~ai-venturebeat | Bookmarks (539)

New approach flips the script on enterprise AI adoption by using input data you already have for fine-tuning instead of needing labelled data.

Users of ChatGPT Team can now add internal databases as references for ChatGPT, making the chat platform respond with better context.

Anthropic has developed a new method for peering inside large language models like Claude, revealing for the first time how these AI systems process information and make decisions. The research, published today in two papers (available here and here), shows these models are more sophisticated than previously understood — they plan ahead when writing poetry, […]

AI security startup Hakimo raises $10.5M to transform physical security with autonomous agents that monitor existing cameras 24/7, detecting threats and saving businesses $125,000 annually compared to traditional guards.

An updated version of the MCP spec introduced key upgrades to make AI agents more secure, capable and interoperable.

The new feature has been widely embraced by users of X, but it raises copyright concerns and goes against Studio Ghibli's creator.

SingularityNET, a founding member of the ASI Alliance, has partnered with ATMTA, maker of Star Atlas, the Web3 space exploration online game.

Groq partners with PlayAI to deliver Dialog, an emotionally intelligent text-to-speech model that runs 10x faster than real-time speech, including the Middle East's first Arabic voice AI model.

Observe.AI has officially launched VoiceAI agents, a solution designed to automate routine customer interactions in contact centers. The latest addition to the company’s AI-driven conversational intelligence platform, VoiceAI agents aim to improve customer experience while reducing operational costs. With this release, Observe.AI is positioning itself as the only complete AI-powered platform that supports enterprises across […]

Rapt AI, a provider of AI-powered AI-workload automation for GPUs and AI accelerators, has teamed with AMD to enhance AI infrastructure.

Microsoft announced Tuesday two significant additions to its Copilot Studio platform: deep reasoning capabilities that enable agents to tackle complex problems through careful, methodical thinking, and agent flows that combine AI flexibility with deterministic business process automation.

Nvidia is updating its computer vision models with new versions of MambaVision that combine the best of Mamba and transformers to improve efficiency.

Gemini 2.5 Pro is now available for Gemini Advanced users and is Google's most capable model with a 1 million token context window.

As AI-generated images become more precise and accessible, GPT-4o represents a significant step forward in the space.

One of the model’s standout capabilities is its strong text rendering performance, addressing a common challenge in AI-generated imagery.

Agentic AI is changing online meeting platforms and now thanks to Otter AI and other leading vendors it’s poised to change it even more.

The security platform built specific large lineage models (LLiMs) to track data lifecycles across users and endpoints and detect shadow AI.

SOCs are seeing false positive rates drop 70%, while shaving 40+ hrs a week of manual triage thanks to the rapid advances in AI copilots.

There's still a lot of juice left to be squeezed, cognitively and performance-wise, from classic Transformer-based, text-focused LLMs.

DeepSeek's free 685B-parameter AI model runs at 20 tokens/second on Apple's Mac Studio, outperforming Claude Sonnet while using just 200 watts, challenging OpenAI's cloud-dependent business model.

Dion Harris of Nvidia helped us piece together the context for Nvidia CEO Jensen Huang's GTC 2025 keynote.

Many technological shifts have coincided with economic downturns. Will 2025 be the year AI not only augments jobs but starts to replace them?