The Archive

EchoCoT systematically extracts hidden chain-of-thought traces from proprietary reasoning models via API interaction using fidelity signals as recovery feedback.

Restricting evidence visibility across multi-module language-model societies improves compositional generalization compared to full-input exposure in relay-based architectures.

Safety Nets framework combines neural network compression with lookup tables to certify 100% correct behavior for AI systems in EASA aviation applications.

DecoVAE decomposes time series into trend and seasonal components using domain-specific inductive biases for lightweight, interpretable probabilistic forecasting.

Empirical evaluation framework audits cross-lingual fairness in LLM watermarking with calibrated thresholds and threshold-independent quality metrics across languages.

Cryptographic Context Injection is only the latest way to break an LLM safety guardrail.

Meta said its Muse Spark model is powering the dictation feature.

Slack is introducing dedicated channels where teams can vibe code together with AI agents instead of jumping between different tools and conversations. The Slack Code launch includes open, project-specific code channels with dedicated user tabs, alongside features that compare coding changes and preview HTML output before the project is shipped. "With Slack Code, when you have an idea or need to build a new feature, update a web page, or fix a bug, you simply tag in a coding agent like Anthropic's Claude or Cognition's Devin, and that agent then spins up a code channel to tackle the task," Sl...

Each day, an airline transports tens of thousands of passengers on hundreds of flights. Often these are not straightforward point-to-point routes, with passengers requiring multiple connections. The airline can consider potentially hundreds of variables to price each of these journeys: demand, season, time of day, current events, global markets, and competitor airline activity to name…

Binance's Agent OS works with tools including ChatGPT, Claude Code, and Cursor.

OpenAI launches Intelligence Age blog exploring AI's impact on power, governance, economy, and freedom.

Z.ai CEO Jie Tang discusses GLM 5.3 model and post-training scaling laws as alternative to parameter growth.

Stampli compressed product launch timeline using ChatGPT Work and Codex for development acceleration.

What does a payments giant want with a startup that routes prompts between different AI models? Stripe says it's because of "the singularity" but it's really for a far more real and powerful reason.

Simon Willison evaluates smolvm as a sandbox for executing untrusted Python/JavaScript with resource limits, demonstrating feasibility for safe user-provided task execution.

Jeremy Morrell proposes LLMs + sandboxing enable extensible web apps where users safely customize core functionality via AI-generated extensions.

Simon Willison discusses productivity metrics for AI coding agents, arguing lines of code can meaningfully measure agent output despite conventional skepticism.

NVIDIA Holoscan is a platform for building real-time AI applications at the edge, from medical imaging to robotics. HoloHub is its companion repository: a... NVIDIA Holoscan is a platform for building real-time AI applications at the edge, from medical imaging to robotics. HoloHub is its companion repository: a growing collection of reference applications and components that demonstrate what’s possible. We wanted to explore how a general-purpose coding agent could use the same examples, documentation, and development tools available to an engineer… Source

A competition is developing between OpenAI and Anthropic over who can provide the best privacy protections for enterprise customer data.

SpaceX was reportedly in talks to buy AI coding startup Cognition. SpaceX has already acquired Cursor as it races to catch up to rivals like OpenAI and Anthropic in enterprise AI.

Bankrupt Spirit accused of selling out workers in massive data sale to Google.

As AI becomes harder to avoid, consumers are growing more wary of the technology — and Silicon Valley is discovering that widespread adoption doesn’t necessarily lead to acceptance.

The launch of the new study features marks Google's latest effort to make Gemini the AI assistant that students turn to when learning and studying, as it continues to compete with companies like OpenAI.

Google Search adds study features using AI; consumer-facing educational tools rather than frontier model capability.

OpenAI expands Zero Data Retention for frontier model APIs and introduces Private Safety Processing to enable content moderation without data logging.

As we're gearing up for back-to-school season, Google is rolling out a new dedicated student hub in Gemini. It's a one-stop repository for collecting research in a study notebook, creating flashcards, taking practice quizzes, and more. Google is also enhancing its study notebooks with support for graphs and images. It can even add test dates and deadlines to your Google Calendar based on your syllabus. Google is also adding Deep Research to Gemini Live. You can ask Gemini to generate complicated research reports and talk through the results. If it's taking a while, you can close the chat and ...

Multiple cybersecurity researchers said they suddenly lost access to OpenAI’s Trusted Access for Cyber (TAC) program, which offers models with fewer guardrails for vetted users.

SPADE: self-play RL framework where LLMs generate adaptive training environments and learn from them, enabling continuous goal-distribution scaling for language agents.

ADEPT: RL framework for sim-to-real dexterous manipulation via pre-training on generic tasks and post-training on downstream behaviors using visuo-tactile perception.

On-policy distillation for long-context reasoning: group-calibrated rewards address token-level teacher bias toward local plausibility over global task constraints.