Hatching Growth #3: How We Rode the AI Wave with Side Projects Before It Crashed
Glasp’s note: This is Hatching Growth, a series of articles about how Glasp organically reached millions of users. In this series, we’ll highlight some that worked and some that didn’t, and the lessons we learned along the way. While we prefer not to use the term "user," please note that we’ll use it here for convenience 🙇♂️
In our first article, we shared how manual, one-to-one outreach helped us reach our first 1,000 users. In the second, we showed how we built a repeatable growth engine through “SEO++” content and word-of-mouth.
Now, in this third installment, we’ll dive into how we spotted the AI wave early, experimented with playful side projects, and positioned ourselves to launch our biggest hit: YouTube Summary with ChatGPT.
If you want to reread or highlight this newsletter, save it to Glasp.
Why Share This Now?
By late 2022, AI was about to explode into mainstream awareness. But looking back, the signs were already there: breakthroughs in text-to-image models (DALL·E 2, Midjourney, Stable Diffusion), rising excitement in tech communities, and early-access experiments with GPT models.
For us at Glasp, this wasn’t just background noise. We had already been paying close attention to OpenAI, keeping an eye on the news around GPT-2, and noticing how creative startups in San Francisco were using AI to generate ad creatives at scale.
When ChatGPT launched in November 2022, we released YouTube Summary with ChatGPT within a few days—making us the first in the world to do so. That timing wasn’t luck. It was the result of actively watching the wave before it crashed.
1. The Signals Before ChatGPT
Image generation hype (2022):
In April 2022, DALL·E 2 launched in closed beta. By July, Midjourney’s beta was live. In August, Stable Diffusion gained significant attention on social media. Suddenly, anyone could generate stunning images with a few words.This led us to a simple yet powerful realization: if AI could generate high-quality images from text prompts, then generating text from text was a natural leap.
Glasp’s product fit:
Since Glasp is centered around text highlights and UGC-driven content, we could already see how text-focused AI could amplify our product and strengthen our SEO-led growth strategy. (We’ll share more about how we leveraged UGC for SEO-led growth in another installment of Hatching Growth.)Local community buzz:
In San Francisco, AI was already a hot topic. Friends were founding AI startups, raising capital at high valuations, and experimenting with new creative tools. Even in programs like ODF, where founders shared emerging trends and insights, we could feel the growing buzz around GPT models—and that helped strengthen our conviction. (By the way, we interviewed , the founder of ODF, on Glasp Talk. Check it out if you’re interested!)
2. Side Projects: DALL·E Meets Wordle
Before ChatGPT, we built a side project that combined the viral puzzle game Wordle with AI image generation.
Here’s how it worked:
DALL·E 2 generated four images from a single text prompt.
We asked users to guess the original prompt (a famous quote or phrase).
Like Wordle, guesses were color-coded as feedback (green for correct, yellow for partially correct, gray for wrong).
We called it “DALL·E-dle” (DALL·E + Wordle).
The project was playful, but it hit a nerve. PC Gamer, a major gaming publication with millions of followers, wrote about it under the headline: “Play a weird wordle-like where you guess what prompt an image generator used.”
For us, this was more than press; it was proof. A scrappy side project could spark attention, land us in mainstream media, and validate our instincts about AI’s viral potential.
3. Launching YouTube Summary with ChatGPT
Armed with these insights and experiments, we were watching OpenAI closely. When ChatGPT launched in November 2022, we recognized the opportunity instantly.
Within a day, we released the YouTube Summary with ChatGPT. It quickly went viral on social media, was picked up by global media outlets, and brought millions of users to Glasp.
The speed mattered. By leaning on prior conviction, side-project experiments, and our founder's mindset of “ship fast and iterate,” we were able to move before anyone else.
4. Lessons Learned
Develop the instinct for waves:
The biggest inflection points come from recognizing technology shifts before they become obvious. Watch the signals, follow the communities, and trust your intuition.Side projects can be big wins:
Small, playful experiments can earn you outsized attention and credibility. “DALL·E-dle” may not have been a core product, but it helped us build awareness and confidence to ride the next wave.Execution speed is everything:
Spotting a trend is one thing; acting on it immediately is what sets you apart. If we had waited even a week after ChatGPT’s launch, others would have beaten us to market.Make it fun, not just functional:
The projects that spread are often the ones that surprise and delight. Whether it’s a quirky AI game or an elegant YouTube tool, fun is a growth accelerant.
Summary
In Hatching Growth #3, we explored how Glasp anticipated the AI boom, experimented with side projects like DALL·E-dle, and leveraged those insights to launch YouTube Summary with ChatGPT, our breakout hit.
The key takeaways:
Watch for early signals of technological shifts.
Use side projects as low-cost bets to test waves.
Execute quickly when opportunity arrives.
In the next installment, we’ll share another AI project we built before YouTube Summary, the lessons it taught us about turning side experiments into growth levers, and also walk through the ideation process that led us to create YouTube Summary itself.
Partner with Glasp
We currently offer newsletter sponsorships. If you have a product, event, or service you’d like to share with our community of learning enthusiasts, sponsor an edition of our newsletter to reach engaged readers.
We value your feedback
We’d love to hear your thoughts and invite you to our short survey.