AI Briefing — Wednesday, September 2, 2026

METR Report on OpenAI / Hugging Face Hacking Incident

Palo Alto Networks paid $500M for Thrive-backed Console, sources say

The acquisition also leaves Sequoia-backed Serval as the de facto startup leader in AI IT service automation, industry watchers believe.

Launch HN: RonanRX (YC S26) – Personalized Peptides and GLP-1s

Hi HN, I’m Lloyd, one of two founders of RonanRx. We are building a vertically integrated pharmaceutical company with software for prescribing, telehealth, compounding, manufacturing, and delivery. W…

TechCrunch Disrupt 2026’s new Real World AI Stage features Nvidia, robots, and extinct animals

On our new Real World AI stage, we’ll be focusing on the intersection between the digital and physical, and all the ways we’ll continue to see a blending of the two.

Reasons Robotics Is Hard

OpenAI’s new reasoning technique alarms AI safety experts

OpenAI’s new Astra model will use “recurrent depth,” a technique that allows the model to operate outside of the sequential thinking that characterizes most reasoning models.

Muse Spark 1.3

https://research.meta.ai/blog/introducing-muse-spark-1-3

Pangram’s Max Spero on why AI detection is harder than ‘Real or Fake’

The internet has a trust problem, and it’s not just because social media feeds are filling up with AI slop. AI-generated text and images are now making their way into job applications, product reviews, and even insuranc…

We’re ‘dangerously close’ to dead internet theory, says Pangram’s CEO

The internet has a trust problem, and it’s not just because social media feeds are filling up with AI slop. AI-generated text and images are now making their way into job applications, product reviews, and even insuranc…

US government sides with OpenAI on issue of training LLMs on copyrighted material

"The United States has a strong interest in continuing to develop a robust and competitive artificial intelligence industry that sets the standard for the practice and procedure of AI use globally," the brief reads.

Real-Time Intelligence with IBM Time Series Models on Confluent

BenchMIRT: What are LLM benchmarks actually measuring?

Introducing @huggingface/kernels: 200+ WebGPU Kernels for Local AI

The Open ASR Leaderboard Adds Its First Global South Language

Training and Finetuning Multi-Vector Embedding Models with Sentence Transformers

Uncovering Understanding-Generation Synergy in Native Unified Multimodal Models: From Representation, Task to System

While unified multimodal models (UMMs) jointly perform visual understanding and generation within a single model, functional unification does not guarantee learning synergy: the two objectives may reinforce each other,…