OpenAI puts ChatGPT Enterprise on AWS Marketplace and explains Control Room, the system that can stop its research models within 35 seconds. Cloudflare makes the first payments from its Vite open source fund, AWS speeds up agent cold starts with AgentCore Runtime V2, and a new OpenAI report finds that most students use AI for homework without being taught how.
AI Must read
On November 9 OpenAI made ChatGPT Enterprise available for purchase through AWS Marketplace. Companies can now buy and manage the product with the AWS account and billing setup they already use, and eligible spending counts toward…
Source: OpenAI — ChatGPT Enterprise now available in AWS Marketplace · openai.com ↗
On November 10 OpenAI published a report on how schools in Europe, the Middle East and Africa are starting to use AI, based on its work with…
Source: OpenAI — Closing the AI guidance gap in education · openai.com ↗
Buying new software in a large company often takes months of contracts and vendor reviews. Selling through AWS removes much of that step for firms that already have a deal with Amazon. OpenAI says the Marketplace version has the same models and features as a direct contract, and new features arrive at the same time, so buyers do not get a reduced edition.
Source: OpenAI — ChatGPT Enterprise now available in AWS Marketplace · openai.com ↗
AI labs train models with agents that run for long periods, and a monitor may spot risky behavior in the middle of a run. A reliable way to cut off output quickly is a basic safety control. OpenAI placed the check in the inference layer, which has only a few code paths, instead of in research code that changes often and may contain bugs.
Cloudflare argues that AI-written pull requests have increased the review load on maintainers, while the number of people doing reviews has not grown. Direct funding helps the small projects that much of the JavaScript tooling depends on. The recipients were picked by core team members of Vite, Vitest, Rolldown, Oxc and Vite+, including maintainers who do not work for Cloudflare.
Source: Cloudflare Blog — Launching the Vite Ecosystem Fund · blog.cloudflare.com ↗
Slow starts and memory billed at peak usage make AI agents costly to run at scale. In AWS tests, P75 cold starts stayed around 2 seconds for container images from 200 MB to 2 GB, while the first runtime went from about 5.4 to nearly 30 seconds. The price per unit is higher, but AWS expects most agents to use far fewer GB-hours and pay less, without promising it.
Source: AWS News Blog — AWS Weekly Roundup, November 9, 2026 · aws.amazon.com ↗
Students are already using AI every day, often without clear rules from their schools. The report argues that teachers become more important, not less, and that governments should give them training and set clear expectations for students, instead of leaving each class to improvise.
Source: OpenAI — Closing the AI guidance gap in education · openai.com ↗
AI Must read
On November 6 OpenAI published BioBasicsBench, a test that checks whether AI agents can read raw results from biology labs. It was built with the…
Source: OpenAI — Measuring how AI agents handle everyday wet-lab science · openai.com ↗
Development Must read
JetBrains released PhpStorm 2026.3 on November 5. The main addition is Air, a new space inside the IDE where developers run and supervise AI coding…
Source: JetBrains — What's New in PhpStorm 2026.3 · blog.jetbrains.com ↗
In a research post published on November 6, Cloudflare explained how it plans to make Privacy Pass resistant to future quantum computers. Privacy Pass…
Source: Cloudflare Blog — The post-quantum future of web privacy · blog.cloudflare.com ↗
OpenAI announced on November 4 that it will acquire Pierre Computer Company, maker of Code Storage and of the open-source Trees and Diffs libraries…
