AI Shrinks Jobs 16% & Meta Loses AI Talent

Hello!

Keeping up with AI is starting to feel like a second job, or that’s what 50% of leaders surveyed by LinkedIn admit. And 40% say the breakneck pace is already affecting their well-being. That’s why in this issue, I’ll be focusing on how AI affects our work. 

We’ll investigate a Stanford study showing jobs in AI-exposed fields have shrunk 16%, why Meta’s Superintelligence Lab is losing talent two months in, and Dropbox and GitHub’s playbooks for achieving AI adoption across engineering. 

And if you need a breather, Google’s nano-banana Gemini 2.5 model is giving Photoshop a run for its money.

Grab your coffee, relax, and enjoy Frictionless!

In the Queue

• One Gemini Prompt = 5 Drops of Water

• Cursor-Powered Migration Tutorial

• Website for the UK’s Top Insurer in 6 Weeks

• Accountability Can’t Fix Everything

• Is SaaS Doomed?

• Lessons from Shopify: Building Agentic Systems

• Guide to Service Reliability Metrics

• The Most Important Question in the C-Suite

Reduce Friction

2 Sprints, 6 Weeks, 3 Platforms Unified: Inside Admiral’s MVP Launch Strategy

Admiral needed a fast and scalable way to unify three previously disconnected services. In 6 weeks, the Pagepro team delivered a modern MVP with Next.js and Sanity, giving their team full content control and faster time to market.

Canaries in the Coal Mine? Six Facts about the Recent Employment Effects of AI

Companies might be facing a future talent drought. A new Stanford study reveals employment shrank 13–16% among young workers in AI-exposed fields ChatGPT arrived.

The Most Important Question in the C-Suite: Can We Trust Our Data?

Chasing after quick ROI can backfire. The successful companies of the future are the ones where the C-suite understands data strategy is a business strategy - and it's their responsibility to get it right.

The Problems That Accountability Can’t Fix

Accountability usually drives ownership, yet failures can’t always be neatly tied to one person. Let's look at situations where accountability doesn’t help to address the problem, and how to approach them instead.

Deepen Your Expertise

Building Production-Ready Agentic Systems: Lessons from Shopify Sidekick

“Death by a thousand tools” is what almost killed Shopify’s AI assistant, Sidekick. As it grew more capable, its architecture became unmanageable. To solve this, Shopify's team had to completely reinvent how instructions, evaluation, and AI training worked.

Build vs Buy in the Age of AI

SaaS is doomed, or so the story goes. AI makes building easy, but the “build vs buy” debate is far from over. Complex business logic still gives SaaS its edge. What’s changing is how we’ll build, integrate, and use it with AI agents.

Refactoring a Next.js & Tailwind app with Cursor

How do you update a two-year-old Next.js app in under an hour? Lee Robinson from the Cursor team walks through migrating a Next.js 13 + Tailwind 3 project to the latest stack with AI assistance.

What Are SLOs, SLAs, and SLIs? A Complete Guide to Service Reliability Metrics

An hour of downtime can cost up to $5 million. That’s why SLAs, SLOs, and SLIs matter. Explore what they are with benchmarks, error budget examples, and learn how they shape engineering trade-offs and customer trust.

AI Corner

Introducing Gemini 2.5 Flash Image, State-Of-The-Art Image Model

Move over, Photoshop! Google’s new image model might be the most advanced tool of its kind yet. It keeps characters consistent across scenes, supports precise edits through natural language, and is already making waves on social media.

Meta Superintelligence Labs is Losing Key Staff Less Than Two Months After Launch

After poaching top AI scientists from rivals, Meta’s Superintelligence Labs are bleeding talent amid hiring freezes and reshuffles. With the Llama “Behemoth” project abandoned, many are wondering what Zuckerberg’s next move will be.

Driving AI Adoption at Dropbox: A Conversation with CTO Ali Dasdan

Is 90% AI adoption possible? Dropbox CTO says yes. Ali Dasdan shares how strong leadership, custom-built tools, and a culture shift turned AI skeptics into daily users.

GitHub’s Internal Playbook for Building an AI-Powered Workforce

According to GitHub, your AI adoption will fail without rewiring how people work. Take a peek into their internal playbook to see how executive backing, grassroots advocates, and clear guardrails turned AI into an everyday tool.

Just Cool

How Much Energy Does Google’s AI Use? We Did The Math

A single Gemini AI query uses about five drops of water. Google’s numbers show a 33× energy and 44× carbon efficiency gain in a year, which makes AI inference less thirsty than me after a good workout.

Let’s Stay in Touch! 📨 

Do you have any comments about this newsletter issue or questions you want to ask? Drop me a message or book a meeting.