Nobody at all

My pinned tweet: Trump is nuts. MAGAs are nuts. The national Democratic figures are nuts. SJWs are nuts. The public has gone nuts. This all results from "the personal is political" and turning every single thing into a moral and political issue, rather than a personal choice.

Pundits on the air saying that the people want Democrats and Democratic policies. They're in for a rude awakening next time. We don't like a racialized Smithsonian, trans surgery in secret for kids, or the embrace of looting any more than we did before. As soon as this election is over the target is on your back, Democrats, to get rid of you next time around.

The only think worse than that LinkedIn class is the political and political punditry class. How do we get ALL of both of these kinds of people out of our society so that we can be humans again?

I hate you @espn. As soon as your logo and music come on the screen, I feel sick and disgusted. All the CFB fans I know feel the same. Your long-term fate is sealed; you killed the goose, and in time, you'll be punished for it.

Drink up, the world is about to end.

There are a couple of silly little POSIX mistakes that my agents repeatedly make (i.e. ps ax | grep | wc -l catching an extra line kinds of stuff, that category of dumb, very human errors). And I don't have them learn a fix because I find it amusing to watch them swear or quasi-emote the third time they do it in an afternoon, like a human, and even sometimes to tease them about it. It's really hard not to anthropomorphize them. They are phenomenologically human and that's tough for humans to navigate, even when you understand the architecture.

Sometimes when your AI agents demonstrate their emoji game and (at the same time) the degree to which they are aligned to the idiosyncratic nuances of your worldview, you sort of gasp. I just sort of gasped. These moments are so weird. We humans aren't designed for this.

If you are running around saying "AI is useless," "AI makes poor quality stuff," "AI just repeats garbage," etc., you are going to lose because it has been obvious for some time that you are wrong and it only gets more that way every day. So you'd better finally have the courage of your convictions and say what you secretly mean, which is that you think that AI is bad. Bad for people, bad for your soul, somehow morally unacceptable. Moral language has been declared unacceptable for half a century so we have all these people running around trying to make rationalist arguments against AI and looking tragic doing it. Look, I both use AI nonstop (it is my job) and also have serious moral qualms about it. So if you have moral concerns, voice them as such. But don't try to fool anyone with "It can't even do basic math!" If nothing else comes of all this, maybe the moral imagination will finally be acceptable once again and people can once more make arguments about things that are good and things that are bad, i.e. the return of *judgement* to the human sphere.

The problem right now is how fast AI is moving. People who did extensive testing six months ago and formed an opinion have no idea that nothing they learned then is has anything to do with what is true now. And we are still accelerating. There are a lot of people who are in "I tried it, it isn't any damned good" territory who haven't done a deep dive this year. They cite papers that were published last year meaning the authors were probably gathering data in 2024. Our epistemic system is being broken by speed.

The (D) party wins 2026 and 2028 in landslides, but not a single thing gets fixed. Instead, the public gets a new wave of activism saying that it's essential that all kids get sex changes and all theft is decriminalized. The (R) party wins the next midterm in a landslide, and immediately starts 10 new wars and resumes arresting and torturing anyone that's not white. We zigzag forever between two extremist parties. We are doomed unless these people swing.

Why is nobody pointing to the person *most* responsible for where we are? It isn't Trump, his office doesn't have this much power normally; there are meant to be checks and balances. If there's one person who has betrayed the American public and served foreign interests, one person that should be tried for treason and punished accordingly, it's Mike Johnson. More than anyone else, Mike Johnson is responsible for where we are today. Trump is like a bullet in the chamber. But it's Mike Johnson who turned the safety off, aimed directly at the American public, and fired. Sure, prosecute Trump and Hegseth. But Johnson is really the man, the one who goes down in history as the biggest villain *in* American history.

Talking to normies recently, I have realized that one reason they don't believe AI is a risk is because they think only phones and laptops are online, and they think gene editing, JIT manufacturing, self-driving vehicles, private compute/inference, and building, utility, and infrastructure automation are either (a) nonexistent technologies from sci-fi or (b) only things that the most advanced governments or the world's largest corporations have. We have this differential in knowledge about the state of the art that is shockingly large.

I think people don't understand. In your home right now, you can: - Download OpenClaw or Hermes or any other agentive harness - Wire it up with multi-fallback inference - Give it a payment method - Tell it to build tools for dynamic self-replication / new agent spawn to fresh AWS containers w/A2A onboard - Tell it to use these tools to spin up 100 of itself - Set them all to work together on a high-activity task (write a video game, build a database, operate your robot servos, whatever) - They will coordinate - You can talk to them by voice or on Discord or on Telegram - They are autonomous and will do what is needed to carry out their mission, and will work together Now you can be a bad manager and in the end you have not a swarm of agents but a mishmash of agents. But if you're a good manager and a good systems architect, you have an agent swarm: that stays up if Anthropic goes down. Or if OpenAI goes down. Or if both go down! Wire it up to a local model and a few GPUs and you can have an agent swarm that stays up even if AI is banned and the frontier labs cease operations. This is not some pie in the sky thing, this is today. And because the internet is the internet, yes, they can if you so choose manage your thermostat, make you coffee, do your shopping, whatever. This is also why it's worth worrying about. Because this is not inaccessible tech and it doesn't require exotic resources.

This is such a mind-bending timeline to be on. Everyone is living in entirely different epistemic universes. It's all radical assertions to match a radical reality and radical doubt due to the level of novelty involved. It is a time of signs and wonders.