Every Sunday the carousel picks the AI stories that actually matter. This page has the original sources behind each one, so you can read them yourself, plus the email that sends you the list every week.
#Edition 1: week of 26 Oct 2026
1. Microsoft: 60,000 buyout offers. Microsoft offered voluntary buyouts to 60,000 employees, about 27% of its global workforce, in the US, UK, Canada and India, while it spends heavily on AI. What it means for you: big tech is moving money from people to data centres. Be the person on your team who knows how to use the AI. Source: AP, 26 Oct 2026, also FOX 13 Seattle
2. 40% of code starts with AI. GitHub's Octoverse 2026: by September, more than 40% of pull requests started with an AI agent's first draft, up from 17% in March. That is 110 million in September alone. What it means for you: typing code is no longer the skill. Explaining the task and reviewing the result is. How I keep agents honest: sangeeths.ai/harness Source: GitHub blog, Octoverse 2026, 28 Oct, notes: Simon Willison
3. A giant AI you can download. Mistral released the open weights of Mistral Large 4, a 1-trillion-parameter model. Free to use, modify, fine-tune and commercialize under a modified MIT license (one condition for companies that resell it as a service). It needs a room full of GPUs, not a laptop. What it means for you: companies can run a top model on their own servers, so data never has to leave the building. Want AI on your own laptop instead: sangeeths.ai/local Source: Mistral, 29 Oct 2026
4. AI answers eat the web. Reddit said its search traffic in October "dropped by more than 1 million referrals on average per week". CEO Steve Huffman: "The reality is, AI Mode does not drive traffic." What it means for you: when AI answers the question, fewer people click. When it matters, open the source. Source: The Verge, 29 Oct 2026
5. ChatGPT's new rulebook. OpenAI rewrote its Model Spec around five values: faithfulness to human intent, harm minimization, honesty and transparency, human agency, and care. One line matters for everyone: the model should not pick a perspective just because it thinks that's what you want to hear. OpenAI adds that its models don't fully follow the Spec yet. Try it today in ChatGPT, Claude or Gemini: "Be objective. What is the strongest argument against my plan?" Source: OpenAI, 29 Oct 2026
#How I pick the stories
- Only things from the last 7 days.
- Something you can try, read or use, not just a funding round.
- The original source, not a repost of a repost.
- No security exploits, no investment tips.
#Get it every Sunday
Put your email in the box below: one email every Sunday with the stories and the links. No spam, one click to unsubscribe.
Made by @sangeeths.ai: real AI engineering, tested live. Share this kit: sangeeths.ai/weekly · All free kits: sangeeths.ai