US Energy Department Launches Genesis Open Models Initiative
The U.S. Department of Energy opened a program to build open-weight AI models for scientific research, starting with Genesis-Science-1, developed with startup Arcee AI.
Read more →AI Research
The science behind the headlines: new models, research papers, and capabilities — summarised in plain language, linked to the original work. We cover academic labs and industry research teams advancing machine learning, language models, computer vision, and AI safety. Each piece connects technical findings to practical implications for developers, businesses, and policymakers.
Combine with:
Jeff Dean Leaves Google to Co-Found Discovery Loop
Jeff Dean and three other longtime Google AI researchers have left to launch Discovery Loop, a startup that aims to automate scientific research with AI.
Read more →Thinking Machines Releases Inkling-Small, Outperforms Larger Model
Thinking Machines Lab's second open-weight model needs a quarter of its predecessor's active parameters yet edges past it on reasoning and coding benchmarks.
Read more →OpenAI Says Unreleased Astra Model Solved Ten Math Problems
OpenAI says an internal, unreleased model called Astra produced machine-checked solutions to ten long-standing problems in mathematics and theoretical computer science.
Read more →Claude Mythos AI Finds Flaw in Post-Quantum HAWK Cipher
Anthropic says its unreleased Claude Mythos Preview model autonomously found a flaw that halves the effective security of the post-quantum HAWK signature scheme, a weakness two years of human review had missed.
Read more →OpenAI Launches ChatGPT for Academic Researchers Program
OpenAI is giving free access to advanced ChatGPT models to 10,000 academic researchers this summer, aiming to expand to 100,000 by 2027, backed by a $250 million research commitment.
Read more →NVIDIA and KAIST Launch $300 Million Joint AI Lab in Seoul
NVIDIA and South Korea's KAIST are building a $300 million joint research lab in Seoul to develop agentic AI models tailored to the Korean language, backed by yearly compute grants and researcher funding.
Read more →Claude Fable 5 Helps Disprove 87-Year-Old Jacobian Conjecture
Anthropic mathematician Levent Alpöge used Claude Fable 5 to find a counterexample to the 1939 Jacobian conjecture, and mathematicians including Terence Tao independently verified it within days.
Read more →Anthropic Launches $50,000 AI Grants for Rare Disease Research
Anthropic is offering researchers and early-stage biotechs up to $50,000 in free Claude access to accelerate rare disease research, with partners including the Monarch Initiative and Every Cure.
Read more →
What Is the Turing Test — and Has Any AI Actually Passed It?
The Turing Test asks whether a machine's conversation can fool a human judge. Here's how Alan Turing's 1950 thought experiment works, and what a 2025 study found when it tested GPT-4.5.
Read more →SAP Completes €1B Acquisition of Tabular AI Pioneer Prior Labs
SAP has finalized its purchase of Prior Labs, the German startup behind the TabPFN model, and will invest more than €1 billion over four years to build a frontier AI lab for structured business data.
Read more →
What Is Google DeepMind? The AI Lab Behind Gemini and AlphaFold
Google DeepMind is Alphabet's AI research lab — the team behind AlphaGo, AlphaFold, and the Gemini model family. Here's what it is and how it formed.
Read more →Moonshot AI Unveils Kimi K3, a 2.8-Trillion-Parameter Open Model
China's Moonshot AI released Kimi K3, a 2.8-trillion-parameter open model with a 1-million-token context window — a system the company calls the world's first open 3-trillion-class AI model.
Read more →DeepMind CEO Hassabis Proposes FINRA-Style AI Watchdog
Demis Hassabis called for a US-led, industry-funded body to test frontier AI models before release, arguing AGI is now only a few years away.
Read more →What Is AI Model Collapse — and Why Does It Threaten Future Models?
AI model collapse happens when generative models repeatedly train on AI-generated data, gradually erasing rare facts, minority voices, and diversity from their output.
Read more →PrismML's Bonsai 27B Becomes First 27B AI Model to Run on a Phone
PrismML has released Bonsai 27B, a 27-billion-parameter open-weight model compressed to under 4GB, small enough to run natively on a phone while keeping most of its reasoning ability.
Read more →Thinking Machines Lab Releases Inkling, an Open-Weight AI Model
Mira Murati's startup released Inkling, a 975-billion-parameter open-weight model built for developers to fine-tune and adapt, rather than to top leaderboards.
Read more →
What Is Synthetic Data — and Why Does AI Training Depend on It?
Synthetic data — information generated by AI itself rather than collected from the real world — is quietly becoming a core ingredient in how AI models are trained.
Read more →
What Is the AI Talent War — and Why Does It Matter?
AI labs are offering researchers pay packages worth hundreds of millions of dollars. Here's why the fight for talent has gotten this extreme, and what it means for everyone else.
Read more →
What Is AI Drug Discovery — and How Does It Work?
AI models now help scientists find and design new drug candidates in weeks instead of years. Here's what AI drug discovery is and how it actually works.
Read more →AI Drug-Design Startup Chai Discovery Raises $400M Series C
Chai Discovery, an AI startup that designs antibodies and other molecules from scratch, raised a $400 million Series C led by Index Ventures at a $3.8 billion valuation.
Read more →