Ask The Game, the Build Log

A Curious Fix for Summaries That Might Be a Product

I wasn’t planning to build a new feature. Now, I think I build a business.

I was reviewing the default summary that Deepgram generated from The Game Podcast episodes, and honestly, it felt like filler. It was generic, vague, and completely missed what made the episode valuable. So I put together a quick script, just a rough little experiment, to see if my entire system, with the help of ChatGPT, could produce a smarter summary.

It worked. Like, really well. It caught me off guard so well.

Why it worked (and why I think this matters)

The summary wasn’t better because of GPT. It was better because of all the context I had already prepared. The transcript wasn’t raw. It was cleaned. The speakers were labelled. I had metadata on who was speaking, what the episode was about, which business topics were covered, and even where the key transitions occurred. All of that came from the ETL pipeline I’ve been slowly building behind the scenes.

The summary worked because the data was rich. Not just formatted, but structured in a way that helped the model reason. And that made me realize: this isn’t just a one-off trick. This could be a service.

The experiment

I ran the summary script on one episode—just one—to see if I was imagining things. The prompt was simple:

The result captured Alex’s core message better than most show notes I’ve seen. It highlighted positioning strategy, named real examples, and surfaced sharp, tweetable lines—all without me tweaking anything.

It felt like a human editor had stepped in and said, “Here’s what you need to know.”

Before and After

That's what the default RSS summary

In this reshare episode, Alex (@AlexHormozi) sits down with the hosts of Iced Coffee Hour to talk about building multi-million dollar businesses, why most founders focus on the wrong things, and the surprising habits that make a top-tier entrepreneur. This is Part 1 of a two-part reshare.

Welcome to The Game w/Alex Hormozi, hosted by entrepreneur, founder, investor, author, public speaker, and content creator Alex Hormozi. On this podcast, you'll hear how to get more customers, make more profit per customer, how to keep them longer, and the many failures and lessons Alex has learned and will learn on his path from $100M to $1B in net worth.

Wanna scale your busine [...] blablabla.

Well, ok. Sure.

That's the Deepgram auto-generated summary

The speakers discuss the challenges of starting businesses and the importance of building a brand and getting paid for it. They stress the benefits of investing in a high ROI business, the importance of speed, ease, and ease in salesperson training, and the importance of belief, attitude, and success in businesses. They also discuss the impact of the gym industry on behavior and the importance of building a network to gain more skills, limiting the amount of time spent on distractions, and finding a job that is rewarding and provides a high likelihood of success. They suggest limiting the amount of time spent on distractions and practicing music for longer periods of time, finding a job that is rewarding and provides a high likelihood of success, and realizing that people should have a belief in themselves to hold on to actions and not use the word "m manifesting."

That could apply to literally any episode. No punch. No precision.

Here the Enhanced Summary

In this 80-minute interview on The Iced Coffee Hour, Alex Hormozi shares his contrarian approach to building multi-million dollar businesses. The conversation reveals why focusing on being "the best" is often counterproductive and how strategic positioning can outperform pure excellence.

Key Discussion Points:

Alex Hormozi on Strategic Positioning: Hormozi explains that success comes from choosing the right market and timing rather than being the absolute best at execution. He shares specific examples of how he built businesses in competitive spaces by finding underserved niches rather than trying to out-compete established players.

The Hosts Challenge Traditional Thinking: Guest_1 pushes back on conventional startup advice about perfectionism, while Guest_2 explores how this applies to different industries. Their questions reveal practical applications for entrepreneurs at various stages.

Business Scaling Insights: Alex details his framework for rapid business growth, emphasizing speed of execution over perfection. He discusses specific metrics and decision-making processes that helped him scale multiple companies to 8-figure revenues.

Actionable Takeaways: The conversation includes concrete strategies for identifying market opportunities, building sustainable competitive advantages, and maintaining growth momentum without burnout.

This episode demonstrates why Alex's unconventional approach to entrepreneurship has generated massive results, offering listeners a blueprint for building successful businesses without the traditional pressure to be "the best" in every aspect.

No tweaking. Just raw output from the prompt and pipeline.

I plan to collect more of these before-and-after comparisons to test consistency. But even at this stage, the difference is clear.

What this unlocks

This changes how I think about summaries. They’re not just recaps. They’re onboarding for new listeners. They’re metadata for search. They’re raw material for social content. And when done right, they make the podcast way more usable.

If I can automate this—across hundreds of episodes (it's really tempting), then I’ve got something bigger than summaries. Not sure why. Or ... maybe a small SaaS hiding in plain sight?!

What’s next

I’m going to try this on more episodes. I want to test if it holds up across different guest formats and episode structures. I’m also considering testing various versions of the prompt. Could I generate different styles? Or ... a tweetstorm version?

I’ll also need to think about tracking the prompt version and metadata used for each run. If this becomes a feature (or a product), I need it to be reproducible.

Why this matters to me

This wasn’t a feature I meant to build. It was just a curiosity. But now it feels like it could be the start of something real.

– Benoit Meunier