SCRH2O.ai has unveiled two new vision-language models, H2OVL Mississippi-2B and H2OVL Mississippi-0.8B, aimed at enhancing document analysis and optical character recognition (OCR) tasks. Despite their smaller size, these models demonstrate competitive performance against larger models from major tech companies, with the 0.8B model excelling in the OCRBench Text Recognition task. CEO Sri Ambati emphasized that …
Perplexity has launched an Internal Knowledge Search feature that allows Pro and Enterprise Pro users to search both their internal files and the web from a single platform. Users can upload important files, limited to 500 for Enterprise Pro, in various formats like Excel, Word, and PDFs, VentureBeat reports. The aim is to streamline the …
OpenAI has launched a Windows desktop application for ChatGPT, enhancing its integration into daily workflows and productivity. This release follows a similar app for macOS and is currently available to select subscribers, VentureBeat reports. The app allows users to access ChatGPT easily via a keyboard shortcut, aiming to streamline user experience without the need for …
Google has updated its AI note-taking and research assistant, NotebookLM, enhancing its audio summarization feature to allow users to guide AI-generated audio conversations on specific topics. This update lets users customize audio overviews using a dedicated control, enabling them to focus on particular content areas. Raiza Martin, NotebookLM’s product leader, emphasized that this customization helps …
Microsoft and Tsinghua University have developed a new AI architecture called “Differential Transformer” that improves the performance of large language models. Furu Wei from Microsoft Research told VentureBeat that the new method amplifies attention to relevant contexts and filters out noise. This is designed to reduce problems such as the “lost-in-the-middle” phenomenon and hallucinations in …
A new text-to-image framework called Sana can efficiently and quickly generate high-resolution images up to 4096 x 4096 pixels. The system uses a deep compression autoencoder, linear attention, and a decoder-based text encoder to optimize performance. According to the developers, Sana-0.6B can compete with state-of-the-art large diffusion models, but is 20 times smaller and over …
Social media platform X (formerly Twitter) is changing its privacy policy to share user data with external partners for AI training. Sarah Perez reports for TechCrunch that users will be able to opt out of this use of their data. The change goes into effect on November 15, and could open up a new revenue …
Salesforce CEO Marc Benioff criticized Microsoft’s AI assistant, Copilot, on his personal X/Twitter account, calling it “disappointing” and likening it to “Clippy 2.0,” a reference to the infamous virtual assistant from the 1990s. He claimed that Copilot lacks accuracy and fails to deliver value, despite its recent upgrades that include new features like vision and …
SynthBee has raised $20 million in funding to develop a new kind of “computing intelligence”, Dean Takahashi reports for VentureBeat. The round was led by Crosspoint Capital Partners. SynthBee plans to use the money to expand its team and proprietary computing intelligence platform. SynthBee’s founder, Rony Abovitz, is known for previous successful startups including Mako …
Microsoft and OpenAI are currently negotiating the future of their partnership. According to several US media reports, OpenAI is demanding more financial support and computing power from Microsoft. However, the technology company is reluctant to invest billions more. OpenAI expects to lose $5 billion this year and is looking for ways to cover its high …