55th Edition Download

This Week in AI:

Klarna just showed that AI’s not a side project anymore. It’s the operating system.

Klarna just hit nearly $1 million in revenue per employee, powered by AI tools replacing thousands of hours of manual work. OpenAI quietly launched an agent that helps developers write and manage software autonomously. And YouTube is now placing ads based on emotional peaks in videos, using Gemini to find the exact moment you’re most locked in.

Let’s dive into how companies are actually using AI to get ahead.

In This Issue:

• Klarna’s AI Efficiency Play → From 700 employees to 2 AI agents. Klarna’s productivity is breaking records. (link)

• OpenAI Launches Dev Agent → Their new software agent can write, debug, and explain code for you. (link)

• YouTube Uses AI to Read Your Emotions → Gemini now predicts emotional peaks in videos to time ads better. (link)

Klarna Hits $1M Revenue Per Employee with AI

TL;DR:

Swedish fintech Klarna revealed that its revenue per employee is now nearing $1 million, thanks to a sweeping internal adoption of AI. They replaced 700 internal support staff with an in-house AI assistant built on OpenAI’s tech, claiming it now handles two-thirds of all customer interactions and is improving faster than human agents ever did.

Our Take:

This is about leverage. Klarna’s model shows how AI isn’t just a “productivity booster,” it’s an infrastructure change. For any business, especially in support-heavy or operations-heavy industries, the message is loud: the companies that scale with AI under the hood will outpace those still doing it manually. If you're not testing internal agents now, you're already behind.

OpenAI Drops Dev Agent to Build Software

TL;DR:

OpenAI just launched its “Codex” software agent, designed to write, edit, and debug code across a wide range of tasks. It can manage entire dev workflows, act as an interactive tutor, and provide detailed explanations, documentation, and error fixes without needing human hand-holding.

Our Take:

Codex isn’t just another assistant. It's OpenAI moving toward full-stack AI teammates. Ones that don’t just autocomplete your code, but manage it, reason about it, and explain it back to you. For startups, this could become your first technical hire. For engineers, it's a new kind of pair programmer. And for product teams? It's the missing glue between idea and MVP.

YouTube’s Ad Placement No Powered by Gemini

TL;DR:

At Brandcast 2025, YouTube revealed it’s using Gemini AI to analyze videos in real-time and identify "peak points", the most emotional, suspenseful, or thrilling moments. It then triggers ad placements immediately after those points, dramatically improving engagement and ROI for advertisers.

Our Take:

This is AI doing media math at scale. YouTube is turning every video into a mapped-out emotional graph and placing ads at the exact moment your guard is down. For marketers, it’s a dream. For viewers, it’s a reminder that attention is now algorithmically priced. The bigger takeaway? Emotional intelligence is no longer just human, and it's already being monetized.