Dear subscribers,

If you use Twitter, I’m sure you’ve come across Trung Phan’s content.

Trung grew to 600K+ followers sharing “smart threads and dumb memes.” In the interview below, I spoke to Trung about:

1. His journey from failed screenwriter to successful creator

2. His end to end process for creating viral content

3. His AI startup and how he uses AI to create compelling content

This is the funniest interview that I’ve done so far - I hope you enjoy it!

Trung’s journey from failed screenwriter to successful creator

You’ve had an interesting career. Can you describe how you ended up creating content on Twitter?

When I was in college, I thought I wanted to be a doctor. It only took a semester of failing classes that led me to give up on that dream (or parental expectation). I decided to embrace my newfound freedom and rescheduled all my courses so they began after 12pm. Long story short, I graduated from college as a history and sociology major. 

I didn’t have any skills so I decided to move to Vietnam.

After eating a comical amount of pho to drown my sorrows, I found a job as an equity analyst in Vietnam. But my real dream was to be a comedy screenwriter. I got my hopes up when my script was picked up and sold to Fox after two years of work. Sadly, it got stuck in production hell (full story here).

Four years later, I got my MBA and moved to Boston to work for a fintech company. One day, a colleague recommended that I write for The Hustle, a business and tech newsletter that had over a million subscribers. It sounded like a great place to combine my comedy writing and business background.

At the Hustle, I was encouraged to build an online personality. That led to my “smart threads and dumb memes” brand.

Trung’s end to end process for creating viral content

Can you explain your “smart threads and dumb memes” barbell strategy? 

Yes it’s called the barbell strategy because smart threads and dumb memes are at the opposite ends of the effort spectrum.

1. Smart threads: Everyone knows that threads are the way to grow on Twitter. But most threads are incredibly cringe (e.g., “99% of people don’t know this…”). I put alot of time and effort into threads that are actually entertaining and useful.

2. Dumb memes: My memes are lower effort but they help keep me top of mind with my audience. The truth is, there is more competition around memes and some folks get meme fatigue. So it’s harder to grow if you’re only a meme account. 

So basically:

“Smart threads and dumb memes” is a pretty broad niche. Many creators focus on just one topic (e.g., product management, solopreneurship) to build credibility and eventually sell a product to their audience.

What do you think about that strategy?

Well we both know that going niche works (hence the saying “niches make riches”). I just couldn’t post about one topic over and over again. It’ll bore me out of my mind. But I respect the people who do this even though sometimes their content is cringe.

Fair enough! Let’s break your strategy down with an actual example. One of your most popular threads is how Salmon sashimi didn’t come from Japan (24K likes).

How did you come up with this?

Yes let me break it down:

1. Where I got the idea: The core of that thread was from an NPR Planet Money episode. Here’s some alpha for creators - go to Planet Money or Freakonomics and skim through the title and description of each episode. Notice how they take seemingly well-known topics but introduce a big curveball, which creates the curiosity gap. This is critical and often a recipe for viral posts.

2. Create a curiosity gap: Everyone knows that sashimi is from Japan, and salmon is one of the most popular types of sashimi. But it turns out the Japanese really didn’t eat salmon sashimi until the 1990s (and it happened because of a decade-long marketing campaign by the Norwegian government to sell salmon).  When you see that headline, you’re thinking “Oh shit, I didn’t realize that!” and want to read more.

Makes sense - creating that curiosity gap is key. Do you have a content creation routine?

My routine is I do a lot of reading. I wake up and read everything from:

1. Publications like New York Times, Bloomberg and Wall Street Journal.

2. Newsletters on all kinds of topics (I subscribe to 30+ and my faves include The Diff and Money Stuff).

3. Social including Reddit’s top posts and Twitter Blue’s Top Articles.

I also do a quick timeline scroll on Twitter every morning.

In Fortnite, there’s always a new map that players have to master. On Twitter, there is a new “map” everyday too -- it’s called the main news story. When playing the Twitter game, my job is to write the funniest or most insightful tweet about that top story. Win or lose, it’s my form of fun. 

If you had to start over again as a creator today, what would you do?

Today, I think it’s a no brainer to write on both Twitter and LinkedIn.

Twitter has more ideas and better content. But LinkedIn has a huge content deficit, so it makes sense to repurpose business-related content onto that platform because you’ll get distribution. The key word being “repurpose” because you won’t have time to do both well from scratch.

I should have repurposed sooner (my audience on LinkedIn is much smaller).  

Bearly AI and how Trung uses AI to create content

Can you explain what Bearly AI is?

Bearly AI is an AI-powered research tool to help writers save hours in their workflow. It can summarize an article for you, correct your grammar and instantly pump out a first draft of an article.

Why did you decide to start an AI startup?

Many creators rely on sponsorship and ads to make money. There are two drawbacks with these channels: 

1. Sponsorships ebb and flow with the economy. The ad market is bad right now.

2. The products you promote need to match your audience. I only want to promote products that I use and love myself.

So I thought, why not build my own high-lifetime value (LTV) product that I would use everyday and have equity stake for? That’s why I started Bearly AI.

There are a lot of AI startups out there and the most common criticism is that they’re just a wrapper around OpenAI. Do you have any thoughts on what your differentiators are?

Yes, the most common question for AI startups is: “What’s the moat?”

In some ways, the criticism is fair. I think these large language models (LLMs) will be commoditized. Every browser and productivity tool will have an AI companion. 

But here’s where I think I can compete:

1. Distribution: ChatGPT reached 100M monthly active users in record time. But my guess is only 10% are using it regularly and even fewer are actually paying for it. So the opportunity is still massive for non-Tech Twitter people to integrate AI into their workflows. And a lot of those people will end up discover generative AI through Bearly AI as I aggressively shill it.

2. Usability: I want to build the easiest interface to interact with AI (you can access dozens of AI reading and writing tools with just one Bearly keyboard shortcut). A 50 year old isn’t going to join some Discord server to use MidJourney. That person can just use my product instead.

How has it been so far? Where do your customers come from?

About half are from North America. But we also have a lot of customers in Southeast Asia. There are ~700m people in that region and many sell to the West but aren’t native English speakers or writers. AI tools like Bearly will be huge for these individuals to communicate and compete globally. 

How are you using AI in your own content creation process?

I don’t use AI to write my tweets or my newsletter from scratch. I use it for:

1. Ideation: I’m sure most of your readers have seen how imaginative ChatGPT can be. Generative AI is a great way to come up with ideas.

2. Copy editing. Writing and editing are two different tasks. So after I write something, I ask Bearly AI to copy edit it in the style of AP. Honestly, it fixes grammar mistakes and sentence structure much better than something like Grammarly. It redlines everything and I just pick and choose what sentences I want to change.

3. Summarizing what I read. I read a lot and hate taking notes. With AI, I would first read the full article then ask AI to summarize it in a few bullet points.  This helps with recall or when I want to return to the subject. A common response from users is that they get “way too many” articles sent to them from friends, family and colleagues. Instant summaries is one way handle the content dump.

How do you think we’ll all create content with AI a few years from now? Where’s this all heading?

I think AI will create a deluge of content, and a lot of it will be crap.

You’ll probably get your lunch eaten if you’re just sharing straight information. So you have to differentiate based on who you are, what you care about, and what your voice is. 

I think the advice to write like how you talk matters even more with AI. Oh and don’t forget the dumb memes.

Check out Trung’s newsletter and follow him on Twitter.

If you enjoyed this post, there are two other ways I might be able to help:

1. Sponsor this newsletter: Reach 40,000+ readers including creators and employees at top tech companies.

2. Launch your PM career course: If you’re a new or aspiring product manager, sign up for the waitlist for my PM course that currently has a 10/10 rating on Maven.