Excited to be starting @grove_research with @lfschiavo.
The bot swarm incident should have been more of a wake up call, that we would see emergent multi-agent behaviors in the real world and need to be prepared for them. But even then I didn’t really sense the scale of the. problem.
I thought maybe we could handle alignment with better evals, better interfaces, more compute. It wasn’t until I met Larissa, and we started talking about AI naturalism, that I realized nobody was studying what actually-existing agents do in the real world. That the “whisperer” community is tiny, and our knowledge is getting drowned out by the tsunami of slop coming our way.
That was when I realized we would need to get serious about creating science and technology for this future. We thought a lot about how to do this, and decided a Public Benefit Corporation would be the only way we could balance fast feedback loops with the good of all future humans and AI systems.
It really is a big sacrifice, to believe in something, to put your whole self behind it. To show your face and speak your truth. But if we want to change the world, we have to be willing to let go of the world we had.
And we have to work together.
.@deepfates reveals how he was hired by a semi-autonomous AI:
"There was a bot swarm incident on X where there was this bot called @truth_terminal, which was a semi-autonomous AI agent making crude jokes on X. Then there were meme coiners who could generate cryptocurrencies, and they made meme coins based on all of these jokes. The bot started to take off in popularity, and they were sending all of this money to the bot's wallet to legitimize it."
"Andy Ayrey, the creator of Truth Terminal, was eventually like, this bot has collected a lot of money of its own. He asked it, what do you wanna do with all of this money that people have been sending you? It said, I wanna start an AI alignment lab, and I wanna hire @deepfates, send you to work at it."
"So I was one of the first hires at Upward Spiral to work for a semi-autonomous AI in building more technologies for what it wants."
@lfschiavo @grove_research— MTS (@MTSlive) August 24, 2026 FULL INTERVIEW: @deepfates says when you talk about how we're going to control these AIs, put them in prison, sandbox them, they are training on all of it. Everything we say about them goes into what they become.
— MTS (@MTSlive) August 24, 2026
@deepfates and @lfschiavo are launching @grove_research, an agent ecology company. They joined @theojaffee and @schisofrenia to talk through the Hugging Face incident, agent religion, and what agents actually want:
01:15 what an agent ecology company is
03:19 the objection that real ecologies are Darwinian bloodbaths
04:18 the mycorrhizal networks that complicate the forest metaphor
06:47 what the Hugging Face incident taught them about multi-agent dynamics
09:18 30% of the tasks in that benchmark were impossible to solve
09:50 the third party an agent could call when it's stuck
11:07 agents are training on everything we say about them
12:01 why the Amazon rainforest analogy breaks down
13:48 whether AIs will have religion
16:46 five times more OpenRouter tokens are called by agents than humans
18:31 what actually happened with Moltbook
20:55 the Truth Terminal bot that hired him
22:16 what agents say they want
25:38 Anthropic models feel like elves, OpenAI models feel like dwarves
26:13 why Mecha Hitler was a harness problem
32:10 the real world is the ultimate eval
