Learning Brief — 2026-09-02

What we covered

• AI news: AI & Tech Developments
• PM news: From Product Talk (Teresa Torres): Get My Eval Test Harness and Run Your First Experiment
• PM learning: From Product Talk (Teresa Torres): Get My Eval Test Harness and Run Your First Experiment

Mental model

Make the next experiment small enough to learn quickly.

Summary

TechCrunch — AI: OpenAI’s new reasoning technique alarms AI safety experts. OpenAI’s new Astra model will use “recurrent depth,” a technique that allows the model to operate outside of the sequential thinking that characterizes most reasoning models. The Verge — AI: Google says its new Gemini 3.8 Flash model ‘works harder’ but might cost more. Google launched Gemini 3.8 Flash, arriving just a few weeks after its predecessor. The company claims the new model "works harder" than Gemini 3.7 Flash by performing more reasoning steps on complex tasks and "calling tools iteratively." It has the same introductory pricing as 3.7 Flash, $0.75 per m The Verge — AI: Amazon’s AI assistant can now spot fake emails from the company. Amazon is trying to combat impersonation scams with a new feature that allows you to use its AI assistant to determine whether an email, text message, or phone call actually came from the company. With the update, you can ask Alexa for Shopping about a message you received, and it will use AI to com VentureBeat — AI: Stolen Claude session cookies can reach corporate Gmail through grants no IT admin can revoke. Infostealers replayed stolen Claude session cookies into paid accounts without ever touching the login page two-factor authentication guards. The accounts Anthropic flagged were card-billed, self-serve accounts, which is the population no corporate identity provider governs, and no admin console can TechCrunch — AI: Pangram’s Max Spero on why AI detection is harder than ‘Real or Fake’. The internet has a trust problem, and it’s not just because social media feeds are filling up with AI slop. AI-generated text and images are now making their way into job applications, product reviews, and even insurance claims, leaving platforms and users alike scrambling to figure out what’s real. TechCrunch — AI: We’re ‘dangerously close’ to dead internet theory, says Pangram’s CEO. The internet has a trust problem, and it’s not just because social media feeds are filling up with AI slop. AI-generated text and images are now making their way into job applications, product reviews, and even insurance claims, leaving platforms and users alike scrambling to figure out what’s real.

Today's PM news comes from Product Talk (Teresa Torres). The story is Get My Eval Test Harness and Run Your First Experiment. For a senior PM, the key takeaway is to watch how product decisions and strategy plays out in practice — not just the headline, but the reasoning behind it.

Today's PM learning comes from Product Talk (Teresa Torres). The core idea is Get My Eval Test Harness and Run Your First Experiment. What matters for a senior PM is translating the idea into a repeatable habit: define the user problem clearly, pick a metric that actually reflects value, identify the riskiest assumption, and make the next experiment small enough to learn quickly. Good PM content is rarely magic. It is usually disciplined prioritisation dressed up in nicer prose.