BREAKING: We just published the most comprehensive report on crypto cards in the industry.
Not because it’s a niche. But because it quietly became an $18B market.
In early 2023, crypto cards were doing ~$100M per month.
Today, they’re doing >$1.5B.
So we spent weeks digging into the data, the infrastructure, and the companies actually building this stack.
Here’s what we learned 👇
Yesterday we launched https://t.co/CR5RB8LamL into early access. Today we're bringing you a deep dive into LLM internals with beautiful visuals crafted by our very own @samwhoo. Discover exactly what gets cached to offer you 10x cheaper input tokens.
https://t.co/qyPqz80lhk
Made my first transaction on @base before the first Onchain Summer. Now, Base is two years in and I am building @GrowlySuite on Base.
Sometimes it’s wild to zoom out and see how far we’ve come.
Met @Nibel_eth at ETH Vietnam today - ngl, he’s BASED!
Bonus: Matching shirts, zero planning. 🫡
"AI is amplified intelligence, not artificial intelligence. The smarter you are, the smarter the AI is."
"AI is practically constrained, because it has to be prompted and verified, and because it does things middle-to-middle rather than end-to-end."
Two key insights I strongly agree with from @balajis in his latest article.
While building @GrowlySuite through ElizaOS and now the LangChain framework, I realized it’s not just about setting up an agent with a long system prompt and a bunch of tools.
Most of my time was spent writing clear descriptions for custom tools (pre-MCP era), and endlessly tweaking system prompts to ensure the agent invoked the correct tool(-chain) in varied test scenarios. It felt like 90% prompt iteration, 10% actual coding. At one point, I almost disliked the term “AI agent” — it felt like I was sidelining my coding skills.
But with time, it became clear: the more intentional I am about the desired output and the middle steps, the faster I move toward meaningful impact and broader capabilities with agentic intelligence.
❌ We were one of nine fantastic teams awarded $40K at Base Batches Demo Day 001.
✅ @balajis and @jessepollak recognize our product as crucial tool for the future Web3 (WE STILL CANNOT GET OVER THIS MOMENT).
Big thanks to our mentors:
@rachit@PowersOfTau
And to @HashedEM , @odisealabs for the continued support during #IncuBase.
Excited for what we're shipping next - the prophecy of New Day One 🟦
Our New Day One on @base was starting from joining #IncuBase first-ever cohort.
The program helped us transform from a hackathon project to a “wait—we can actually do this” moment.
Here are some of the key lessons we learned along the way 👇
Suite: An AI-powered customer success engine for onchain user insights.
The next billion onchain users will be anonymous. Suite helps Web3 protocols truly understand their users for seamless onboarding and lasting growth.
Proud to join #IncuBase 001, where we've shaped Suite through constant iteration and the guidance of our amazing mentors.
Grateful to @buildonbase, @HashedEM, and @base for the opportunity to be part of such a prestigious program. LFB! 👀💙
Suite: An AI-powered customer success engine for onchain user insights.
The next billion onchain users will be anonymous. Suite helps Web3 protocols truly understand their users for seamless onboarding and lasting growth.
Proud to join #IncuBase 001, where we've shaped Suite through constant iteration and the guidance of our amazing mentors.
Grateful to @buildonbase, @HashedEM, and @base for the opportunity to be part of such a prestigious program. LFB! 👀💙
It’s D-Day! 🚀
We’re going LIVE on @Base at 8 AM PT to pitch Suite, an AI-powered engine enabling onchain user insights
Proud to be part of the #IncuBase 001 Batch by @buildonBase & @HashedEM.
Join us and LFB! 💙