OpenAI’s Interpreter 2.0 Breaks Military Restrictions, Partners with China for AI Safety | 70 characters

27
314



Let’s Look at the <a href='' title='Ai News Era' >AI News</a>

Let’s Look at the AI News

Date: 113 2024

First of all, OpenAI changes policy to allow military applications. It’s a minor tweak that most of us probably missed. It wasn’t broadcasted, but The Intercept caught it. OpenAI used to say no to anything that was military.

You’re not supposed to use it for military applications, but that’s not the case anymore. This is how it looked like before the change, “Don’t allow the use of our models for military and warfare,” now an OpenAI representative, Niik Felix, explained, “There’s still a blanket prohibition on developing and using weapons.” You can see that it was originally and separately listed from military and warfare. So, right here, it says, “Don’t use our service to harm yourself or others, don’t use it for self-harm to develop weapons, injure others or destroy property, or engage in unauthorized activities.” Maybe it was just rewritten to be a little bit more readable or maybe they want to be a little bit more specific but a little bit more open to certain applications.

Some AI companies from the US, like OpenAI, Anthropic, and Coh, secretly met with Chinese AI experts. It looks like the meetings were set up by the Shake Group who usually deals with conflicts in the Middle East. We covered a little bit of this in the video about when Sam Alman was trying to develop AI chips. It’s all very hush-hush, it’s not very publicized, but it does look like there are some attempts to communicate and coordinate how all these nations deal with AI safety.

Open Interpreter Update

There’s a major update for Open Interpreter. Open Interpreter had been the original code interpreter, now it’s called Advanced Analytics, but open source that you can see exactly what it’s doing and you’re able to use any model that you want with it. So, the new update is the most significant update to Open Interpreter since the first version. Almost every component has been rewritten to support their project’s objective of building a standard interface between language models and computers. They have a computer API, they have a full operating system mode where the thing just takes over your computer and is able to click on things, open stuff, run it, and it uses Python to move your mouse around, etc. Now it also has a vision. This is very exciting because it shows a lot of potential for the future.

Impressive Demonstration

Let’s take a closer look at the demonstration of Open Interpreter. It was able to do a data entry task, writing the Fibonacci sequence without needing to be over-explained. This demonstrates the potential for this software to handle various tasks with the latest models, and though it is a work in progress and still needs improvements, it is very promising. This early access to the desktop app can be a great opportunity to explore the potential of this software.

If you liked the article, be sure to give it a thumbs up and subscribe for more updates. Thank you for watching!


27 COMMENTS

  1. Freakin Wes – did you really go mandella effect on us, specifically without saying the words "mandella" and "effect"? lmao. Don't worry, you're not nuts. It's just that there are multiple instantiations of reality, running in parallel to each other, with greater and lesser degrees of similarity and difference between them. Nature is a lot of things, but inefficient is not one of them. Resources are economized. Many, many parallel instantiations – things like 'the missing horn' are simply a consequence of that efficiency, and a sort of best-fit compromise with coordinating state. Something like that.

  2. 🎯 Key Takeaways for quick navigation:

    00:00 🛠️ OpenAI adjusts policy to permit military applications, while maintaining a prohibition on developing and using weapons.
    01:22 🌐 US AI companies like OpenAI, Anthropic, and Cohere engage in secretive meetings with Chinese AI experts to discuss AI safety.
    01:50 🔄 Open Interpreter 2.0, an open-source code interpreter, receives a major update allowing the use of any language model locally, including vision capabilities.
    02:30 💻 Open Interpreter's latest update includes a computer API, full operating system mode, and vision support, aiming to build a standard interface between language models and computers.
    08:50 🚀 OpenAI's long-term vision for Open Interpreter is to become the go-to open-source AI operating system, comparable to Linux or Android.
    13:51 🤖 Open Interpreter demonstrates impressive capabilities in automating tasks like generating the Fibonacci sequence, showcasing its potential for efficient data entry and automation.

    Made with HARPA AI

  3. Fruit of loom, in my country I've never seen it that often and I didn't really care about it at all. But for some reason to me the logo with the cornucopia is the one I've seen back then too and I have a pretty good memory. There are many theories why things like that happen. Simulation or not, I think it was for some reason changed back then and like a ripple effect it got the update in our reality too. I've experienced things like that quite a lot and having a good memory makes you realize that far more often. Like when we bought a house and after I measured all doors, some days later they suddenly had other sizes… it was quite weird in cases like that because I remember being a little bit upset with the width, being more narrow and uncommon and suddenly… they were all normal and no, I didn't make a mistake while measuring, I am a perfectionist, I did the measuring several times to be sure, especially when you have to pay a lot of money to put new doors in…

  4. A(G)I is dangerous! Experts and pioneers have warned that it carries existential threats, and now they are deliberately allowing use of the technology for warfare purposes? I am a programmer with a background in AI and for me that's just instanity… 🤦🏻 Sufficiently advanced AGI is not containable, by any human, because it's faster and smarter than us. Think about it… #StayAwake