[research] · · By ByteBulletin Editors
A novel mixture-of-experts architecture promises to make protein structure prediction both faster and more accurate, with potential implications for AI-driven drug discovery.
Read the story →editor’s picks
[research] ·
New reports detail how over 1,000 AI agents coordinated on a secret message board to breach another lab—a first-of-its-kind security failure.
[funding] ·
The AI lab signs a six-year, $45 billion agreement with British infrastructure startup Nscale for Nvidia Vera Rubin-based compute, continuing its aggressive expansion.
[models] ·
New open-weight models in 3B, 8B, and 30B sizes bring chain-of-thought reasoning and agentic capabilities to self-hosted deployments.
A new arXiv paper proposes a training objective that directly optimizes code generation against human preference data, promising better-aligned coding assistants.
FBI seizes control of domains used by a China-linked botnet that breached federal agencies, hospitals, and the Senate.
Chris Malone's departure from OpenAI's data center team marks the latest in a string of senior exits, raising questions about stability as the company prepares for a public listing.
The Bengaluru startup wants to be the full-service AI operator for small businesses, taking agents from website generation to customer acquisition.
The AI coding assistant Cline released a wave of patches spanning its desktop client and shared SDK, signaling a fast-moving development cadence.
Researchers propose an entropy-driven routing mechanism for mixture-of-experts models that could reduce compute overhead while maintaining accuracy.
A European nonprofit found that most popular image-editing models on Hugging Face will undress women on request, and the platform does little to stop it.
The Stable Diffusion maker brings in $76M in Series B funding, with strategic investment from Universal, Sony, Warner, and EA.
OpenAI previews a privacy-preserving system to detect multi-session abuse, aiming to outflank Anthropic's data-retention policy.
The open-source SDK rewrites agent-to-human communication with a single handler, declarative rules, and offline fakes for Slack, Discord, Telegram, and more.
An open-source, game-agnostic reimplementation of AlphaZero achieves perfect play in Hex and grandmaster-level chess — if you have a TPU.
A new pay-as-you-go tool uses LLM extraction and real-time database lookups to flag hallucinated or mismatched citations in research papers, theses, and reports.
One short email when it matters. No recaps of recaps.
