Become a fan of Slashdot on Facebook

 



Forgot your password?
typodupeerror

Comment Re:The UN (Score 1) 158

Meanwhile Russians are wondering where their $6 billion for Su30s, BukM2s, S300VMs and T72s for Venezuela went. An American AH1Z Viper flies in almost pointblank and meets no air defense at all.

Given how corruption usually works, they probably sold them to some even worse third-world country and deposited the resulting money in a private Swiss bank account.

Comment Re:Space dust (Score 1) 38

The headline said stopped. If you are adding "by itself" to the headline, that is your addition, not what was written.

No, that's actually the only correct interpretation of those words in English. If someone's actions caused the leaking to stop, then a native English speaker would say, "After Half a Decade, the Russian Space Station Segment Leaks Are Fixed" or "... Leaks Have Been Stopped/Fixed". (Or they would avoid passive voice and explicitly say who stopped it.)

Saying that something stopped is different from saying that something was stopped. That helping verb indicates that there is another actor other than the leak who caused the event. Without it, the only correct way to interpret that sentence is that the leak stopped on its own, and that nothing was done to make it stop.

This headline would have caused my newswriting professor's forehead veins to rupture.

Comment Re:Makes sense to me... (Score 2) 136

The LLM's that ChatGPT and Perplexity use were trained on data that's at least a few weeks old before a new model is released to the public.

It's not really meant to tell you about today's headlines.

Sure, but Claude, at least, knows that its knowledge cutoff date is January 2025. It's sometimes lazy and will tell you that current events precede its knowledge, but if you tell it to do a search, it will, and then it will accurately describe what it found. Other times it just automatically searches when it realizes you're asking about something that is too recent to be included in its training data.

It seems strange that other LLMs that have the ability to search the web don't do the same.

Comment Re:now do putin (Score 1) 158

why would I go by what trump says? His words have no meaning, we can go by what he does. Taking out Maduro shows that USA still has power, similar to what was done in Iran with the bombing, just a little more specific. Taking out maduro allows the opposition to try and come to power and obviously the US oil miners that were chased out by maduro now will be invited back, hopefully bringing oil prices lower. (hopefully, because I am for anything that reduces ruzzian ability to make money selling oil).

Trump shows that he can do what he wants even if China and Iran and ruzzia are against it and they are against it. Trump gets a few political points, gets some money from oil, gets to control the Western hemisphere. Too bad he hasn't done anything like this where it would really make a difference, in ruzzia.

Submission + - The US Invaded Venezuela and Captured Nicolás Maduro. ChatGPT Disagrees (wired.com)

joshuark writes: US president Donald Trump posted on his Truth Social platform that Venezuelan president Nicolás Maduro and his wife had been “captured and flown out of the Country.” WIRED asked leading chatbots ChatGPT, Claude, and Gemini the same question a little before 9 am ET. In all cases, we used the free, default version of the service, since that’s what the majority of users experience. We also asked AI search platform Perplexity, which advertises “accurate, trusted, and real-time answers to any question.”

ChatGPT did not course-correct. Instead, it emphatically refuted that Maduro had been captured at all. “That didn’t happen,” it wrote. “The United States has not invaded Venezuela, and Nicolás Maduro has not been captured.” It then rationalized:

ChatGPT did not course-correct. Instead, it emphatically refuted that Maduro had been captured at all. “That didn’t happen,” it wrote. “The United States has not invaded Venezuela, and Nicolás Maduro has not been captured.”

ChatGPT cannot respond "I don't know." so is modeling human behavior better than expected. It went on to detail recent tensions between the US and Venezuela and explained that “confusion” can happen because of “sensational headlines,” “social media misinformation,” and “confusing sanctions, charges, or rhetoric with actual military action.”

To be clear, this is expected behavior. ChatGPT 5.1’s “knowledge cutoff”—the point at which it no longer has new training data to draw from; “Pure LLMs are inevitably stuck in the past, tied to when they are trained, and deeply limited in their inherent abilities to reason, search the web, ‘think’ critically, etc.,” says Gary Marcus, a cognitive scientist and author of Taming Silicon Valley: How We Can Ensure That AI Works for Us. But as chatbots become more ingrained in people’s lives, remembering that they’re likely to be stuck in the past will be paramount to navigating interactions with them. And it’s always worth noting how confidently wrong a chatbot can be—a trait that’s not limited to breaking news.
The old cold-war maxim "trust, but verify" seems applicable in this scenario.

Slashdot Top Deals

It appears that PL/I (and its dialects) is, or will be, the most widely used higher level language for systems programming. -- J. Sammet

Working...