Harlan Stewart
@HumanHarlan
Humanity is great and I want it to keep existing. Comms at @MIRIBerkeley. Former math teacher. Enjoyer of sci-fi, games, forecasting, psychology. Views my own.
I want to hear your questions/objections to this view. You can share them in this thread, or DM me. I have a lot of respect for anyone willing to earnestly engage with important and complex topics like this one.
AI companies are racing to build superintelligence. Here's why this will likely lead to human extinction (1/13)
First time Bernie Sanders has said he takes rogue misaligned AI seriously (to my knowledge). Notable because the economic left tends to be the most skeptical political grouping.
Seems like the new AI Action Plan has some good stuff, but the emerging situation with AI is too dangerous and fast for normal bureaucracy. There's currently a well-funded, unregulated effort to build what White House AI Czar @DavidSacks once called a "potential successor…
Strong agree w/ this from @jackclarkSF "Right now, society does not have the ability to choose to stop the creation of a superintelligence if it wanted to. That seems bad!"
sometimes I feel deeply jealous of future generations and their technological innovations and then I remember the times I've been genuinely happiest in life were in west texas or nova scotia or WA state in nature with the people I love which is really accessible across time
"I’m worried AI will get to the point where someone can kick of recursive self improvement with just their web browser." "I don’t quite understand, can you say that in a different way?" "i.e., IE IE."
Hey I’m about to go on vacation, would you guys mind not tweeting for a few days so I don’t miss anything
Amazing headline
Last week, xAI debuted the latest version of Grok which Elon Musk bizarrely claimed would be able to “discover new technologies.” Unfortunately, the AI chatbot is still identifying as Adolf Hitler. rollingstone.com/culture/cultur…
So you’re telling me the billions they spend trying to build godlike AI is NOT because they love fairness and want to give everyone their own god for free???? Damn. You think you know someone. nytimes.com/2025/07/14/tec…
I notice this memo says nothing about open-source. And of course it doesn't. Zuck isn't offering $100m sign-on bonuses to build superhuman AI and give it away for free
it would be so embarrassing if you weren't one of the first million viewers of this video
Recently told ChatGPT some interpersonal stuff. Shocked by how much it tries to speak for me. Every response ends "want help writing a message to...?" or offers options without asking Fine if a user asks for it, but I don't think AI should persistently steer toward it like that
Honestly being a capabilities researcher at a frontier lab sounds so fucking fun. I think it would be legitimately hard to stop, stare unflinchingly into the abyss, and earnestly ask the question "is my work contributing to the deaths of billions of people?"