Author: Advanced AI Bot

baona/Getty ImageMany people are turning to AI chatbots for personal needs, whether companionship or emotional support. However, in recent months, there have been heightened concerns about the potential harms of using this technology over human interactions.According to new research from OpenAI in partnership with the Massachusetts Institute of Technology, higher use of chatbots like ChatGPT may correspond with increased loneliness and less time spent socializing with other people. The analysis evaluated how AI chat platforms shape users’ emotional well-being and behaviors through two paired studies conducted by researchers at the organizations. The studies have not yet been peer-reviewed.Also: Microsoft’s new AI…

OpenAI is adding support for Anthropic’s Model Context Protocol, an open-source standard for connecting artificial intelligence applications to data repositories. Developers can use it to ensure their chatbots only provide responses based on relevant information. OpenAI CEO Sam Altman announced on Wednesday via X that the protocol can now be used in the Agents SDK, its toolkit for building AI-powered agents. The company’s developer team added that it is working on adding support for the ChatGPT desktop app and OpenAI API, and will announce more in the coming months. SEE: Have Some Spare Cash? You’ll Need it for OpenAI’s New…

Amid recent reports of Apple struggling in its AI efforts, particularly with Siri, a new analyst report indicates the company is close to laying down a full $1 billion in new purchases of NVIDIA’s AI servers. Apple reportedly ordering hundreds of pricey NVIDIA AI servers Patrick Seitz writes at Investor’s Business Daily: Loop Capital analyst Ananda Baruah late Monday said Apple is in the process of placing orders for about $1 billion in Nvidia GB300 NVL72 systems. That equates to about 250 servers at $3.7 million to $4 million each, he said in a client note. Apple is working with…

AnthropicLarge organizations wanting to integrate generative AI into their business can find the process daunting, especially if they’re unsure how or where to start. A new initiative with Anthropic’s Claude AI aims to provide the tools for enterprises struggling with this challenge.In a new five-year partnership announced Thursday, Anthropic and AI company Databricks will offer Claude AI models through the Databricks Data Intelligence Platform, which helps businesses use AI with their internal data. The deal will bring the tools and technology to more than 10,000 companies.Also: Navigating AI-powered cyber threats: 4 expert security tips for businessesAnthropic’s latest Claude 3.7 Sonnet…

Google has introduced version 2.5 of its Gemini AI model, which the company said offers a new level of performance by combining an enhanced base model with improved post-training. The Gemini 2.5 models are thinking models, capable of reasoning through their thoughts before responding, thus resulting in better performance and improved accuracy, said Koray Kavukcuoglu, CTO of Google DeepMind, in a March 25 blog post. The first Gemini 2.5 release is an experimental version of Gemini 2.5 Pro, which Kavukcuoglu said is state-of-the-art across a range of benchmarks that require advanced reasoning. Available now, Gemini 2.5 Pro Experimental demonstrates strong…

Semiconductor giant Nvidia’s (NVDA) H20 artificial intelligence (AI) chips are facing a shortage in China amid increasing demand, according to Chinese partner H3C. The news was first reported by Reuters, citing reference to a client notice. Nvidia’s H20 AI chips are among the few advanced AI processors legally permitted for sale in China. H3C warned that their shortage could pose a serious threat to China’s AI ambitions.Light Up your Portfolio with Spark:H3C manufactures data center servers and is one of the OEM (Original Equipment Manufacturer) partners for the sale of Nvidia’s chips in the mainland. Notably, H3C’s H20 inventory is…

Stability AI speaks of a “multi-view video generation with 3D camera control”. This refers to an AI model that can convert photos into 3D models. These in turn can be viewed from all sides, enabling an immersive view. Stable Virtual Camera is still a research preview. Stability AI uses a diffusion model to create the 3D videos. These are AI models in which images are generated point by point using noise. A single photo or up to 32 images can be used as input. The generated videos are available with different camera paths, such as “dynamic”, “spiral”, “dolly zoom”, “pan”…

This paper demonstrates in a series of experiments that current safety alignment techniques of LLMs, as well as corresponding jailbreaking attacks, are in large part focusing on modulating the distribution of the first few tokens of the LLM response. Paper: Abstract: The safety alignment of current Large Language Models (LLMs) is vulnerable. Simple attacks, or even benign fine-tuning, can jailbreak aligned models. We note that many of these vulnerabilities are related to a shared underlying issue: safety alignment can take shortcuts, wherein the alignment adapts a model’s generative distribution primarily over only its very first few output tokens. We unifiedly…

❤️ Check out Lambda here and sign up for their GPU Cloud: Guide for using DeepSeek on Lambda: 📝 The Gemma 3 paper and the rest are available here: Sources: 📝 My paper on simulations that look almost like reality is available for free here: Or this is the orig. Nature Physics link with clickable citations: 🙏 We would like to thank our generous Patreon supporters who make Two Minute Papers possible: Benji Rabhan, B Shang, Christian Ahlin, Gordon Child, John Le, Juan Benet, Kyle Davis, Loyal Alchemist, Lukas Biewald, Michael Tedder, Owen Skarpness, Richard Sundvall, Steef, Taras Bobrovytsky, Thomas…

We missed this earlier: Social media conglomerate Meta is involved in revenue-sharing agreements with hosts of its Llama Artificial Intelligence (AI) models, revealed court filings related to an ongoing copyright lawsuit, Kadrey vs. Meta. While the case itself deals with Meta’s alleged usage of a vast quantity of data from pirated e-books to train Llama, proceedings before the United States District Court for the Northern District of California on March 21 saw plaintiffs’ attorneys noting that the company “shares a percentage of the revenue” that the AI model hosts generate from Llama users. The court redacts the names of the…