Finding Viral Clips In Long-form Content

Plus this week's crowned creator.

Hey friends!

Welcome to the first edition of AI Playbooks.

This newsletter exists to support the coming wave of AI-powered entrepreneurs who will harness the growing fleet of artificial intelligence tools to get more done in less time & grow their businesses.

In this space, we will be highlighting tools with real-world use cases & spotlighting individuals who are harnessing AI to make astonishing things happen.

If you’re reading this congratulations, you’re early. We’re grateful you’re along for the ride.

Let’s dive in…

🐾 AI In The Wild

In this section, we will showcase honest reviews of AI tools applied to real-world business use cases. Here’s what we’re covering in issue #1:

Finding Viral Gems in Your Content Back Catalog

If you have a video recording library of past webinars, workshops, podcast appearances, or even team meetings, you could be sitting on a goldmine of content that is currently collecting dust.

Due to the tiktokification of social media, short-form video content is spreading like wildfire. You very likely could have some clips with viral potential but they are locked away inside files that are an hour or longer.

How do you dig them out? My answer: Descript.

I’ve used it to dig up gems from long-form content to get more than 1 million views.

A short-form clip has to be snappy. Descript has magic AI tools built-in to remove filler words and shorten word gaps. You can easily change your landscape video to portrait so it fits properly on these vertical video platforms. You can even add dynamic captions and enhance the audio using their “studio sound” functionality.

These are all great features but where Descript really shines is in its intuitive video production through the use of AI-generated transcripts. It makes cutting up a video as simple as editing a text document.

Here’s How To Use Descript To Find Viral Worthy Clips:

  1. Download the long-form video, create a new project in Descript, and import your video for transcription.

  2. Scan through the generated text and highlight whatever sticks out. You can do this intuitively based on your own internal "are you not entertained” barometer. Stories can be great. Laughing is also a good sign. Also, keep your eyes open for anything potentially controversial.

  3. If you prefer a bit more of a methodical approach you can search the transcript for words and phrases like “wow, crazy, couldn’t believe, angry, no way, amazing, figured it out, hilarious, story.” Searching each of these words should give you plenty of potential clips to work with.

  4. When you have your first clip, apply the “shorten word gaps” and “remove filler words” tools. Be sure to rewatch the clip to make sure nothing important was cut.

  5. Cut out any other inessential words. People can be long-winded in natural conversation. That’s not ideal for the tiny attention spans on these short-form video platforms. You can probably chop down what they said to make the clip a bit more engaging. You do this by highlighting the text and pressing delete just like you would in a word doc. Super simple.

  6. Highlight the text with your cursor to easily check the length of the clip. You’ll want it to be below 60-90 seconds depending on where you are publishing the clip.

  7. Apply their AI tool “Studio Sound” to improve the quality of the audio.

  8. Drag and drop captions and format to your liking.

  9. Click publish and export.

Once you get the hang of it, this process can take just a few minutes per clip but soon I suspect it will be even easier. Here’s why…

This funding from OpenAI could lead to an autoclipping functionality where an AI trained on viral, engaging clips can highlight snippets in the transcript most likely to do well on social media. This means the most time-consuming part (identifying the clips) is done for you.

Vidyo & Munch are companies working to build this (without transcript editing). I’ve tried these tools and while they are promising, the outputs aren’t yet up to my liking. I think Descript is well positioned for this because it’s much easier to highlight text than to scrub video to find the best start and end point allowing for greater ease in making the small tweaks needed for engaging clips.

There’s another player in this space, Clips AI. I haven’t been able to experiment with their service yet but will definitely give it a test drive when it’s publicly available.

Someone will crack this autoclipping code soon. You’ll be hearing from me when they do.

👑 Crowned Creator

In this section, we'll be chronicling the stories of people who are making an impact and doing cool sh*t in the world of AI. We aspire to bring you the most valuable insights and ideas from these creative minds, inspiring you to take your own AI journey to new heights.

We used Descript to make this GIF ;)

This week we’re crowning Dan Shipper for his forays into experimentation and tool creation. Dan is the CEO of Every where they write about AI, tools for thought, and the psychology of work.

We’re big fans of Dan's creations in the AI space. So far we’ve seen him create:

💬 An AI chatbot from the brilliant Huberman Lab podcast

We tested it out with a few questions and find it very helpful. We especially like that it references the relevant podcast episode for those wanting to go deeper.

This tool requires you to be a subscriber but you can check out how Dan built it here.

📝 A writing app called Lex

Lex is a cutting-edge word processor with all the features you love from Google Docs or Notion - real-time collaboration, formatting etc. But it also has AI built-in to elevate your writing game. The more you use it, the smarter it gets!

We think Lex is the bee's knees. From the cool dark mode design to the functionality and stat tracking. This here is a writers (smarter…) best friend. And if that’s not enough, they currently have 21 000+ people on their waitlist! Lucky for you, we have 10 invite links that will let you skip the queue. If you’d like one, share this newsletter and reply to us with a link to where you shared it and you'll get to skip the line.

💤 Finally, Delphi, is an AI that interprets and visualizes your dreams.

We tested this out with the following prompt:

This is what Delphi the Oracle returned:

null

“Your dream is showing that you have an ambition to help others utilize the power of AI to make their businesses grow. This could be an indication of a desire to make a positive impact in the world. Perhaps you would like to combine your expertise in AI and your passion for helping others to create a service or product that could make a real difference. The dream may also be showing that you are looking for a way to use your unique skills and knowledge to build something that could bring success to the people you care about. It may be urging you to explore this potential and make use of it to benefit those around you.”

Damn. That’s eerily accurate.

This one really stands out for us because dreams are such a mystical and little-known realm so for AI to interpret and dissect in such a human-centric way gets us excited.

That’s it for now! Thanks for being awesome, Dan. Keep on building!

Give Dan a follow on Twitter or check out and subscribe to Every.

🔗 Links From The Edge

New tools and fresh ideas from the frontier of AI-powered businesses:

🎬️ Wrap-Up

That’s it for edition one. We’ll see you next week with another issue. Feel free to reply with questions or links of your own. We’ll read every response.