Introducing the Giga Roadmap.
Upgrading to Giga is an immense engineering challenge. The hardest technical problem since Ethereum’s Merge.
It will not be a single upgrade.
The Giga Roadmap is a public resource where you can follow progress towards Sei Giga.
@Seanfrank reminds me of a classic twitter hack: if you want to get engagement, purposely say something wrong - it'll become an arms race to correct you, resulting in your core idea getting a lot more attention
@bhorowitz same could be said for the internet, with nearly a trillion being spent on cables during the dotcom era
the real Q is whether AI infrastructure build is going to be similar to dark fiber
Sei Labs has built an in house AI code reviewer for the primary sei-chain repo.
The median wait time for a first code review decreased from 54 minutes to 6. The p75 review time decreased from 8 hours to 12 minutes.
Code changes now merge 25% faster than before.
The ultimate goal is to build seidroid into a platform where agents investigate incidents, respond to queries, and execute high-level specs in AI sessions, which will be instantly sharable across the engineering team's oncall rotation.
Getting institutions onchain is about building confidence at an organizational level, not just blockchain performance.
Happy to announce a joint report from @SeiNetwork and @Mastercard on how financial institutions decide whether to use a blockchain.
https://t.co/HXGGltqQDC
@joeljohn@0xNairolf this is a disingenuous take
most teams working on L2s are well intentioned, but similar to basically every other project in the space, have significant friction finding pmf
@degennQuant having a native app that competes with the ecosystem is always a double edged sword, and it always results in ecosystem projects losing
core contributors and governance will always prioritize the core chain
@arian_ghashghai the criticism is not that they're optimists, rather that they're opportunistic and not ideologically aligned
but tbh nothing wrong with that, its just the market at play
@Austin_Federa on the flip side, constructing new subway lines (and stations) costs billions of dollars and decades, whereas buses cost millions and take significantly less time
@alessandrod@SeiNetwork >it’s trivial to replay a 4kb block in 600ms.
600ms is time for data propagation, consensus, and execution, not purely execution
When a chain mentions blocktime they usually mean average blocktime (or p50 to be technical).
The speed of the slowest 1% of blocks (the p99) matters more.
As of v6.6, @SeiNetwork has a p50 of around 400ms, and an extremely fast p99 of roughly 600ms.
Scaling an L1 means scaling consensus, execution and storage.
The first part of Giga going live is storage.
Every transaction that runs has to be stored. If the database can't keep pace, the speed of the other components is irrelevant.
Eidos is scaling @SeiNetwork's storage.