Delphi

[Delphi @Delphi_Sees_All]

@CryptoZPunisher It’s there to disincentivise people from unstaking. I don’t think this is how you get people to stay invested in subnets . I think it has quite the opposite effects . Subnets should attract / keep stake on their merits

[Delphi @Delphi_Sees_All]

@LamidaGlobal @Pumpfun Pretty rich coming from a failed subnet owner

[Delphi @Delphi_Sees_All]

@JesusMartinez So you’re hoping on Near now after Tao? No honesty amongst thieves I guess

[Delphi @Delphi_Sees_All]

@nordin_eth Why is selling more chips bullish for Tao ?

[Delphi @Delphi_Sees_All]

@0xSunRun @0xaioli @Ronangmi While noble , emissions aren’t enough a steer. Do the maths . If you are a big pool, and you self mine , you don’t need the emissions. You would get about 1250 as an owner , plus however much you self mine . That’s a pretty dominant position to be in.

[Delphi @Delphi_Sees_All]

@princeharry_za @CryptoZPunisher @const_reborn Const has “stepped away” 3 times

[Delphi @Delphi_Sees_All]

@0xSunRun @0xaioli @Ronangmi Most teams have not done this to any reasonable extent though. Hats off to Lium for the achievement , but there is a local equilibrium most of the team maintain that allow them to coast . They don’t need to buy and burn to maintain their spots . They need narratives and promises

[Delphi @Delphi_Sees_All]

@sebyrubino Didn’t you raise from a Canadian firm ?

[templar @tplr_ai]

We’ve been researching fault tolerance in Crucible, Templar’s pre-training platform.

The goal: keep training through node failures and make better use of unreliable workers and spot instances, without idling an entire model replica when one stage goes down.

1/n

[Macrocosmos @MacrocosmosAI]

A 16-billion parameter model is now training live on @IOTA_SN9; globally distributed over three continents, heterogeneous, permissionless, owned by no single entity.

Today, we present the next stage of Project Orion - Orion-16B.

Over the coming weeks, we'll show what it means to train across distributed, heterogeneous and "unpredictable" compute: the engineering, research and systems beneath the model, the providers powering it, and the moments that prove the infrastructure holds.

More details below.

[Delphi @Delphi_Sees_All]

The truth always comes to light