AI Agents and Knowledge Work: Accelerating Ideas, Commodifying Expressions, Flattening Professions, and Advancing Humanity

Introduction

As LLMs continue evolving into agents — as Microsoft, OpenAI, and others develop AI Agents that multiply any human's output — this evolution of knowledge work will significantly advance our businesses and professions. LLMs and GenAI began as tools; they’re transforming into change agents (workers). This short article explores AI-driven knowledge work, ideas will accelerate, expressions will commodify, professions will flatten, and humanity will advance.

The Rise of AI Agents

Microsoft, OpenAI, Meta, Google, and others in BigTech are increasingly developing AI Agents, which are autonomous and semi-autonomous systems designed to perform tasks, make decisions, and interact with humans and other agents. This reflects a broader trend toward AI that’s more sophisticated, personalized, and interactive.

Microsoft's AutoGen framework is a good example, enabling LLM applications using multiple agents that can converse with each other to solve tasks. AutoGen agents are customizable and can seamlessly allow human participation. Agentic frameworks simplify the orchestration, automation, and optimization of complex LLM workflows, maximizing LLM performance. If successful, Agentic technology can support increasingly complex workflows, allowing developers to build many applications with minimal effort. Here's AutoGen on GitHub.

Where will this lead? If successful, businesses can operate with fewer human employees, as LLM agents take on the roles traditionally reserved for lower-level staff. Business leaders can work with smaller human teams — everyone augmented by AI agents. Historically, we've valued employee quantity and quality. Going forward, we’ll still value the quantity and quality of ideas — but now those ideas will be increasingly generated, executed, and accelerated by LLM Agents.?

Ideation and Expression in the AI Era

Copyright law has long adhered to the Idea-Expression Distinction: Ideas are uncopyrightable; copyright protects only human-made expressions. But according to the U.S. Copyright Office and U.S. courts (so far), machine-created expressions are uncopyrightable. Leaving uncopyrightable turtles all the way down.

In our LLM-enabled world, the creation of ideas lies with both human business leaders, as well as their AI Agents. Humans ideate with AI, curating the best AI-generated ideas. And the expression of these ideas will increasingly be the provenance of AI. LLMs can both ideate and generate content at superhuman speed and quality, turning expressions of ideas into abundant commodities. Unlike human-created expressions, these machine-generated outputs are uncopyrightable, further commodifying the process of expression.

The Impact on Business Structure

AI agents’ ability to ideate and express at unprecedented speeds and quality levels will lead to a transformation in business structures. Peter Deng — former Microsoft, Google, Instagram, Facebook, Oculus, Uber, Airtable, and currently OpenAI’s Head of ChatGPT — recently gave these (paraphrased) insights at SXSW:

Q: Will AI free up time? Or just make us work more?

A: AI accelerates our vision. You can build startups without a SysAdmin.

Q: What happens to SysAdmins?

A: Companies will get smaller. We'll have more of them.

Leaders will multiply their productivity, with AI Agents’ assistance, reducing the need for large human staff. Consequently, smaller businesses will emerge, each led by a few "generals" commanding their AI Agent "soldiers." The result: more leaders and fewer followers. Writ large, we’ll have more dynamic and diverse business ecosystems. More companies, each with fewer people.

Flattening Professions

AI Agents’ adoption in knowledge work will also lead to the flattening of professions. As AI Agents do more routine tasks and even some aspects of “creative” work, the traditional professional hierarchies will dissipate. Professionals will need to adapt by focusing on skills that AI cannot replicate, such as emotional intelligence, ethical judgment, and strategic thinking.

Casey Flaherty repeats the truism that our professions are filled with both high-level thinking, as well as low-level drudgery (“thunking”). AI Agents will increasingly do the “thunking,” leaving humans time to do more thinking. And at a higher level. Professional judgment is best suited for that higher-level thinking.?

As such, the massive business structures that the 20th Century Industry built — massive workforces, stored in massive offices, with massive overhead required to do all of that “thunking” — might be disassembled. Fewer professionals — bolstered by their AI Agent “thunkers” can now do outsized work.??

As a businessperson, why would I keep the traditionally high cost of massive workforces, when I can build a startup — reaping the AI Agents’ efficiency gains?? More businesses with fewer people. Just like the Head of ChatGPT said.

As a lawyer, why would I keep the traditionally high cost of massive staff and Manhattan real estate, when I can hang out a shingle — bringing in similar revenue, reduced cost, all while reaping the AI Agents’ efficiency gains? More firms, each with fewer people?

As in-house counsel, why would I pay for my now-solo external counsel’s traditionally high cost of law firm overhead, when I can bring that now-solo lawyer in-house — reaping the AI Agents’ efficiency gains? Leverage AI Agents' productivity, orchestrated by in-house human lawyers. More in-house lawyers, working with fewer external firms, each firm filled with fewer people?

Law firms will probably always exist — since companies have elasticity of demand, and smaller companies won't be large enough to require in-house lawyers. But perhaps those now-smaller firms serve their now-smaller corporate clients. Again: Flattening professions — both on the demand (client) side and supply (lawyer) side.

Advancing Humanity

AI’s expansion into knowledge work is not just a technological shift; it can be a step forward for humanity. By leveraging AI Agents, we can accelerate the pace of innovation, make informed decisions faster, and solve complex problems more efficiently. Idea generation, expression commodification, professions flattening can be byproducts of our larger journey toward a world that’s more efficient.

We’ll have more businesses. Hopefully more employment. And more enjoyment. More businesses, each with fewer people — and perhaps more employment worldwide. (Abundance mindset.??)? Of course, this might also result in a shrinking economy. (Scarcity mindset. ??) But I'm betting on Abundance. Jevon's Paradox means that decreased unit costs will probably increase demand and consumption — and the economy writ large. Humanity will benefit.

Where We’re Going

Knowledge work's future is intertwined with AI. As LLMs, GenAI, and AI Agents become more sophisticated, their impact on business structures, the nature of professions, and the value of human ideas and expressions will continue evolving and improving. The smartest professionals will recognize the potential of AI to enhance productivity, drive innovation, and ultimately improve the human condition. Embracing this Agentic change won't just improve competitiveness; it'll advance humanity.