OpenAI opens Decisions API decision model to all customers
OpenAI opened Decisions API to all customers on Thursday. It is not a model specially trained for decision-making, but a version of ChatGPT Luna reconfigured for speed and cost, according to an OpenAI product lead; performance metrics vary by workload.
OpenAI opened Decisions API, a reconfigured version of ChatGPT Luna optimized for speed and cost, to all customers; Speak says it cut the time between actions from 1.4 seconds to 0.13 seconds.
02INDUSTRY
Bloomberg: DeepSeek raised about $15 billion
Chinese AI company DeepSeek raised about $15 billion (100 billion yuan), Bloomberg reports, naming CATL and Tencent as potential investors. The company is reportedly increasing spending on AI infrastructure and open models; DeepSeek did not immediately respond to a request for comment.
DeepSeek raised about $15 billion in investment after U.S. restrictions, Bloomberg reports; CATL and Tencent are among potential investors.
03PRODUCTS
NVIDIA adds training and simulation tools to Cosmos
NVIDIA expanded the Cosmos platform with tools for physical AI data, training, simulation, and deployment. The update includes five Data Factory templates and Cosmos 3 Nano models for testing autonomous-vehicle and dual-arm robot policies in simulation. NVIDIA says up to 2x faster inference in NIM applies to validated configurations.
NVIDIA added data-preparation templates, model-adaptation tools, and new models for testing physical AI policies in simulation to Cosmos.
04DEVELOPER TOOLS
OpenShell offers an isolated runtime for AI agents
OpenShell is a runtime designed to run autonomous AI agents in isolated environments and restrict file, system-call, and network access through policies. The project says it can review policy changes with formal verification and require human approval for risky access; Windows support via WSL 2 is experimental.
OpenShell 0.1.x adds new isolation mechanisms, expanded plugin support, and new APIs.
05PRODUCTS
NVIDIA introduces Minerva observability platform for AI factories
NVIDIA says Minerva combines signals from GPUs, networks, storage, and workloads in a knowledge graph to identify anomalies and possible root causes. The company says the platform offers a natural-language assistant, IREN uses Minerva, and other companies plan to deploy it.
NVIDIA Minerva was introduced as an AIOps platform that combines infrastructure telemetry with AI workloads to identify problems early.
06POLICY & LAW
Anthropic, OpenAI, and Nvidia set lobbying spending records
According to federal disclosures, Anthropic, OpenAI, and Nvidia spent about $6 million combined on lobbying in the July–September quarter; their respective spending was $2.1 million, $1.5 million, and $2.2 million. Topics included cybersecurity, AI safety, export controls, and data-center energy costs.
Anthropic, OpenAI, and Nvidia made their highest-ever quarterly lobbying expenditures in July–September, according to federal disclosures.
07DEVELOPER TOOLS
OpenAI CLI runs API operations from the terminal
OpenAI CLI provides commands for accessing the OpenAI REST API from a terminal or shell scripts. Users can generate images, upload files, send batch requests, and inspect results; Codex CLI is described as an agent that writes and edits code in projects.
OpenAI CLI lets users generate images, upload files, and send batch requests to the OpenAI API from the terminal.
08SECURITY
Author says LLMs produce misleading PoCs for PHP flaws
The author argues that online PoCs claiming to exploit a PHP vulnerability do not work on a standard PHP installation. In the author's view, some LLMs produce misleading results by making the test server vulnerable as well; the article still recommends patching servers.
An author claims that LLM-assisted PHP vulnerability PoCs do not work on standard installations and that some create a separate vulnerability on the test server.
✓
That’s it for today.The rest of your time is yours.
Your daily technology digest, every morning in English. Free. Unsubscribe anytime.
Sources, in the open.
Stories are selected from Hacker News; accessible original articles are summarized and translated with AI. Summaries may contain errors; check the source beneath each story for details. Preparation starts daily at 06:55 Türkiye time (UTC+3), targeting publication before 07:00.
This edition covers articles shared on Hacker News on the date shown above.
Now on the App Store for iPhone
Your daily digest. Now in your pocket.
Read or listen to a new edition every morning with Kısa for iPhone, save stories for later and keep your streak.
