I Let AI Make A Podcast About My Career

I decided to run a quick experiment the other day about how AI might help coach me in self promotion. The goal wasn’t to create some amazing podcast just about me everyone would want to hear.

The main goal was to hear how someone else might sell and promote my experience and success. I wanted to hear new ways or angles of promoting myself. I did produce a podcast episode too though, it’s pretty awesome and I thought NotebookLM sold me pretty well.

The two tools I chose to run this experiment through were NotebookLM and Bench.io. I added the same documents to both systems – my resume, LinkedIn profile, and this website. These were documents that highlighted my accomplishments throughout my career. They both had very different outputs when it came to the audio. NotebookLM used a conversational set-up between 2 virtual hosts to have a back and forth discussion about my work, while Bench just had a single voice storytelling about my career and accomplishments.

Hearing is Believing

More than Just a Pretty Voice

While the difference in audio output is interesting, I think the more interesting aspect here is the difference in the way the experience approaches next steps.

NotebookLM gives a few actions of what to do next such as adding a note, or creating a mind map. While these can be helpful, they’re not contextual to the content or smart about what someone might want to do next.

Bench takes a different approach and thinks about context management. It anticipates next steps you might want to take with what it made. It gives you instant actions to make it shorter, provide a transcript for posting, and create assets for social media. Bench takes it to the next logical level, anticipating what you might want while also showing you what it’s capable of helping with, adding to the discovery of capability.

Best of Both

I preferred the output of NotebookLM but I appreciate the interaction and design of Bench more. I think what Bench is exploring is really smart and will help people get more value out of the tool. Its context-driven actions help expose what the system can do instead of just being an open prompt of unknown possibilities.

I’ll need to experiment some more and see if I can push Bench into more of the conversational output NotebookLM produced. I really just ran with the first output from each on this test.

This experiment revealed something important about the evolution of AI tool design: the format of output matters, but how tools anticipate your next steps matters more. NotebookLM created better content for me this time but then stops short. Bench creates something adequate but takes it further and helps with next steps to actually use it.

As these tools evolve, the winners won’t just be those with the best generation capabilities – they’ll be the ones that understand the full workflow of what users are trying to accomplish. Context-aware actions beat feature lists every time.