy servers on Anyproxy.net and 5socks.net. The two domains were managed by a Virginia-based company and hosted on servers globally.
On Wednesday, the FBI also issued a flash advisory (PDF) and a public service announcement warning that this botnet was targeting patch end-of-life (EoL) routers with a variant of the TheMoon malware. The FBI warned that the attackers are installing proxies later used to evade detection during cybercrime-for-hire activities, cryptocurrency theft attacks, and other illegal operations. The list of devices commonly targeted by the botnet includes Linksys and Cisco router models, including:
— Linksys E1200, E2500, E1000, E4200, E1500, E300, E3200, E1550
— Linksys WRT320N, WRT310N, WRT610N
— Cisco M10 and Cradlepoint E100
Comment Re:Waive the egress fee (Score 1) 59
Since you as a customer owns the data it could also mean costly court cases if the data isn't released so it's easier to just waive the fees.
Submission + - A simple question tripped up a North Korean spy (yahoo.com)
smooth wombat writes: Over the past year there have been stories about North Korean spies unknowningly, or knowingly, being hired to work in western companies. During an interview by Kraken, a crypto exchange, the interviewers became suspicious about the candidate. Instead of cutting off the interview, Kraken decided to continue the candidate through the hiring process to gain more information. One simple question confirmed the user wasn't who they said they were and even worse, was a North Korean spy. From the article:
The interview was scheduled for Halloween, a classic American holiday—especially for college students in New York—that Smith seemed to know nothing about.
“Watch out tonight because some people might be ringing your doorbell, kids with chain saws,” Percoco said, referring to the tradition of trick or treating. “What do you do when those people show up?”
Smith shrugged and shook his head. “Nothing special,” he said.
Smith was also unable to answer simple questions about Houston, the town he had supposedly been living in for two years. Despite having listed "food" as an interest on his résumé, Smith was unable to come up with a straight answer when asked about his favorite restaurant in the Houston area. He looked around for a few seconds before mumbling, “Nothing special here.”
The interview was scheduled for Halloween, a classic American holiday—especially for college students in New York—that Smith seemed to know nothing about.
“Watch out tonight because some people might be ringing your doorbell, kids with chain saws,” Percoco said, referring to the tradition of trick or treating. “What do you do when those people show up?”
Smith shrugged and shook his head. “Nothing special,” he said.
Smith was also unable to answer simple questions about Houston, the town he had supposedly been living in for two years. Despite having listed "food" as an interest on his résumé, Smith was unable to come up with a straight answer when asked about his favorite restaurant in the Houston area. He looked around for a few seconds before mumbling, “Nothing special here.”
Comment Re:Reinstall Windows 11 (Score 1) 2
Sorry, this is a reinstall of Windows 11 Enterprise that isn't using the Dell system and that doesn't matter in this case. This could happen to any installation linked to intune.
I can't choose to not use intune in this case since that's decided above my level.
Submission + - Ask Slashdot: Trouble with Microsoft Intune Autopilot 2
Z00L00K writes: At the company where I work we have now three times encountered cases where some Windows 11 computers have either been reassigned to another company or removed from the company where I work and ended up being a generic Microsoft account computer.
This is first visualized on the Let's add your Microsoft account page where the logo is incorrect over the Sign In field. In my most recent case the computer has been removed from Intune, but has still been present in Entra. The end result has been a computer with a quite weird behavior limiting its usability.
All computers have been Dell computers of various models, but I doubt that it's a Dell only issue.
So my question is if anyone else has encountered the same phenomenon?
This is first visualized on the Let's add your Microsoft account page where the logo is incorrect over the Sign In field. In my most recent case the computer has been removed from Intune, but has still been present in Entra. The end result has been a computer with a quite weird behavior limiting its usability.
All computers have been Dell computers of various models, but I doubt that it's a Dell only issue.
So my question is if anyone else has encountered the same phenomenon?
Submission + - High tariffs become 'real' with our first $36K bill (adafruit.com)
ptorrone writes: We're no stranger to tariff bills, although they have definitely ramped up over the last two months. However, this is our first 'big bill', where a large portion was subjected to a 125%+20%+25% import markup. Unlike other taxes like sales tax where we collect on behalf of the state and then submit it back at the end of the month, or income taxes, where we only pay if we are profitable, tariff taxes are paid before we sell any of the products and are due within a week of receipt which has a big impact on cash flow.
In this particular case, we're buying from a vendor, not a factory, so we can't second-source the items (and these particular products we couldn't manufacture ourselves even if we wanted to, since the vendor has well-deserved IP protections). And the products were booked & manufactured many months ago, before the tariffs were in place. Since they are electronics products/components, there's a chance we may be able to request reclassification on some items to avoid the 125% 'reciprocal' tariff, but there's no assurance that it will succeed, and even if it does, it is many, many months until we could see a refund.
We'll have to increase the prices on some of these products, but we're not sure if people will be willing to pay the higher cost, so we may well be 'stuck' with unsellable inventory that we have already paid a large fee on.
In this particular case, we're buying from a vendor, not a factory, so we can't second-source the items (and these particular products we couldn't manufacture ourselves even if we wanted to, since the vendor has well-deserved IP protections). And the products were booked & manufactured many months ago, before the tariffs were in place. Since they are electronics products/components, there's a chance we may be able to request reclassification on some items to avoid the 125% 'reciprocal' tariff, but there's no assurance that it will succeed, and even if it does, it is many, many months until we could see a refund.
We'll have to increase the prices on some of these products, but we're not sure if people will be willing to pay the higher cost, so we may well be 'stuck' with unsellable inventory that we have already paid a large fee on.
Comment Re:Poll please (Score 1) 29
I can definitely agree that more and more searches are ending up in irrelevant results so you have to refine your searches.
Unfortunately the prompt interpreter isn't smart, so it have a hard time to understand.
Comment How about 2 hours? (Score 1) 78
I had a college roommate that slept 2 hours a night at most. He took 2 15 minute naps during the day and that's all he slept. He had two full time jobs and maintained a straight A college record at one point for Electrical Engineering (oh, and had time for a girlfriend). The guy he roomed with (same apartment, 2 rooms, 4 students) had an eidetic (photographic) memory and could read a page he just looked at in a textbook back to me. Stuff I studied for 10+ hours to do the homework he did in 30 minutes, but he only had 1 full time job because he needed 8 hours of sleep. The guy in my room needed like 14 hours a sleep at night if an alarm didn't wake him. I'd sleep probably 3 hours, be up for 3, then slept another 3.
I imagine if they did a case study for sleep on that apartment, we'd have broken it. I transferred after that year and had normal roommates except maybe one, but he slept a lot after getting really stoned every night, so I don't know if that counts.
Comment Re: Interesting! (Score 1) 78
I was going to say the same thing. In fact, my mom and I both sleep like that - 3 to 3 1/2 hours asleep, up for an hour or two, then another 3 to 3 1/2 hours. My dad and brother go to sleep at like 9:30PM and are up at 6AM like clockwork. I've heard that my (and mom's) sleep pattern is actually historically normal. Sometimes I do need 8-9 hours of sleep and sleep the entire night, but ~3 1/2 hour cycles is much more normal.
Comment Re:Captcha (Score 1) 26
Not with the AIs of today.
We need something that's better than the captchas that we have seen so far.
Slashdot Top Deals
The herd instinct among economists makes sheep look like independent thinkers.
