Jack Dorsey Companies Lay Off 40% of Employees Due to AI Productivity

Another canary has given its life in the coal mines. This is signal. Let’s discuss.

The desire to get more work done from fewer employees is evergreen. GenAI has had software development in its sight from the beginning and with the maturing of the tools the wave is here. Not coming. It is upon us.

Let us first discuss the disaster roll outs. Two name two: Klarna and Salesforce, but there are many others. Companies that ASSUMED there would be AI productivity gains, laid off work force in mass prior to justification, and suffered from it. Real stories, real regrets, real failure of roll outs of AI to serve that siren song of saving money on employee cost via automation. Don’t hang your hat on these stories as proof that the wave isn’t real and won’t sweep you away. Consider the failed internet companies that crashed when the bubble burst. Then asked yourself - what happened to retailers who took comfort in the failures and thought they were safe after all?

The Dorsey example is one where they deployed first, trained and learned, and saw the performance gains. He is restructuring in light of the reality that developing software is now optimized by smaller teams using agentic AI tools effectively. His example is signal to add to Andreas Kling, David Heinemeier Hansson, Christian Lattner and many others. And with Dorsey, it’s not just the coders and engineers.

What to do other than pray this isn’t so and bury heads in the sand hoping it goes away? Prepare to be one of those who is effectively using agentic AI to boos productivity and quality. What others can learn, so can you. We are still early in the transformation, it is not too late.

I wrote a where to start tutorial: https://forum.lunduke.com/t/its-time-for-humans-to-pair-up-with-ai-agents

There is good news. All of us who took the time to build the skills in our career have developed what AI doesn’t replace, but amplifies. There are people who have not programming background using these tools to “Vibe code” apps. They can’t compare to what we can do with our systems architecture, problem solving and other high level skills that we’ve been calling “coding” - but always was far more than that.

I have found learning GenAI and agentic tools tickles all the same parts in my brain that learning data architecture, object oriented programing, web development, business intelligence dashboards - and every other challenge that I’ve enjoyed for more than 40 years. I have tried, at times in vein, to communicate the counter message to “AI gets the credit for all you create with it”.

I write songs, make music, am writing a book. In this arena I’m the “vibe coder”. Creating works that I require AI to help me do because I do not have the knowledge, talent or skill. Even so, I know the part that I do add. My vision, passion, taste, heart and my lived experience. In coding - I have decades of experience that help me instantly bypass those who are “vibe coding”.

I am living what Dorsey is describing. A team of one who is producing the output of many people - anchored by the benefit of 40 years in the industry solving problems. I have gone from speaking predictions of what would happen when I saw ChatGPT 3 years ago - to witnessing what I am experiencing.

This is not a one day wave. Most of us are not going to lose our jobs this week. Here is what you should be considering.

The job market is going to be rough as companies downsize before the new opportunities arrive that will follow. The cost of creating software is going down, the demand will rise for it. But there may well be a rough transition.

Landing that new job will either require or greatly benefit from you having experience and skill using agentic AI tools.

Those who find shelter in the few companies who are taking an anti-ai stance will find tough competition to get those jobs. And those companies will eventually be out competed and those jobs will go away.

Turn your skill in finding the weaknesses in AI coded projects into a marketable skill for how to help companies increase quality WHILE using those tools.

Use AI to learn AI. Find the weaknesses, build your skills and solutions that address those weaknesses.

Or not. This is simply my read and I’ve been directing my efforts to prepare for this wave for three years and counting. What is one project you’ve had in mind all these years that you could use as the way to learn agentic AI?