The teams that move fastest with AI in regulated environments don’t try to bypass compliance.
They bring them in early so governance shapes what gets built — not slows it down.
What’s been your biggest surprise when implementing AI in complex or regulated settings?
Two years ago, we tried to reduce institutional onboarding at Coinbase from 70+ questions to under 10.
We succeeded. Time-to-activation went from 15+ days to under 2.
But the hardest part wasn’t building the AI. 🧵
Compliance didn’t care about 97% accuracy. They wanted to know what happens in the 3% — and how we’d catch it.
That forced us to change how we built things. Governance stopped being a final checkpoint and became part of the design process from day one.
Coinbase's evolution from the bottom of the bear to now will be studied for generations.
Competing at nearly every layer of the crypto stack...and winning.
I just contributed to @_cubik Grants Round 1
Join me in supporting solana projects.
Let's together shape the future of public goods and open source on solana https://t.co/snPaPvGf1G
Anyone else constantly oscillate between complete conviction that they can and should make a dent in the world, but also the realization that we are but a speck of dust in the universe
We're delighted that Imperial has risen one place to sixth in the @worlduniranking for 2023, putting it third in the UK and top in London.🏆
Congrats to @OurImperial community! 🎉
Read more about the results: https://t.co/sLGPkY82wB
📷: Dr Pradeep Luther
#QSWorldRankings