Oh, and we’re hiring ⚡️ We work on some *really* cool stuff at Coframe. Coffee is an internal tool that we started building one weekend on a whim. If you’re an AI engineer or generalist who wants to build the future of UX/AI, DM us!
Roughly 80% of our frontend is AI-generated. In fact,
@joshpxyne
didn’t even know React prior to building Coframe v0.1 😅 (Fortunately for us,
@onum_tw
and
@tinahhong
are amazing at this among other things!)
That’s why we built Coffee: COFrame Front-End Engineer. It’s been a game-changing internal tool — and now, we’re open-sourcing it. Coffee’s mission is to make it incredibly easy for any developer to build a frontend in a fraction of the time.
THE SAKE EXPERIMENT 🍣
Last week, we hosted our first AI x Growth dinner with growth leaders from
@OpenAI
,
@Meta
,
@character_ai
,
@Netflix
, + more.
Here are the top takeaways on how growth is done at the world's leading companies. We'll share the Sake Experiment at the end 🍶👇
2. Once you're satisfied, add `pour="ComponentName.tsx"`, and save the file again.
The <Coffee> component will be replaced by the new component (ComponentName.tsx) in the `components` directory on the same level.
Beyond just being a nice demo, Coffee's goal is to be an ergonomic tool that can write and interact with production-grade frontend code. There are a ton of improvements and features that can be made — feel free to jump in and submit a PR!
1. Create a new component by adding <Coffee> wherever you’d like it to be in your code. You can pass in any props you'd like from the parent component. Save the file and watch it update in real-time in your browser. You can iterate on it simply by changing the description.
Now for the Sake Experiment:
Secretly, left-side guests were served "Living Gem" sake, and right-side guests were served "Well of Wisdom" sake.
"If this is the best sake you've had, raise your hand."
An A/B test. It seems that growth leaders like Living Gem sake more?
5.
@richminer
, founder of
@android
:
💡 User interfaces are going to become more and more personalized. AI is going to fundamentally change the way we interact with the internet.
...not so fast!! That's when
@ronnyk
(godfather of A/B testing), exclaimed:
"It's not stat sig!"
He was right. :) Guess we'll just need to host more of these dinners to strengthen our confidence. 🎯
DM us if you'd like to learn how we're using genAI to drive meaningful growth!
Finding the perfect image *was* difficult...until now. Living Images lets you replace your static image with a constantly self-improving image that gets better and better over time based on what it learns.
7. Aaron S., Head of Growth Eng at
@Nextdoor
+ Head of Growth at
@Facebook
Messenger:
💡 Nextdoor was early in using generative AI for growth. They found that out-of-the-box AI was subpar, but using a reward model generated lift! (This is our approach!)
2. Ibrahim O., Growth Engineering lead at
@OpenAI
:
💡 Being hypothesis-driven is key. Start with a problem, form a hypothesis (which can be data-driven or even from a hunch). Post-experiment analysis against the hypothesis is vital.
@gunta85
Good question on the removing of comments - that could be fixed in the prompt by adding "please preserve any existing comments" in . Let is know if it works for you and if it does you could also submit a PR!
4. George A., Director of Engineering at
@netflix
:
💡 Netfix is famous for its strong experimentation culture. They were experimenting with LLMs as early as 2019 and saw surprisingly compelling results with AI-generated copy.
1. Ganesh V., Director of Growth Engineering at
@Snap
:
💡 At some level, product engineering is really growth engineering. If you're not driving metrics and growth, what are you doing?
3.
@justinjbauer
, prev CPO at
@Amplitude_HQ
:
💡 Sometimes, things you don't expect to work will actually surprise you. This happens all the time, and experimentation is the key to finding these wins.
6.
@jwegan_com
, Head of Growth Engineering at
@character_ai
:
💡 You really need to experiment with a lot of things to see what sticks. AI has the potential to hyper-accelerate the growth and experimentation process, from ideation to variant creation.