is an agent an intern with urgency?

i’ve been sitting with this for a while. in the past, the role of an intern was simple; they worked under the CEO, and did offloaded work to them that they didn’t want or have the time to do. think buying a cup of coffee, respond to emails, take meeting notes, or learning new projects.

nowadays, i’d like to think some of that is still true to an extent—but now we have artificial intelligence and agents that submerge the internet to do these things for us.

the definition of an agent is this: a software system that uses artificial intelligence to pursue goals and complete multi-step tasks on its own. not to be confused by chatbots that sit passively to feed you information crowdsourced from data on the web, but takes actions. the fundamental core of an agent was built to do these things:

1. automate workflows. agents are connected to your existing tech stack to execute multi-step processes without human intervention. this is based off of trigger actions with conditional logic.
2. research and assist. agents browse the web, parse pegabytes of data and synthesize the core insights from it, so you don't have to.
3. execute transactions. agents interact with APIs and payment gateways to make purchases based off of pre-approved parameters or policies that a human sets.

you have granola to synthesize meeting notes upon activation, claude to run your tasks for you, clay to reach out to leads on LinkedIn.

but the thing about an agent is, it is only able to do what is set within its given premise. the context that the prompt sets for it. that’s where a lot of human input comes in to avoid breaking things. similarly, you tell your intern to go out and do a specific task, which they can either do right after stockpiling you with many questions or asking another reddit thread as a workaround because you’re too busy. other times, they’ll just execute on the task completely wrong. additionally, most agents don't have the typical guardrail to possess persistent memory or continual learning, which should be self-improving over time.

i think that agents shouldn’t be completely interchangeable with interns, but it should be augmented into an intern's everyday life (in the case they master AI literacy) so that they can focus on owning work that is far more valuable and on ROI in skill building. maybe experiential learning, strategic thinking, and creative problem-solving domains to execute on for thought.

because if ai agents are the new intern, what does that leave the interns with?