Bookmarks (570)

The company developed DeepSeek-R1 by using pure reinforcement learning on top of DeepSeek-V3-Base, and matched or beat o1 on some benchmarks.

Beyond performing tasks, machines will integrate into our social fabric, requiring us to navigate new relationships with technology.

AutoGen represents Microsoft’s latest attempt to address the challenges of building multi-agent systems for enterprise applications. This article explores what this means for the state of agentic AI today, and how it compares to other major frameworks like LangChain and CrewAI. This article unpacks the implications of AutoGen’s update, explores its standout features, and situates it within the broader landscape of AI agent frameworks,...

As LLMs become more capable, many RAG applications can be replaced with cache-augmented generation that include documents in the prompt.

Devin 1.2 will help developers with accelerated workflows and less cognitive load in file searches, understanding codebases or fixing code. 

AI or Not, an AI fraud detection platform, has raised $5 million in a seed funding round to accelerate its use of "AI to detect AI" in images.

Whether for brand-consistent marketing visuals or detailed character art, the Flux Pro Finetuning API enhances precision and adaptability.

Titans architecture complements attention layers with neural memory modules that select bits of information worth saving in the long term.

Here's how robotics manufacturing company Baca Systems and corporate food ordering platform ezCater are using Salesforce's AI agents.

Enterprises can prepare for AI becoming faster, cheaper and more capable in 2025, and take the lead on developing applications.

Let's dive into the many reasons security leaders are embracing AI-native security operations centers (SOCs) as the future of defense.

AI-powered code development has come a long way in recent years. Here's where it's headed as 2025 gets underway.

The infrastructure to support AI inherits many challenges that defined industrial factories, from power to scalability and reliability.

This article is part of VentureBeat’s special issue, “AI at Scale: From Vision to Viability.” Read more from the issue here. Enterprises can look forward to new capabilities — and strategic decisions — around the crucial task of creating a solid foundation for AI expansion in 2025. New chips, accelerators, co-processors, servers and other networking and […]

Successful edge AI deployments require deep integration between edge and cloud, complex orchestration and new approaches to data management. 

Borderless AI launches an enterprise HR platform powered by AI that streamlines employee data management, policy compliance, and HR workflows, securing $32M in funding to challenge traditional HR software giants.

Nvidia is optimizing AI safety controls for multi-agent systems with minimal performance impact through new NeMo Guardrails microservices.

Botika, a started focused on AI-generated fashion models, has raised $8 million and launched its new mobile iOS app.

Impressively, all the motions in the example videos appear lifelike and fluid — with subjects moving much faster and more naturally.

Discover how AI threat modeling is helping CISOs redefine zero trust in 2025 by combating identity-based attacks.

Unlike traditional deep learning, which interpolates between data points, program synthesis searches for discrete programs that explain data.

Microsoft's AutoGen for AI agents is popular among agent developers, and a new update addresses observability and flexibility issues.

Microsoft wants to make using Copilot a daily habit and take on Google's aggressive push of Gemini for Workspace.

LangChain’s founder Harrison Chase is looking to advance agentic AI with the concept of ambient agents, which might well be the next step on the path to generalized intelligence.

Colossal BioSciences has raised $200 million in a new round of funding to bring back extinct species like the woolly mammoth.

Vellum polled 1,250 AI developers to understand what’s happening in the AI trenches, including techniques, use cases and challenges.

LLM MiniMax-Text-o1 is of particular note for enabling up to 4 million tokens in its context window — equivalent to a small library.

Hallucinations are a persistent problem in healthcare. Here's how GSK is using test-time compute scaling to improve its gen AI systems.

OpenAI will let some users schedule task reminders on ChatGPT, a sign the company may be readying itself to release an agent.

Cerebras Systems has teamed with Mayo Clinic to create an AI genomic foundation model that predicts the best medical treatments.

PC makers are betting everything on the AI bandwagon. At the CES 2025 tech trade show in Las Vegas last week, AI technology was everywhere.

Europe's Imec.xpand is spinning out memory chip firm Vertical Compute in a seed investment round worth $20.5 million.

A raft of U.S. unicorns -- those valued at $1 billion or more -- are likely to go public in 2025, according to the PitchBook/NVCA.

Even when it comes to non-reasoning LLMs such as Claude 3.5 Sonnet, there may be room for regular users to improve their prompting.

The newest version of Mistral's code generation model Codestral aims to be more efficient and generate code twice as fast.

LlamaV-o1, a groundbreaking AI model from MBZUAI, revolutionizes multimodal reasoning by providing transparent step-by-step explanations across text and visual tasks, outperforming competitors in speed and accuracy.

Agentic AI helps SOCs automate decision-making and adapt to evolving threats, streamline workflows, and improve incident response.

Up Network, the user-powered AI agent operating system, has announced its partnership with DreamSmart to create Web3 AI glasses powered by Google Gemini. This product integrates state-of-the-art industrial design, AI, XR (extended reality) capabilities, and Web3 incentives, with the aim of redefining human-machine interaction and advancing the post-smartphone era. Developed under DreamSmart’s StarV brand, the […]

Nvidia CEO Jensen Huang answered all of the press questions at an event at CES 2025. Then he took questions from financial analysts.

Adobe's new Firefly Bulk Create streamlines the tedious task of editing dozens of images for websites and social platforms.

Nvidia and Krafton unveiled a groundbreaking on-device AI that will enable smarter AI characters for PUBG and inZoi.

Unbound Academy will replace teachers with AI tools, representing the latest chapter in a 60-year quest in computer assisted instruction.

As one of 6,000 journalists at the show, I walked around a lot to find the coolest tech. My conclusion: The robots are coming.