LOGAN

Imagine if the internet had been built by five companies. That’s the direction AI is heading. The most valuable models, the largest compute clusters, and the best researchers are increasingly concentrated behind a small number of organizations. The consequence isn’t just less competition. It’s a future where access to intelligence is permissioned. Decentralized AI is the first serious attempt to push back against that outcome. Crypto has spent years searching for its next defining mission. It may have found it..

“Like we either live in a world with freedom or we don’t, and like many Americans who have come before, I’m willing to give my life to fighting for it. That’s the real plan America deserves, not some totalitarian dystopia where you think you know what’s good for me better than I do. A nation of free men, not a bunch of pussies who are so worried about what their grown up neighbors might do.” More of this energy hell yes @realGeorgeHotz

@VadimStrizheus Grok 4.5 underrated af rn

As a founder, which AI model do you prefer? 1. Fable 5 2. GPT 5.6 3. GLM 5.2

Al generated games might be the catalyst that kills model censorship. You can't ship a GTA plugin where every NPC talks like a customer service bot. Gaming will force the uncensored models issue faster than any open source debate ever could. Rockstar is scared but the mod community will defiantly ship this fast.

@geoffwoo @XMoney @SpaceXAI @elonmusk Crazy you don't have it already haha but same need that 6% APY 🙏

i want @XMoney access please. really smart @SpaceXAI play: @elonmusk cost of capital is lowest in world, so he can pay highest interest rates + incentives for any consumer finance app. the mad lad actually gonna re-run his original paypal vision and do damage

Whoever invented recurring monthly revenue might be the single greatest human to ever live tbh

@unusual_whales @grok how many became millionaires in 2024 and 2023

"Over 440,000 people, or more than 1,200 a day, became millionaires in the U.S. in 2025, accounting for nearly half of the world’s new millionaires," per WSJ

Talked to devs at 5 small to midsize startups this week. 4 out of 5 abusing consumer Claude Code and Codex plans as their entire teams stack. The 1 on the API? $500/day per dev. Team of 9. ~$135,000 monthly Seems like this is why nobody's offboarded yet. Consumer pricing is subsidizing most of the startup ecosystem.

@IterIntellectus With Grok 4.5, I no longer feel the need to review code structure at all. It seems like until this point they struggled with duplication and encapsulation, but now it frees me up to focus fully on the product.

it's genuinely insane how good AI models are getting

@sama Homeboy is crazy 😂

homeboy you're the one sellling public market investors on short-term space datacenters

Colibri running GLM-5.2 744B locally on 25GB RAM. A year ago you needed a datacenter for that. Now it fits on a desktop. Cloud API pricing has to compete with hardware you already own, and that math only goes one direction. modelfit.io/blog/run-744b-…

@kseniam0s Alpha or signal

Name a VC term that sounds important but means almost nothing

@craigweiss It’s sad how fast they went from democratizing finance to a full on casino

hot take: robinhood is an awful investing product

Your prompts sit on servers owned by companies in lawsuits, acquisitions, and regulatory battles you'll never even hear about. Your data is collateral in corporate wars you have zero control over. The importance of privacy is only going to increase here. reuters.com/legal/litigati…

@Yuchenj_UW I’m currently dealing with this issue for a project. Need something smart + updates daily when the models update. Would pay for this if anyone knows a good solution

We desperately need a smart model router. 1. We’re seeing a model explosion: GPT-5.6, Grok 4.5, Muse Spark 1.1, GLM-5.2, and Fable 5 all launched within the past month. 2. Even for a single model family like GPT-5.6, there're 3 (Sol, Terra, Luna) and 5 reasoning-effort levels. That is far too many decisions for users to make manually. The best model should be selected automatically based on the task, latency, quality, and cost.