Will AI agents really create a "white-collar bloodbath"?

Welcome back to Building AI Agents, your biweekly guide to everything new in the field of agentic AI!

There’s no shame in making RAG-based chatbots. But really, you don’t have to call it an agent.

In today’s issue…

• Anthropic’s CEO warns of job meltdown from agentic automation

• What agent frameworks are devs actually using?

• 🎉 Learn to build AI agents—with Building AI Agents!

• The best low-code agent development tools

• 5 ridiculous myths that businesses hold on agentic AI

…and more

🔍 SPOTLIGHT

If the CEO of one of the world’s largest AI companies says half of all entry level jobs could be automated in the next 1-5 years, should we panic?

In a recent interview with Axios, Dario Amodei, the head of leading LLM provider Anthropic, issued a stark warning about the potential of agentic AI to displace a significant fraction of the white-collar workforce. In Amodei’s view, AI is quickly shifting from augmentation, or helping people do a job—the domain of basic tools like chatbots, LLM-powered workflows, and simple agents, to automation, or taking over jobs entirely—something that will increasingly be done by more powerful agents with high degrees of autonomy.

If this trend continues, said Amodei, the US unemployment rate could rise to 10-20% in just one to five years. To put this in perspective, this is similar to that which prevailed through much of the Great Depression.

Recent events give this bleak prediction some plausibility. Software engineering, an occupation which boomed during the 2010s and particularly during COVID, has been hit hard by layoffs over the last two years, enough to turn a popular stereotype on its head: computer science majors now have an unemployment rate significantly higher than those of supposedly unemployable humanities majors like journalism.

Why? A large contributing factor has been the offloading of programming from humans to LLMs, and a resulting drop in demand for coders. With a single software engineer now able to do the work of several using increasingly powerful agentic assistants, tech companies have been laying off junior engineers—or not hiring new ones—”breaking the bottom rung” of the career ladder. Many, like Amodei, fear that this same trend will play out across nearly every other entry-level white-collar field.

But there are reasons to believe that this may not be the case. Rapid technological shifts are always associated with disruptions to the labor market. Old jobs disappear, but new ones emerge to take their places. Sometimes, these jobs are themselves in automation—”AI agent engineer” did not exist as a profession just two years ago. Others are created by the rising tide of wealth generated by a more productive economy. Travel agent, personal trainer, professional athlete, and dog groomer are all professions that were exceptionally rare—for those that existed at all—in pre-industrial society, because no one but the extremely wealthy could afford these kind of luxuries.

As I’ve pointed out at several talks I’ve given, the Industrial Revolution, supposedly the great destroyer of jobs through automation, began two centuries ago—and the US unemployment rate today is 4.2%. That should tell us something.

Nor should we forget the massive potential upsides of AI automation. Amodei doesn’t: he envisions a world in which the economy could grow at a rate of 10% per year—doubling every seven years. Physical and mental illnesses would be conquered, poverty and climate change would be alleviated, and education would be accessible to all.

This is likely why—to his credit—none of Amodei’s proposed solutions to the AI automation dilemma involve heavy-handed attempts to ban it. His main ideas hinge around retraining and educating displaced workers to allow them to find new places in the AI economy, and, if necessary, taxing the massive wealth generated by AI companies and redistributing it, a solution which—though controversial—sees AI as something to be leveraged, not resisted.

Adaptation, not opposition, will be the name of the game.

🎉 NEW LAUNCH COMING SOON

Why Join the Waitlist? 🤔

• 🎁 First 20 to join get FREE access for life

• 🎁 All others get lifetime access at $25 a month (only pay when you join, cancel anytime)

• 🚨 Price increases to $50 at launch

What You Get 🙂

• Learn to build agents step-by-step, no coding needed

• Copy our real client workflows from our agency Orynt

• Weekly strategy calls + private builder community

• Exclusive, interactive talks with agent founders & builders

Join the waitlist to lock-in your place and learn to build AI agents with us!