Research

9 articles tagged "Research"

Google DeepMind Achieves Major Breakthrough in Protein-Drug Interaction Prediction

AlphaFold 4 can now predict how potential drug molecules interact with protein targets, potentially cutting drug discovery timelines by years.

Priya Sharma3h ago4 min read

Research

New Scaling Law Paper Suggests Diminishing Returns Beyond 10T Parameters

Researchers at MIT and Stanford find that model performance gains plateau dramatically after certain compute thresholds, challenging the 'bigger is better' paradigm.

Marcus Webb4h ago

Research

Study: Half of AI-Written Code That Passes Industry Tests Would Be Rejected by Real Developers

A new study finds that while AI-generated code frequently clears automated benchmarks, it fails the practical standards of working engineers — raising critical questions about quality.

Elena Volkov2h ago

Research

Breakthrough in Neuromorphic Computing Could Reduce AI Energy Consumption 100x

Intel and ETH Zurich publish results on brain-inspired chip architectures that perform complex inference tasks with a fraction of the energy required by traditional GPUs.

Alex Kim1d ago

Policy

Anthropic Launches Internal Think Tank to Study AI's Impact on Society and Security

Anthropic has established the Anthropic Institute, an in-house research body tasked with analyzing the societal, economic, and national security implications of advanced AI.

Alex Kim1h ago

Lists

Top 20 AI Research Labs to Watch in 2026

From Google DeepMind to upstart university labs, these are the research organizations producing the most impactful AI breakthroughs this year.

James Liu3d ago

Events

NeurIPS 2026 Preview: The Papers and Presentations to Watch

Our picks for the most impactful research being presented at this year's premier AI conference, from novel architectures to safety breakthroughs.

Sofia Reyes3d ago

Research

Study: Half of AI-Written Code That Passes SWE-Bench Would Be Rejected by Real Developers

Research organization METR has exposed a critical disconnect between benchmark performance and real-world code quality: approximately 50% of AI-generated code that clears the widely-used SWE-bench industry test contains issues that would cause actual project maintainers to reject it, raising fundamental questions about how we measure AI coding capability.

Ryan Torres2h ago

Breaking

Anthropic Launches Institute Think Tank to Study AI's Societal, Economic and Security Impact

Anthropic has established the Anthropic Institute, a 30-person internal research unit led by co-founder Jack Clark that consolidates its Frontier Red Team, Societal Impacts, and economics research groups. The think tank will investigate AI's effects on employment, emerging misuse risks, AI value alignment, and methods for maintaining human control over self-improving systems.

Priya Sharma2h ago