#21: AI + Work

The age of productivity

Here’s a quote that’s worth internalizing.

“AI won’t take your job. But someone who understands how to use AI will.” – Scott Galloway

It’s no secret that AI is rapidly changing the way we approach work. From hiring decisions to day-to-day tasks, AI applications enable employees to dramatically increase their level of productivity. The good news is that most of the tasks that can be automated are tasks that you probably didn’t enjoy doing to begin with. And the even better news is that if you’re a reader of this blog, you’re going to fall into the group of people who understand how to use AI.

Now, AI is challenging the status quo of all areas of work, but let’s call out a few key areas and discuss them in detail.

Meetings

My least favorite type of meeting is the one where I’m responsible for presenting and taking notes. It’s an undertaking that is nearly impossible to do without either the presentation or the quality of the notes suffering. Go heads-down on the presentation? Awesome. But unless you have a perfect memory and jot down notes in the middle of your afternoon back-to-back meetings, your audience likely won’t get much out of those notes. And neither will you.

Thankfully, several meeting notetaking software companies have emerged over the past few years. Companies like Fireflies, Otter, and Granola have popped up to ease this burden.

I’ve used Fireflies before. It does a decent job taking notes, though sometimes it misses the context or nuances in a back-and-forth discussion.

If I were to make a switch, I’d make it to Granola. I haven’t tried this out yet as the software is only compatible with iOS (Macs only, smh), but its notetaking ability looks unmatched. Granola considers the notes you took (say in a digital notepad) and pairs it with the audio from your meeting to output an incredibly extensive sheet of information.

But that’s just where it starts. Granola allows you to ask questions about the notes and provides answers based on the context of your meeting. Forget what John really meant by his point on achieving that KPI this year? Ask Granola, it’ll tell you all about it. AI notetakers are the ultimate software win, making employees more effective and more efficient.

Interviews
AI’s impact on interviews is funny, fascinating, and eerie, all at the same time. Listen to this.

Roy Lee founded Interview Coder, an AI application that provides interviewees live answers during programming interviews. Want a high-paying coding job, but don’t have the chops yet? Enter Interview Coder to help you “fake it ‘til you make it”.

Yeah, so Lee was a college student at Columbia University when he launched his app and was subsequently kicked out for “academic dishonesty”. Which I understand. Lee did create software that allows users to cheat on interviews, leading to companies hiring unqualified candidates.

But is that the employers’ fault for letting technology evolve faster than their hiring process? Leveraging AI is soon to be a requirement for most jobs (if not already), so it might as well fool proof the process. I’ll let you decide where the moral dilemma lies, and who should bear the blame.

Also, I think the types of questions we ask in white-collar interviews are going to shift. In 1-2 years from now, I could see myself asking rather AI-heavy questions of candidates. Here’s a few I’ve come up with:

· You realized that you’re spending a lot of time taking meeting notes, researching takeaways from the meeting, and drafting follow-up emails. Walk me through a way you can automate most of this workflow.

· Talk to me about the differences between two different LLMs (your choice) and which tasks you prefer to use one versus the other for.

· Tell me about a time you used an AI agent builder to automate a task or uncover insights.

But wait… what’s stopping an interviewee from leveraging software like Interview Coder from answering the above questions and completely fooling you?

Fair point. I’ll likely need to pick and prod if answers sound too polished or templated. However, what may be interesting is if in-person interviews become more of a norm again. From what I’ve seen, in-person interviews seem to be saved for the final round. If AI continues to provide a glitch in the interview matrix, employers may be forced to require candidates to show up in-person for additional rounds of the interview process. This will certainly extend recruitment cycles, as well as challenge remote or hybrid work.

Note: Don’t worry about Lee not graduating from Columbia, as he’s doing just fine. He recently raised $5M to build Cluely AI, which provides sales teams with real-time intel to guide them through the call.

Workflow Automation via GPT

Customized generative pre-trained transformers (GPTs) will become the information hub for most companies. Don’t be intimidated by the name. Customized GPTs are pretty simple to explain. High-level, they are a version of ChatGPT that is tailored to your specific needs. Companies or individuals can leverage an Open AI model and introduce specific information for the model to train and learn from. Users can combine their proprietary knowledge with the power of Open AI’s slate of models, in a safe and secure way.

Here's an example of a GPT I’m thinking of building in the context of e-commerce.

The goal here is to build a model that quite granularly analyzes trends in keyword-level performance on Amazon to help inform future pricing, promotion, and merchandising decisions.

I’d like to feed weekly Amazon keyword-level performance data, going back a few years and subsequently analyze changes in keyword performance over the course of weeks, months, and years. To aid in GPT knowledge development, I’ll also provide context on the brands, products, and any large performance swings that came due to internal actions (not related to macro moves as I assume the GPT has knowledge of this).

By training the model on company-specific data (but not financial data, out of an abundance of caution), I’m essentially creating a data analyst on my team. I can feed this “data analyst” new information and ask for its recommendations to improve performance.

…

In one line, AI is changing work. Change how you work before work changes you.