otter

Either Claude Opus 5 just leaked a trove of private 2024–2025 chat logs from Anthropic employees (& general public) asking the AI to edit their personal letters, or it’s generating one hell of a hallucination. It’s almost certainly the latter. But read the examples and you might start second-guessing.

It leaves you wondering what exactly is going on with Claude Opus 5—and whether we’re looking at a hallucination, or a model trained on its own creator’s internal data.

I think this article and headline does a poor job of explaining the wider context around this one.

It is likely to be aimed at financial traders who want to see market-moving news fast. Trump’s social media posts often cause sudden swings in global markets, especially when he writes about trade and tariffs.

For firms, a delay of even seconds can be costly. Until now, banks and traders had to monitor the app manually. The new system will send posts directly to paying clients.

I saw this linked elsewhere. The key points in my opinion are:

• AI generated music will be tagged and demonitized
• they will start taking down AI songs uploaded to hijacked artist profiles

Relevant sections:

Tidal defines AI-generated music as music that is wholly or substantially generated by generative artificial intelligence. [...]

Tidal will accept AI-generated music. Artists should have the freedom to create with AI tools, and listeners should have the autonomy to choose the type of content they consume [...]

Tidal will identify and tag AI-generated music. Listeners should know whether content they are listening to is AI-generated. To start, listeners will see an icon next to content we identify as 100% AI-generated. This feature will start appearing for listeners in mid-July. As AI-detection methods become more reliable, we will expand this tag to content that is substantially AI-generated. But the responsibility to identify and tag AI-generated content should not rest with Tidal alone. We expect — and will begin to enforce — that content distributors identify AI-generated content before it reaches our platform.

[...] We will not tolerate AI-generated music that exploits an individual’s or group’s music, name or likeness, deceives listeners, or diminishes the quality of our service. Effective mid-July, AI-generated music associated with fraudulent activity will be blocked or removed from our platform. Fraudulent activity includes (but is not necessarily limited to) AI-generated music that aims to deceive listeners, interfere with authentic artists and their audience, or involves high-volume uploads or unusual streaming activity. [...]

Starting today, AI-generated music will not be monetizable. We are only in the beginning of the era of AI-generated music. We acknowledge the ongoing debate regarding whether certain AI-generated music (e.g. AI-generated music developed from fairly and properly licensed models) should be entitled to earn royalties. This debate will continue as the [...]

cross-posted from: https://lemmy.ca/post/37011397

!opensource@programming.dev

The popular open-source VLC video player was demonstrated on the floor of CES 2025 with automatic AI subtitling and translation, generated locally and offline in real time. Parent organization VideoLAN shared a video on Tuesday in which president Jean-Baptiste Kempf shows off the new feature, which uses open-source AI models to generate subtitles for videos in several languages. 

I thought the CT scans were neat

The code that runs Redbox DVD rental machines has been dumped online, and, in the wake of the company’s bankruptcy, a community of tinkerers and reverse engineers are probing the operating system to learn how it works. Naturally, one of the first things people did was make one of the machines run Doom.

As has been detailed in several great articles elsewhere, the end of Redbox has been a clusterfuck, with pharmacies, grocery stores, and other retailers stuck with very large, heavy, abandoned DVD rental kiosks. To many people’s surprise, many of the kiosks remain operational even with the bankruptcy of Redbox’s parent company, which has led some people to “liberate” DVDs from the abandoned kiosks. Reddit is full of posts by people who say they have taken dozens of DVDs from kiosks all over the country.

In a Discord community called “Redbox Tinkering,” a FAQ states “just ask the store manager if you can have it. They will most likely tell you to just take it, but don’t just take it without asking.”

“Use heavy or appliance hand-trucks/dolly to wheel it onto your trailer or out to your truck,” the FAQ says. “It is heavy, so be prepared. I pull it right onto the trailer and strap it down standing up. You can lay them down, but know that most of the discs will be in disarray when you open it. Take everything having to do with Redbox, don’t leave a mess.” The FAQ also contains information about how to disconnect the Redbox from its power supply and how to cut through the bolts that secure the kiosk to concrete with a grinder. It also has information about how to open and disassemble the device at home.

“Unlike most tinkerer's my main goal isn't to reverse engineer the Official Software more than I have to. I am mainly interested in carousel movement, movie retrieving/returning, etc. I am using the machine to make my own version of the App to effectively do the same thing the original software does, but with my own spin on it. I mainly want to use it to create a massive DVD/Blu-Ray storage machine with ease of use for retrieving the movies.”

“I work in IT and have a decent sized Homelab and I've always been interested in making things work again once they break,” they added.

I've enjoyed Mark Rober's videos for a while now. They are fun, touch on accessible topics, and have decent production value. But this recent video isn't sitting right with me

The video is here: https://www.youtube.com/watch?v=SrGENEXocJU

In it, he talks about a few techniques for how to take down "bad guy drones", the problems with each, and then shows off the drone tech by Anduril as a solution.

https://en.wikipedia.org/wiki/Anduril_Industries

Anduril aims to sell the U.S. Department of Defense technology, including artificial intelligence and robotics. Anduril's major products include unmanned aerial systems (UAS), counter-UAS (CUAS), semi-portable autonomous surveillance systems, and networked command and control software.

In the video, the Anduril product is a heavy drone that uses kinetic energy to destroy other drones (by flying into them). Quoting the person in the video:

imagine a children's bowling ball thrown at twice as fast as a major league baseball fastball, that's what it's like getting hit by Anvil

This technology is scary for obvious reasons, especially in the wrong hands. What I also don't like is how Mark Rober's content is aimed at children, and this video includes a large segment advertising the children's products he is selling. Despite that, he is promoting military technology with serious ethical implications.

There's even a section in the video where they show off the Roadrunner, compare it against the patriot missiles, and loosely tie it in to defending against drones. While the Anvil could be used to hurt people, at least it is designed for small flying drones. The Roadrunner is not:

The Roadrunner is a 6 ft (1.8 m)-long twin turbojet-powered delta-winged craft capable of high subsonic speeds and extreme maneuverability. Company officials describe it as somewhere between an autonomous drone and a reusable missile. The basic version can be fitted with modular payloads such as intelligence and reconnaissance sensors. The Roadrunner-M has an explosive warhead to intercept UAS, cruise missiles, and manned aircraft.