News & analysis, rated.
Breaking AI developments, in-depth guides, real-world case studies, and analysis — each one rated so you know what matters.
Gartner Mumbai Conference: Day 2 Ops and Cloud Priorities
Gartner's IT Infrastructure conference in Mumbai wrapped day two with sessions on operations strategy and cloud deployment. Here's what infrastructure leaders discussed.
Gartner Summit Day 3: Security Leaders Debate AI Risk and Compliance
Gartner's 2026 Security & Risk Management Summit in National Harbor wrapped with talks on AI governance, threat modeling, and regulatory pressure. What practitioners need to know about enterprise security in an agentic world.
Holo3.1 runs computer-use agents locally, cuts step time to 3.3 seconds
Hugging Face's Holo3.1 adds mobile support, quantized weights, and local inference options. The 35B model improves mobile accuracy from 67% to 79.3% and cuts latency 2× on consumer hardware.
Reachy Mini robots can now call web tools without downloading code
Hugging Face Spaces now work as remote tools for the Reachy Mini conversation app via Model Context Protocol. Add weather, search, or custom logic by referencing a public Space — no local Python needed.
DPO cuts text degeneracy 59% in OCR models without human labels
Hugging Face shows Direct Preference Optimization reduces repetition loops in structured document extraction by using a model's own failures as training signals. Results consistent across five architectures.
NVIDIA's Task-Seeded Data Lifts GPQA by +11.1 Points in Nemotron Training
NVIDIA synthesized Q&A from 70 public task families to improve reasoning and knowledge tasks. GPQA jumped 11.1 points; MMLU-Pro +1.8, code +1.9. Here's the pipeline that works late-stage.
OpenAI maps policy push on AI safety, youth, and workforce shifts
OpenAI published a public policy agenda covering AI safety oversight, youth protection online, worker transition support, and international standards. What the company is actually prioritizing—and what it's leaving for others to solve.
OpenAI proposes federal framework for frontier AI safety and security
OpenAI released a blueprint for U.S. governance of frontier AI systems, proposing federal oversight on safety, resilience, and national security. Here's what the proposal covers and who it affects.
Wasmer Built a Node.js Edge Runtime 10x Faster With Codex
Wasmer used OpenAI's Codex to accelerate Node.js runtime development, shipping in weeks instead of months. How code generation cut engineering timelines and what that means for edge infrastructure teams.
OpenAI Adds Biology and Chemistry Tools to GPT-Rosalind
OpenAI has expanded GPT-Rosalind with enhanced biological reasoning, medicinal chemistry analysis, genomics work, and experimental design capabilities. Life sciences teams can now run more complex workflows in a single model.
Endava builds AI-native delivery using ChatGPT Enterprise and agents
Endava is automating software workflows with OpenAI's agents and ChatGPT Enterprise. See how the consulting firm is reshaping internal culture and client project timelines around AI-native practices.
Google funds virtual power plants to power data centers
Google and Voltus are building a 100-megawatt virtual power plant in PJM, the US East Coast grid, paying customers to shift EV charging and thermostat use. The project launches in 2027—but participation rates suggest incentives alone may not be enough.