Well, this one will be short, but heck, whatever. No one reads this blog anyway.
As someone who has been working for almost a decade—sheesh, I feel old typing this sentence—I believe the benchmark for any knowledge worker or junior engineer should be redefined after the emergence of AIs like ChatGPT or Claude.
If you’ve paid attention in your professional life, you must have worked with colleagues who just slack around but somehow manage to stay in their positions. This phenomenon is also known as “quite quitting,” as news put it. I get it—people sometimes lose their drive and passion the longer they work in the same company. And sometimes, people treat a job as a job, which is fine, btw. But these people annoy those who genuinely enjoy what they’re working on, and their pure existence is a financial burden to the company.
Companies should consider hiring university students as interns to replace those workers. There are three reasons:
First, they’re cheaper and eager to learn.
First, because they are interns, their salaries are expected to be lower than those of full-time employees. In addition, their contracts are easier to terminate if there is no cultural/skill fit. Fun fact: about ten years ago, you could even hire interns for free (at least in Taiwan)– I don’t think that’s a thing anymore.
Back in college, I was driven and eager to learn new things during my internship. Not only because I wanted to make it in the industry but also because I aimed to transition to being a full-time employee after the program. Time may differ, but each generation must have this kind of people, regardless of macroeconomic conditions.
Second, the mentoring cost is now reduced.
In the past, hiring an intern meant sudden productivity drops from his/her supervisor, as interns needed someone to guide them in familiarizing themselves with the company/culture/industry.
Companies can now use AIs trained with internal documents, industry reports, or news to help interns get off and running. Additionally, with a proper feedback process, async collaboration tools like Loom can help interns and their managers discuss things without disrupting their schedules.
Third, the talent pool is bigger now.
In my intern days, most companies (at least in Taiwan) had never experienced remote work before—I doubt any company did it back then. However, because of a mysterious virus with an unknown origin, many companies have tried or used it since then.
With a remote working scheme set up, the talent pool for potential interns is higher. Heck. If the position doesn’t require knowledge of a specific market, you can hire someone in countries close to your time zone.
This all sounds neat on paper. Are there any proven results?
Well, the company I’m working with is hiring an intern based on this belief. I will be the manager for this role, so I guess I will find out soon.
Update in September 2024:
After working with at least two interns, I’m confident to say that my theory is correct, with a little caveat. First, you still need to hire capable and proactive interns to make this work. Otherwise, less capable ones will create more garbage due to the AI hallucinations. Second, interns’ experience with AI plays a significant role. They tend to be less productive if they haven’t used it before. In my expereicne, some pick it up fast while some somehow struggle to write prompts (like how?).
Parting Thoughts
But what does this mean for a knowledge worker?
Before I do any task these days, I always ask myself, ” Is this job a college student with an AI can do?” If the answer is positive, I will delegate the tasks to my subordinates or junior colleagues. If they’re occupied, I will try to use AI to reduce my time on this task—the 80/20 rule, you know.
Another thing is that current AIs are more like auto-complete systems, which can be affected by the majority of the opinions on the internet when the neural network is trained; you can be contrarian with your domain knowledge/data analysis from internal data sources to provide value to companies, albeit AIs can do the same thing by modifying the prompts.
