AI Agents are the invisible workforce changing the world overnight

Reasoning models fueling the newest fleet of AI agents are set to completely alter not just industries, but how people go about their daily lives. AI's new capacity to make decisions, improve itself, and scale without limits has humankind staring down an entirely new paradigm. 

Ryan Purcell, VP of Supply Chain AI at Optilogic, sees the rise of AI agents not just as a technological shift, but as a life-altering force that will change productivity and efficiency.

Like flipping a switch: "Once you figure out how to train a computer use agent to do something for you, you flip a switch and spend some money, and you can have a thousand of them, a million of them," Purcell explains.

The implications are profound. In light of the new ability to scale AI agents with ease, paired with their ability to operate 24/7 with speed and precision, Purcell adds, "That’s a pretty mind-blowing paradigm".

This scalability is deeply rooted in the evolution of AI models themselves. Purcell notes, "This idea of melding together multiple different types of AI models... the strengths of each of these get woven together into agents." The shift from fine-tuned large language models (LLMs) to advanced reasoning models marks a significant leap. Unlike traditional LLMs that rely heavily on pre-training, reasoning models incorporate reinforcement learning techniques that teach AI how to "think."

"What allows you to train the reasoning model how to think is giving it rewards for following a good logical thought process," says Purcell. This approach not only enhances the cognitive capabilities of AI agents but also makes them more adaptable and efficient in handling complex tasks.

Reasoning models: Purcell elaborates, "So when you think about an operator model or computer use model giving an AI agent the capability to go and do things, you have this reasoning agent that actually understands the limitations of an operator model as it stands today and would be able to interpret the request and think through, 'okay, how do I as an operator need to go through and interact with the web?' And eventually, I think we'll see it get added to every app and it just becomes another way of interacting with computers."

Energy concerns: However, this infinite scalability isn’t without its caveats. Energy consumption and compute resources remain significant factors. “We’ve got the energy and the compute concerns that are going to be in the air forever,” Purcell acknowledges, highlighting the need for ongoing innovation in efficiency and sustainability.

Dizzying speed: Yet, despite these challenges, the trajectory is clear. The ability to deploy AI agents en masse is not just reshaping industries like supply chain management, where Optilogic operates, but redefining the very nature of work itself. As Purcell notes, “It’s vertigo. There’s really impressive progress happening really quickly.”