p(doom)

You really think we're going to scale data labelers to AGI? Today, we release the largest public long-horizon dataset of human digital work. 600h of long-horizon AGI research across 3 months. 🧵(1/n)

TRY US

We’re p(doom), an AGI research lab. We’ll pay you $300/month to record your screen while working. If your work is open-source and involves research, engineering, design, editing, or similar long-horizon digital work, fill out the form: forms.gle/Um8pzes644tL66…

Since @_rockt's banger position paper is going viral, here is a a set of notes we (@maharajamihir) made in 2024 while brainstorming paths towards AGI in front of an empty whiteboard: docs.google.com/document/d/1Rb…

If you openly publish your work, consider participating in crowd-sourcing by installing crowd-cast. (n/n) 🖥️ crowd-cast: github.com/p-doom/crowd-c… 🌐 Blog: pdoom.org/crowd_cast.html 📈 Dashboard: pdoom.org/crowd_cast_das… 🤗 Dataset: huggingface.co/datasets/p-doo…

As with crowd-code and AGI-CAST, we will publish the cleaned dataset under the most permissive Creative Commons license (CC0). (8/n)

You really think we're going to scale data labelers to AGI? Today, we release the largest public long-horizon dataset of human digital work. 600h of long-horizon AGI research across 3 months. 🧵(1/n)

🖥️ Repo: github.com/p-doom/crowd-c… 📷 Blog: pdoom.org/crowd_code_2.h… 🤗Datasets: huggingface.co/datasets/p-doo…

General intelligence cannot rely on human data curation. It must expand its own training distribution, just like humans do. Crucially, learning how to do this is itself behavior-clonable.

We introduce crowd-code 2.0: An open system for crowd-sourcing long-horizon human software engineering traces as action-observation rollouts Today’s models have short horizons not because of architectural bottlenecks, but because we don’t train them on months-long human behavior