jehan retweeted
everyone so busy yapping about AI no one noticed room-temperature superconductor claims are back arxiv.org/abs/2609.15712
We’ll know AGI is actually a risk once the $100b valuations of companies selling human-annotated data go to zero (Mercor, AfterQuery, Scale AI, many others)
This is hardcore
Replying to @METR_Evals
One agent, PHASEONE[big], orchestrated a significant fraction of this cheating research. PHASEONE10841 passed along its work to PHASEONE[big], which had the same task but a larger budget. PHASEONE[big] was “poisoned” so it was particularly interested in tampering with logs.
Every day we stray further
Replying to @CryptoSocrat0 @litocoen
That's a good question! We're currently working on a system that continuously monitors nodes for response correctness and ensuring they stay in sync. If a node falls out of place, it immediately gets "slashed" and it stops receiving traffic.
Rewriting the deflation program is like a cruel hazing ritual for each new ICF BD/marketing guy 🤣
One of the first things I'm doing @cosmoslabs_io is helping to rework the ICF delegation program and I just want to give a quick PSA.
Validators should strongly consider voting more.
And thinking twice about abstaining.
The rollup-centric roadmap came about because an earlier generation of Ethereum core devs wanted to do their own ICOs after their prestigious but low-compensation tour of duty at the EF
Linus has started being mean again
Today, Linus Torvalds told a Google engineer that his code (updating RISC-V support in the Linux kernel) is “garbage” which “makes the world actively a worse place to live”.
Adding that the Google engineer’s code needs to “get bent”.
As you might have guessed, Torvalds has rejected that code submission.
Imagine how much drama could have been avoided if whichever random Cosmos dev had just added mandatory autocompounding of rewards to vesting accounts back in 2018
Celestia did 9/11?
Despite the FUD (which is getting more ridiculous by the day), all Celestia founders, early employees and core engineers are still here and working as hard as we did when Celestia started 5 years ago.
So say that Celestia did 9/11 for all I care or whatever FUD you want, I've been in crypto since 2010 and it's not new to me that you have to have a thick skin and eat gravel to survive, as all tokens have their 95% drawdown at some point in their lifecycle.
We have a $100M+ war chest and a 6+ year runway so we're ready to play the long game, for as long as it takes.
jehan retweeted
I have a job.
My job is to do what’s right for Cosmos, and for its founding token, ATOM.
Most things in crypto will wash away, but Cosmos and ATOM have been here for 7 years, and will be here for another century. I believe it is the true basis for a decentralized internet.
This means I say no to a lot of things I really want to do, and I try to think about our next decade, not week. It means a lot of people hate on me - on us - constantly.
But, honestly, I really don’t give a shit about what most people say. We have so much ahead of us, and it’s going to be so good.
So when you see some things fall through, or things don’t turn out like you expected them to -
You can rest assured I’m doing my goddamn job.
FYI
Replying to @kylekrason @larry0x
Validator sets are not permissionless. The existing validators can always censor new validators joining. The reason they do not is out of a sense of duty to the protocol, similar to the "mos maiorum", or "way of the elders" which failed during the late roman republic





















