AI news, November 2, 2026: ChatGPT and Codex to share one allowance, Google Cloud spend caps, DeepSeek's open super-agent prototype
Five stories from Friday and Saturday, October 30โ31, 2026 (US time) for solo makers who build with AI. Two change how you pay for and budget AI work, one is an open-source agent design you can study today, one renames the Claude plans many makers use, and one opens a large market for Apple's AI features. Figures are as the companies and outlets state them.
ChatGPT and Codex will draw from one token-based allowance starting in December
An updated OpenAI Help Center page says that from December, Plus, Pro and Business plans will share one usage allowance across ChatGPT and Codex, as OpenAI folds the agent features of Work into the main ChatGPT app. Usage will be counted in tokens, so stronger models and harder tasks use up more of it, while Instant chat will not count at all. OpenAI says it expects most people won't hit resets more often than today, and combined remaining usage and reset times will appear under Settings โ Usage. The page gives no exact start date and no allowance sizes.
Why it matters for makers: if you lean on Codex for coding and also chat with advanced models all day, both now eat the same budget, so it is worth watching the Usage page once the change lands and picking cheaper models for routine chats.
Source: OpenAI Help Center
Google Cloud spend caps are now generally available, with monthly auto-reset
Google made spend caps on Cloud Billing budgets generally available on October 30. A spend cap budget covers one project and one service, and once the estimated cost passes the amount you set, new requests to that service are paused for the rest of the month without deleting any resources; at the start of the next month the cap resets and the service resumes on its own. In this release the eligible services are the Gemini API, Gemini Enterprise Agent Platform (formerly Vertex AI), Cloud Run and Cloud Run functions, and caps can now be managed through the Budgets API. Google notes that enforcement is not instant, so overages during the delay are still billed, and fixed costs such as storage keep running.
Why it matters for makers: a leaked Gemini API key or a runaway agent loop on a side project can now be stopped by a hard monthly limit instead of a surprise bill; set the cap a little below the amount you can actually afford.
Sources: Google Cloud Blog, Google Cloud documentation
DeepSeek open-sources dsh-superagent, a research prototype for an always-on coding agent
DeepSeek's Harness team published dsh-superagent on GitHub on October 31 under the MIT license. It is a coding agent for a single GitHub repository that listens for requests on GitHub (and optionally Feishu), opens a work item, does the job in its own session and directory, and replies in the original thread, for example with a pull request, while a web panel lists what needs a human. The README stresses that it is a research prototype, not an official DeepSeek product, that much of it was written by coding agents, and that it differs from the internal version described in the team's design article. Setup is meant to be done together with your own coding agent by pointing it at SETUP.md.
Why it matters for makers: even if you never run it, the documents on authorization, execution and keeping human attention low are a useful blueprint for anyone wiring an agent to their own repos.
Source: GitHub
Anthropic renames Claude Max 5x to Max and Max 20x to Ultra
Anthropic renamed its two top consumer plans: Max 5x is now just Max at $100 a month, and Max 20x is now Ultra at $200 a month, with prices unchanged. Claude's help center says existing Max 20x subscribers move to Ultra automatically, and that Ultra includes 10 times the Pro plan's weekly usage, Claude Code, monthly Claude API credits, and no five-hour limit for verified accounts. Accounts that are not yet verified get a temporary limit until verification happens automatically.
Why it matters for makers: if you are on the old Max 20x plan for Claude Code, your bill stays the same, and you can check Settings โ Usage to see whether the five-hour session bar is already gone for your account.
Sources: The Verge, Claude Help Center
Apple Intelligence comes to China in iOS 27.3, with Alibaba's Qwen built in
Apple said on its Chinese newsroom that the first Apple Intelligence features will arrive in China with iOS 27.3, starting with a beta on November 4 and a full release in December. The launch covers earlier features such as notification summaries, live translation, writing tools, visual intelligence and object removal in photos, but not the new Siri AI. Apple also worked with Alibaba to add its Qwen chatbot to some experiences, which The Verge compares to the ChatGPT integration iOS added in 2024.
Why it matters for makers: if your iPhone app has users in China, features that rely on Apple's system writing and translation tools will start to work there, so it is a good moment to test those flows on the beta.
Source: The Verge