Two months ago, 5-minute crypto markets didn't exist on Polymarket.
Today, short-duration crypto markets drive over $500M in weekly transaction volume, over 20% of the platform's total.
That category wasn't feasible before.
5-minute and 15-minute markets across 7 crypto assets mean thousands of new markets per day, each resolving onchain the moment it expires. That's not a manual job.
Bots drive 50%+ of the volume in these markets, viable only because settlement is near-instant.
Goldsky Compose automates these markets.
Compose is the framework for writing durable onchain operations in TypeScript. Wire a function to a trigger (cron, onchain event, HTTP call) and Compose handles the rest: wallets, gas, retries, state, execution traces across 100+ chains.
The same pattern repeats across onchain finance. Banks shipping stablecoin rails. Tokenized funds pricing 24/7. AI agents paying for API calls. In each case, the onchain part is 20% of the system. The other 80% is a backend that happens to have an asset attached.
Compose is for that backend.
Live today: https://t.co/dU1GJjltAU
Every successful onchain app is built on vital offchain components that teams spend months to get right.
@Polymarket just launched instant 15 minute markets, built in just weeks.
It's powered by Compose: the framework handling painful onchain <> offchain interaction👇
A protocol I'm super excited to see grow is @liminalmoney; the delta-neutral yield generation platform built on Hyperliquid (and soon #HyperEVM).
'Earn sustainable, market neutral yield' on your stables.
Here's a short primer 👇🧵
🧵👇
You simply need to be farming the HyperEVM ecosystem right now.
Project points and a potential secret season 2 $HYPE airdrop are on their way!
Here’s my strategy to stack points across @kinetiq_xyz, @hyperlendx and @prjx_hl
Integrate all Tenderly-supported EVM chains into @ERPC with just one line!
✨ One URL to rule all your EVM chains
✨ Reliable RPC failover & optimal routing (hedge, retry, data integrity, etc.)
✨ Easy RPC performance/usage monitoring
Learn more here: https://t.co/6ys2KmrRSU
I just launched https://t.co/CIrqNtUMm6 on Product Hunt, it’s an AI-powered mock interview tool that helps devs practice real coding interviews (like LeetCode, but with voice + detailed evaluation).
It's live today, and every bit of support helps!
https://t.co/cCR3BvX7r1
«CNN for predicting and hiding personal traits from face images» by @rasbt presenting CORAL «Consistent rank logits for ordinary regression» and more below at #DLISS19
most common neural net mistakes: 1) you didn't try to overfit a single batch first. 2) you forgot to toggle train/eval mode for the net. 3) you forgot to .zero_grad() (in pytorch) before .backward(). 4) you passed softmaxed outputs to a loss that expects raw logits. ; others? :)
Practical Deep Learning for Coders, 2019 edition, is now available. With a shiny new video player with searchable transcripts.
This course is 100% new material, including some new techniques and results never previously published.
https://t.co/upLaVgfwSg