Just caught up with Vitalik Buterin's latest post about something that's been bothering him for a while now. Turns out that whole Shiba Inu windfall from 2021 - you know, when the SHIB creators randomly dumped a ton of tokens into his wallet trying to ride his name for clout - ended up creating roughly $1 billion in donation capacity. Wild origin story, right? Dog coin, a Canadian closet, and some 78-digit numbers he had to get his stepmother to read out over the phone.

Here's where it gets interesting though. Vitalik split that massive haul between a few initiatives. GiveWell got $50 million upfront. CryptoRelief took one half and used it for medical infrastructure in India plus some research work. The other half? That went to the Future of Life Institute, an org focused on existential risks from AI, biotech, and nuclear weapons.

But FLI apparently did something Vitalik didn't expect. They managed to liquidate around $500 million from their SHIB allocation - way more than he thought possible given how thin the liquidity was. And then they pivoted their entire strategy.

This is the part that's got Vitalik concerned. FLI basically shifted focus toward aggressive political and cultural campaigning on AI policy instead of the broader existential risk roadmap he initially backed them for. His worry is pretty straightforward: when you've got hundreds of millions in a coordinated political war chest, things can easily go sideways. He's flagging that regulation-first strategies tend to centralize power, create backlashes, and end up authoritarian even when that wasn't the original intention.

He gave some specific examples too. FLI's biosafety approach of embedding guardrails into AI models? Vitalik called it fragile because jailbreaks and fine-tuning make those restrictions easy to bypass. The logical endpoint of that thinking is basically 'ban open-source AI' or 'let one good-guy company dominate globally' - and approaches like that backfire hard, he says.

There's also a structural issue he highlighted: when governments restrict dangerous tech, national security orgs always get exemptions, and they're often part of the problem themselves anyway.

That said, Vitalik did give FLI some credit for recent work like the cross-ideological pro-human AI declaration that apparently brought together conservatives, progressives, and libertarians. And they're researching ways to prevent power concentration from AI, which he sees as valuable.

But the core tension is clear. A donation he never planned to make, from tokens he never wanted to hold, ended up funding an organization that went in a direction he's uncomfortable with. He shared his concerns with FLI privately several times before going public about it.

In other market moves, ICP, NEAR and UNI were among the bigger movers today, with altcoins outperforming bitcoin overall. Coinbase bounced back 10% from session lows after yesterday's earnings miss. And SEC Chair Paul Atkins signaled openness to rules around onchain trading and blockchain settlement infrastructure.