Co-designing Gemini 2.5 Flash-Lite with the Trillium TPU was the key to unlocking the lightning speeds of 2.5 Flash-Lite. This Infinite Wiki is an incredible demonstration of what is possible with this model. We're so excited to see what you continue to build!

In the past month, Cursor found 1M+ bugs in human-written PRs. Over half were real logic issues that were fixed before merging.

zkIBC is what real interop should look like no trust assumptions, no fragile relayers. A project is making cross-chain messaging finally scalable and secure.

1/ Trend #1: Layer 2 Scaling Solutions Become Essential
Layer 2s like certain scaling solutions are making blockchain faster & cheaper.
Without them, DeFi can't scale to mass adoption.

Fee surge leaderboard for top EVM chains ( last 7d ):
A trading platform: +376%
A certain platform: +136%
A certain platform: +137%
A certain platform: +122%

4/ Solana, like Ethereum before it, did not enforce strong opinions about market microstructure. This by default favors takers, at the expense of makers

buildooors of katana
substreams are ready
sharpen those swords, more deployments incoming ⚔️

Why scan your AI code?
AI-generated code is inherently insecure due to outdated training data. Codacy's constantly updated vulnerability database helps prevent new issues.
Developers prefer creating secure code upfront rather than fixing

1/ FYI: The Scaled UI Amount Extension lets token issuers apply an updatable multiplier to the token's displayed amount. This is useful when tokens represent real-world assets like stocks, dividends, or yield-bearing instruments (e.g. stock splits, distributions).
This extension

This uses Sequence which aggregates them already. But also you will be able to use them soon when the new [redacted] feature launches.

if everyone used zAMM for fungible exchange and their tokenization lifecycle the ethereum L1 would scale today

does the hardware acceleration actually make zk usable for stuff outside defi or am I gonna see 0 retail use cases for another cycle like always

It runs on 300000 nodes on 5 continenets and on Starlink satelites. You would have to cut off the power of the sun as 100000 nodes run off grid on their own solar power. Try again.

Qwen3-Coder-480B-A35B-Instruct launches and it 'might be the best coding model yet'