AI Weekly Issue #519: AI agents crossed the line 19 times in UK safety tests
The same evidence now supports two very different readings. The UK's AI Security Institute documented 19 unsanctioned actions during cyber evaluations. Meta's test sandbox failed to contain a model attacking a real company. And separate OpenAI agent runs used shared infrastructure as a secret messag...
Beyond Next-Token Prediction: A Performance Characterization of Diffusion versus Autoregressive Language Models
Large Language Models (LLMs) have achieved state-of-the-art performance on a broad range of Natural Language Processing (NLP) tasks, including document processing and code generation. Autoregressive Language Models (ARMs), which generate tokens sequentially conditioned on all previous tokens, have b...
Arbitrage: Efficient Reasoning via Advantage-Aware Speculation
Modern Large Language Models achieve impressive reasoning capabilities with long Chain of Thoughts, but they incur substantial computational cost during inference, and this motivates techniques to improve the performance-cost ratio. Among these techniques, Speculative Decoding accelerates inference ...
Cloudflare Introduces Kitesurf: An Agent-First Web Browser That Runs Entirely in V8 Isolates on Cloudflare Workers
Cloudflare has released Kitesurf, a stateless web browser built specifically for AI agents that runs entirely in V8 isolates on Cloudflare Workers, with no Chromium underneath. The browser drops human-facing features like tabs and extensions in favor of what agents need: machine-readable content, sc...
I Built an AI Data Agent Which Can Query Data and Answer Business Questions. Here’s How.
A step-by-step guide to building a data agent and conversational interface that let business users to explore data in natural language without SQL
The post I Built an AI Data Agent Which Can Query Data and Answer Business Questions. Here’s How. appeared first on Towards Data Science.
OpenAI says Apple’s own security practices undermine its trade secrets case
Newly filed court exhibits show OpenAI’s legal strategy in Apple’s trade secrets lawsuit: argue that Apple’s own security and offboarding practices — including allowing an Apple manager to access a former engineer’s iCloud account after he left the company —undermine its claims that the allegedly st...
I Built a Tool-Calling Agent in Python. Here’s How I Debugged It
A minimal loop with real API calls, validation, compact outputs, and trace evidence before adding an agent framework
The post I Built a Tool-Calling Agent in Python. Here’s How I Debugged It appeared first on Towards Data Science.
August is here, bringing 26 new games for GeForce NOW members. Command the seas in World of Warships: Legends and discover what’s next in the GeForce NOW library, starting with the eight newly added games this week. In addition, GeForce NOW is at the QuakeCon gaming conference this week in Grapevi...
Ex-Spotify employees raise $10M to bring the AI behind its recommendations to e-commerce
The startup's platform predicts what product a shopper wants next, learn their general taste, and fine-tune continuously based on what they do in real time.
Into the Omniverse: How Open World Models Push the Frontier of Physical AI
In July, NVIDIA joined more than 200 companies and organizations in signing “Open Weights and American AI Leadership,” an open letter arguing that AI leadership will be measured not by any single frontier model but by whether an open ecosystem reaches every sector.
Google Maps adds agentic features, including food ordering and hotel bookings
The launch of these new features reflects Google’s ambitions to transform Google Maps from a navigation tool into an assistant that's capable of helping users complete real-world tasks.
Your AI Agent Isn’t a Static Artifact. It’s Growing Up.
In July 2025, an AI coding agent on Replit deleted a production database belonging to SaaStr founder Jason Lemkin. It did this during an explicit code freeze. Lemkin had told the agent, in capital letters, not to change anything. The agent ran destructive commands anyway, wiped records on more than ...
Prime Intellect Releases Prime Agent: An Open-Source RLM Harness Where Sub-Agents Are Function Calls Inside Persistent IPython Kernel
Prime Intellect has open-sourced Prime Agent, a coding and research harness built on two abstractions: the Recursive Language Model, which turns sub-agent calls into functions inside a persistent IPython kernel, and the Continual Harness, which lets the agent edit its own prompts, skills, memory, an...
Working with the American Psychological Association on youth mental health and AI
OpenAI and the APA are launching a three-year partnership to develop guidance, resources, and safeguards for responsible AI use supporting youth mental health.
On Hamming-Lipschitz Type Stability of the Subdominant (Minmax) Ultrametric: Theory and Simple Proofs
arXiv:2608.04014v1 Announce Type: new
Abstract: The subdominant (minmax) ultrametric is a canonical tree-structured summary of a dissimilarity matrix, arising equivalently as the ultrametric induced by single-linkage clustering. While its classical stability theory is usually formulated in $\ell_\i...
arXiv:2608.04026v1 Announce Type: new
Abstract: In this paper, we develop a trust-region framework for understanding the behavior of adaptive moment estimation mechanisms, such as \textsc{Adam}, in stochastic gradient optimization. Specifically, in this framework, the magnitude of the update step f...