PhantomFill: When the Form Demands an Answer, Language Models Invent One
arXiv:2607.20492v1 Announce Type: new
Abstract: Language models in production do not write prose. They fill forms: JSON fields, function arguments, extraction templates. We show that the form itself causes hallucination.
We ask thirteen models the same question about the same input and change onl...
DataPrep-Bench: Benchmarking LLMs as Training Data Preparators
arXiv:2607.20465v1 Announce Type: new
Abstract: The quality of training data fundamentally determines the capabilities of large language models (LLMs), yet no unified benchmark exists to measure how well LLMs, agents, and data-centric workflows actually prepare training data end to end. We view LLM...
How AI guardrails are impeding the work of offensive cybersecurity researchers
We spoke with several cybersecurity researchers, who look for unknown vulnerabilities and develop tools to exploit them, about how OpenAI’s and Anthropic’s guardrails affect their work.
Policymakers, academics, healthcare providers, AI developers, and patient advocates convened by Stanford HAI identify critical gaps in how we regulate AI tools used for therapy and emotional support.
LEAD: Breaking the No-Recovery Bottleneck in Long-Horizon Reasoning
Long-horizon execution in Large Language Models (LLMs) remains unstable even when high-level strategies are provided. Evaluating on controlled algorithmic puzzles, we demonstrate that while decomposition is essential for stability, extreme decomposition creates a “no-recovery bottleneck”. We show th...
Andrew Ng Just Released OpenWorker: An Open-Source, Local-First Desktop AI Coworker That Returns Finished Deliverables Instead of Chat
Andrew Ng has released OpenWorker, an MIT-licensed desktop AI agent that returns finished deliverables instead of chat replies. It runs a local Python agent server under a Tauri shell, supports 30 curated tool-calling models plus fully local Ollama, and gates every write, shell command and off-machi...
You Probably Won’t Read This Article…and That’s OK
“Help! There are too many [LLM bug reports, blog posts about LLM bug reports, books, treatises, codices, scrolls, papyri, cuneiform tablets]! How do I choose which to read?” —Many people, presumably Stop there! If you are reading this, ask yourself how you got here. Did Substack’s algorithm recomme...
The line in the response object You call the API. You pass model: “claude-fable-5”. You get back a completion, a token count, and a field that reads “model”: “claude-opus-4-8”. Nothing errored. Nothing retried. The request was classified before generation began, matched a sensitive category, and was...
Runway launches AI model router as generative media gets crowded
Runway no longer wants to be just another AI model company. It wants to become the infrastructure layer for generative media. On Thursday, the startup launched Runway Media Router through Runway Dev, its developer platform, released earlier this month, that provides API access to a growing roster of...
The first expensive agent run doesn’t look like a governance problem. It looks like a billing problem. A team opens its first agent invoice after the meter turns on, sorts the runs by cost, and finds one that cost 40 times the median. The provider meter shows tokens and a total. The application logs...
Most RAG Hallucinations Are Extraction Errors: Seven Patterns for a Typed Generation Contract
Enterprise Document Intelligence [Vol.1 #8ter] - Naming the RAG error correctly matters: model reads the context, so a wrong answer is an extraction error, not a hallucination. Seven typed-contract patterns keep the generation brick honest, with a decomposition rule for small models
The post Most RA...
AI chip startup Etched defies skeptics, hits $10.3B valuation from big-name investors
Etched, founded by three Harvard dropouts, has created new chips and memory components that speed up inference on any AI model -- no GPUs required, it says.
OmniVoice Studio is built on a premise that everything runs on your hardware. Voice cloning, video dubbing, real-time dictation, voice design, all of it local, all of it free for personal use, no API key required, no usage counter.
StrongestLayer Raises $9.3M Total Seed to Combat AI Email Attacks
New round led by Inovia Capital backs a reasoning-based architecture designed for attacks that bypass traditional email security StrongestLayer, the AI-native email security company, today announced it has raised $4.1 million in new funding, bringing total seed funding to $9.3 million. The round was...
This continued expansion has increased the regional customer base 72% year-over-year Kong Inc., a leading developer of API and AI connectivity technologies, today announced strong growth across Europe, the Middle East, and Africa (EMEA), driven by the rising enterprise demand for API and AI connecti...
EMeRG Announced the Launch of Hospital Intel Suite™
One of the world’s largest hospital intelligence platforms, powered by AI and predictive analytics. EMeRG today announced the launch of its advanced SaaS intelligence platform, built with a mission to make every hospital visible and every MedTech innovation more accessible. Commercial teams often la...
GeForce NOW Sets Sail With ‘Path of Exile: Curse of the Allflame’ Joining the Cloud
Lock in and load up the cloud. GFN Thursday brings fresh updates and new adventures, all ready to play without waiting for downloads. Set sail in Path of Exile: Curse of the Allflame and charge in Battlefield 6 Season 4 both launching major content for members this week. Then revisit Capcom legends ...
West Monroe Research Reveals AI Transformation Gap
Research identifies three imperatives for building an AI-native enterprise based on a survey of 417 U.S. business leaders West Monroe, an AI-native global consulting firm, today released Building the AI-Native Enterprise, new research examining how organizations are progressing from AI adoption to e...
How AI helps scientists design the next generation of medicines
Designing and developing a new medicine is an expensive, failure-prone scientific challenge. A new drug can take many years to develop, at the cost of a significant investment. And even then, most possible candidates never reach the patient. For biologic medicines, therapies made from engineered pro...
Anthropic Releases Claude Security Plugin for Claude Code in Beta: A Multi-Agent Vulnerability Scanner That Runs in Your Terminal
Anthropic has released the Claude Security plugin for Claude Code in beta. The plugin runs a multi-agent vulnerability scan of a repository from inside an existing Claude Code session, then turns the findings you select into patch files that you review and apply yourself. Anthropic emphasized the to...
FormulaSPIN: Self-Play Fine-Tuning for Natural Language to Spreadsheet Formula Generation
arXiv:2607.19354v1 Announce Type: new
Abstract: Spreadsheet applications are used by hundreds of millions worldwide, yet writing formulas remains a significant barrier. Existing approaches rely on static supervised data, which quickly saturates on limited annotations. In this paper, we introduce FO...