← Back to archive

Signals for 2026-07-31

Published 2026-07-31T08:16+02:00

10 geselecteerde signalen uit de lokale hybride Daily Signal Brief pipeline.

Investigating three real-world incidents in our cybersecurity evaluations

Simon Willison

It happened again! This is turning into something of a pattern.

#evals #research-evals

Frontis-MA1: Training an AI4AI Model towards Recursive Self-Improvement in Machine Learning Engineering

arXiv reasoning / agents / evals

Recursive self-improvement (RSI) requires AI systems that improve the process of building AI (i.e., AI4AI); machine learning engineering (MLE) offers a concrete, executable testbed for studying this capability.

#agent #builder #evals #research-evals #systems-framing

Advancing the price-performance frontier with GPT‑5.6

Simon Willison

Huge price drop from OpenAI today: GPT-5.6 Terra got a 20% reduction, and GPT-5.6 Luna got a massive 80% drop.

#builder #implementation #implementation-adoption

AI systems and the reproduction of (standard) language ideologies in World Englishes

arXiv reasoning / agents / evals

The rapid growth of large language models (LLMs) has resurrected age-old questions in sociolinguistics and world Englishes, such as who decides what counts as legitimate English, whose English is suspect etc.

#builder #evals #implementation #research-evals

PAIChecker: Uncovering and Checking PR-Issue Misalignment in SWE-Bench-Like Benchmarks

arXiv reasoning / agents / evals

SWE-bench-like benchmarks are widely used for evaluating LLM's issue resolution capability. They typically follow a common construction pipeline: each PR (Pull Request) is paired with its linked issue by extracting issue references from the PR description; the issue description is used as the problem statement, and the PR patch serves as the test oracle.

#agent #evals #research-evals

Okta buys AI security startup Permiso — source says for about $200M

TechCrunch AI

The deal gives Okta identity threat detection capabilities as enterprises seek to secure AI agents and other non-human identities across cloud environments.

#agent #agentic-workflows #implementation

A fundamental flaw leaves LLMs strikingly vulnerable to attack

MIT Technology Review AI

It is impossible to make large language models fully secure against hacks because of a fundamental flaw in how they work, a team of researchers argue in a paper presented at the International Conference on Machine Learning, a top AI conference, this month. The claim has huge implications for the safety of this technology, which…

#evals #research-evals

Language models can't spark scientific revolutions, but world models might

The Decoder

Can language models spark a scientific revolution? In a position paper titled "LLMs can't jump," Google Deepmind's Tom Zahavy argues they can't.

#evals #research-evals

Pangram says its new AI text detector makes only one mistake per 24,000 documents

The Decoder

Pangram 4 detects 99.66 percent of AI-generated text with just one false positive per 24,000 documents, the company claims. The model also resists "humanizer" tools that disguise AI writing as human.

#builder #tooling-runtime

Forward-deployed engineers are the AI industry’s latest talent obsession

TechCrunch AI

A new study estimates only 2,000 U.S. engineers have the expertise to deliver meaningful AI ROI, as enterprises race to hire forward-deployed engineers to implement AI at scale.

#implementation #implementation-adoption