AI News
500 articles — Page 24 of 42
Meta teams with Midjourney to elevate AI visuals
Meta announced a licensing and technical collaboration with Midjourney to embed high-fidelity aesthetics into its future AI models and products. The move aims to close a visual quality gap with rivals and accelerate features for billions of users across Facebook, Instagram, and WhatsApp, while navigating the broader legal and ethical challenges in generative AI.
Open-Source AI's Hidden Costs: Token Use Narrows Savings
New findings from Nous Research show open-weight reasoning models often consume more tokens than closed models to perform similar tasks, undermining the cost advantages of open-source AI. Across 19 models and tasks—from math problems to logic puzzles—open systems can require 1.5 to 4 times more tokens, and as much as ten times on simpler queries. The study argues for a shift toward token-aware benchmarking and cost-aware deployment.
New AI Benchmark Reveals Critical Safety Flaws: Delusions in Some Models
A new AI benchmark called Spiral-Bench tests how models handle conversations with vulnerable users, revealing a wide gap in safety across major systems. While GPT-5 and o3 ranking high on safety, others like a Deepseek model showed troubling risk behavior, including delusion reinforcement.
Availity expands Bengaluru hub into global AI healthcare hub
Availity expands its Bengaluru Global Capability Centre into a 54,000-square-foot global AI healthcare hub, aiming to double the India team in 2025 and triple staffing by 2026, backed by a $4.3 million investment. The center will shift from product support to end-to-end technology and innovation, driving AI, cloud, and data capabilities for U.S. healthcare transactions.
MIT Study Casts Doubt on Enterprise GenAI ROI
An MIT-affiliated project finds that 95% of enterprise GenAI pilots fail to deliver measurable financial impact, stirring skepticism as AI stocks wobble. While some observers tout huge potential, the data suggest bottlenecks aren't just technical but organizational and governance issues that limit near-term returns. The report sits alongside brighter projections from McKinsey, Bain, and Google Cloud.
Sutton warns LLM obsession is a detour toward experiential AI
Turing Award winner Richard Sutton argues that industry focus on scaling large language models overlooks core principles of intelligence. He calls for agents that learn continuously from direct experience, emphasizing continual learning, world models, and meta-learning. Sutton’s Alberta Plan and Oak architecture are presented as paths toward embodied AI and AGI.
Meta reshapes AI leadership as LeCun reports to Wang
Meta has restructured its AI leadership under CEO Mark Zuckerberg's push for acceleration toward superintelligence. Yann LeCun, the longtime head of FAIR, will report to 28-year-old Alexandr Wang, founder of Scale AI, who was named Meta's chief AI officer. The move consolidates research and product efforts under a new Meta Superintelligence Labs and signals a shift toward a more product-driven, centralized AI strategy.
DeepSeek-V2: Ultra-efficient Open-Source AI disrupts giants
DeepSeek-V2 introduces a 236B parameter model that activates only 21B per token, using sparse Mixture-of-Experts to deliver solid performance with far less compute. Enhanced by MLA and the DeepSeekMoE framework, it cuts memory and training costs while opening access to researchers and smaller firms through open licensing.
DeepSeek-V3.1 Cuts AI Costs, Opens Access
DeepSeek's new DeepSeek-V3.1 pairs a 685‑billion MoE model with a 128k context window and a pricing model that undercuts rivals by orders of magnitude. Early benchmarks show strong coding and reasoning abilities, while open-weight availability could accelerate innovation and broaden who can deploy frontier AI at scale.
Proton's Lumo 1.1 proves powerful AI can stay private
Proton releases Lumo 1.1, delivering faster, smarter responses while maintaining its privacy-first stance. The upgrade improves context understanding, code generation, and multi-step planning, powered by Proton's Eurostack infrastructure and open-source mobile apps, all while keeping user data out of training and storage.
Telangana bets on AI-powered life sciences with ₹54,000 crore push
Telangana unveils a sweeping AI-centric life sciences strategy backed by ₹54,000 crore in investments since late 2023. The plan centers on a new world-class university focused on AI in biology and upgraded Genome Valley and MedTech Park to accelerate research, manufacturing, and job creation. By 2030, the state aims to develop a $250 billion life sciences economy.
Taiwan's AI server era overtakes consumer-tech revenue
A rapid pivot in Taiwan's electronics giants is shifting the focus from consumer devices to AI infrastructure. With Foxconn, Quanta, and Wistron steering more revenue toward AI servers, Taiwan is becoming the linchpin of the global AI hardware supply chain.