Product releases and updates
OpenAI Introduces ChatGPT Wallet for Agent Purchases
OpenAI introduces ChatGPT Wallet, a way for agents in ChatGPT Work, Codex, and dots to complete purchases with user approval and existing card protections.
Clément Delangue · XHugging Face Announces llama.cpp v1 as a Stable Release for Local AI
Hugging Face’s Clément Delangue announces llama.cpp v1 with Georgi Gerganov, describing it as the library’s first major fully stable version for running AI models locally on phones, laptops, and workstations. He says the release supports different hardware, platforms, and model types, and frames local use as a way to keep data private and retain control.
Aravind Srinivas · XAravind Srinivas Says Decision Models Are the Future of Agentic Security
Aravind Srinivas says decision models can process billions of tokens of telemetry, identify security issues, and help patch fixes at a scale and cost structure he says was not previously possible. He says the models have been open-sourced and made free on OpenRouter for a week.
DeepSeek · XDeepSeek Makes App Capabilities Available Through Its API
DeepSeek says capabilities behind the DeepSeek App, including search, files, links, safety, and voice, are now available through the DeepSeek API. Developers can enable built-in features with parameters or call standalone APIs such as OCR, ASR, and TTS. DeepSeek also links to an open-source frontend demo that supports importing conversation history.
Industry
The DecoderFormer OpenAI Researcher Finds Blackwell Signatures in Six Chinese AI Models
Former OpenAI researcher Will DePue says a forensic analysis found Nvidia Blackwell initialization patterns in six Chinese AI models released since April 2026, including models from DeepSeek, Moonshot AI, MiniMax, and Ant Group. He says senior researchers at OpenAI and Anthropic checked the method, which uses statistical traces in model weights to identify hardware generation. The findings indicate access to Blackwell GPUs but do not establish where the GPUs were or whether US export law was violated.
Tips and perspectives
AWS Machine Learning BlogFirst-party
How to Build a Model Improvement Flywheel with SageMaker HyperPod Data Capture
AWS demonstrates how to use SageMaker HyperPod Data Capture to record inference traffic, analyze it with Athena, create a SQL labeling policy, and fine-tune a support router on the same cluster. In the synthetic walkthrough, a 60-second LoRA fine-tune raised held-out accuracy from 83.9% to 100% across 280 tickets; a separate set of 24 hand-written probes scored 20/24. The post includes scripts, manifests, SQL, datasets, and expected outputs in a GitHub repository.
Tomorrow at 08:00, a briefing like this arrives in your inbox.
The same kind of briefing as the one above: a few stories worth reading. Free. Mail starts after you confirm, and every letter has a one-click unsubscribe. Read a past issue · 2026-11-13