Here are the top trending news from the world of AI and technology during last 24 hours.
1)
Third Parties can now access your X Data for training its AI model
Your X data is now up for grabs. The Elon Musk led company announced a major change to its policy on Thursday that allows it to share users’ data with third party collaborators in order to train its AI models. This new policy will come into effect from November 15, 2024. The policy does give the users the option of opting out of this arrangement. They probably will have to do this manually by going into settings. Apparently, X has adopted the Reddit model, which recently inked a content licensing deal with Google ahead of its IPO. For X, creating new monetization opportunities is critical in the wake of top advertisers continuing to boycott the microblogging platform, which has clearly impacted its topline.
2)
ChatGPT debuts on Windows with a dedicated app
ChatGPT has finally arrived on Windows. OpenAI on Thursday announced that it has released the preview version of ChatGPT’s windows app for it paid users. At least for now, free users won’t get the access. Although the app would offer all the standard features including access to its latest model o1, it comes with many limitations. For instance, it does not support voice functionality yest including advanced voice model. Furthermore, some of the GPT store integration aren’t quite functional either. ChatGPT’s Windows app are optimized for running on Windows 10. OpenAI’s hasn’t revealed when it plans to launch a full-fledge functional version of the app.
3)
NotebookLM launches a new paid version NotebookLM Business
Google’s experimental research tool NotebookLM has unveiled a new business version called Notebook LM Business. Notebook LM Business will come with some extra valuable features and customers will obviously have to loosen their wallet to avail these valuable features. However, it is currently available only in pilot. Companies and organizations that wish to access the features will have to apply for its pilot program. According to Venture Beat, some of its features include higher usage limits and new features such as customization and sharing notebooks with team members. Separately, NotebookLM has launched a new feature called Audio Overview, which allows you to create a narrated study guide.
4)
Nvidia quietly launches a new AI model that has outperformed OpenAI’s ChatGPT 4
With much fanfare, GPU maker Nvidia earlier this week launched a new AI model that has surprisingly surpassed industry leaders like ChatGPT 4 and Anthropic’s Claude model on several key benchmarks. The model is called Llama-3.1-Nemotron-70B-Instruct and is available for download on the Hugging Face. As per Nvidia, its latest open-source LLM has outperformed OpenAI’s ChatGPT 4 and Anthropic’s Claude 3.5 on quite a few important evaluations, like scoring 85.0 on the Arena Hard benchmark, 57.6 on AlpacaEval 2 LC and 8.98 on the GPT-4-Turbo MT-Bench. Nvidia’s Llama-3.1-Nemotron-70B-Instruct model is based on Meta’s Llama 3.1 model but has been enhanced with RLHF training. This new open source model represents Nvidia’s ambition to create its own niche in the highly competitive LLM space, which is dominated by the likes of OpenAI, Mistral, Google and Anthropic.