Welcome back to another week covering what’s happened in the AI space. With the release of the rabbit R1 finally getting out to the initial purchases. It was interesting to see this Marques Brownlee video doing the rounds where it seems like the AI is trying to deny that it knows his location specifically. Jump to the comments and there are posts about the device knowing its own location but not specifically knowing where Marques is which may be the case, but it opens up a whole lot of other privacy potential concerns if they didn’t ask permission to use current location for insights. What makes it worse is the denying of knowledge of this by the device (you can see that in YouTube shorts here and other users have also found the device gaslighting them – see here).
The full video review is below.
Another interesting thing to come out of the news this week was this 60 Minutes interview with NVIDIA CEO Jensen Huang. They talk about the new GPU, Blackwell, and they showcase the range of work with AI that NVIDIA is powering from movies to drug development.
Is there a new OpenAI model out there? Some people think GPT4.5 is out there as this appeared on the leaderboards but it went by the name GPT2-Chatbot.
But then, just as fast as it appear, it was gone, now no longer available
Sam Altman even mentioned it which led to people thinking it was an OpenAI model. Tweet here.
More here from ArsTechnica in this article “Mysterious “gpt2-chatbot” AI model appears suddenly, confuses experts“.
There was also this piece on the different stages that are likely to happen with AI taking over roles. The fight we have right now over jobs getting taken away by AI only seem to think about the jobs we have now. We don’t yet know what new jobs get created based on new capabilities so this is not a negative only thing to think about.
For the music fans out there, here’s a great Cold Fusion take covering the ChatGPT moment that the music industry is going through.
If you’re into your AI documentaries, this one should be fun given it’ll set you back 3 hours (very Christopher Nolan-esque). This is on the history of AI and ‘how it was stolen’.
We’ll leave you with this piece which you may have missed from one of our Directors, Arturo, which originally came up on his personal blog but which we’ve republished. It’s all about the stochasiticity problem in AI outputs and what we’re doing to combat it with SIKE.
Reducing Stochasticity in Corporate Applications of Language Models Through Semantic Programming
Get in touch
Stay tuned for more insights from us next week and if you’re interested in SIKE (our business friendly AI tool), then you can check out more on what it’s about here.
Or reach out to Nick Bishop CFA Arturo Rodriguez or 📈 Mark Monfort here and get us on email here.
