Your model can code, so what?
In a new era of harness engineering where cheap models can code "like" humans, what is stopping Artificial General Intelligence?
Error-AccumulationAgenticone-shot
Read
Centre for AGI · Independent research
Tracking the rise of artificial general intelligence — and whether today’s methods can get us there.
We measure progress toward AGI and superintelligence with transparent, citable evidence: capability trends, benchmark saturation, and the open gaps between the systems we have and the systems we’re trying to build.
Our mandate
One question, rigorously pursued: are we on a path to AGI?
An independent research institute measuring progress toward artificial general intelligence and superintelligence — and assessing whether today’s methods can get us there.
Research
All articles
Featured analysis
Reading list
All papers
Papers worth your time
Paper · 2020
Scaling Laws for Neural Language Models
The paper that turned capability into an engineering forecast: loss falls as a smooth power law in compute, data, and parameters. The empirical backbone of the "scaling is enough" thesis we continually stress-test.
ScalingFoundational arxiv.org
Paper · 2022
Training Compute-Optimal Large Language Models
The "Chinchilla" result recalibrated the scaling recipe: most large models were badly undertrained on data. It reset everyone's compute-to-data ratios and remains essential for any honest forecast of where the curves go next.
ScalingCompute arxiv.org
Paper · 2023
Sparks of Artificial General Intelligence: Early Experiments with GPT-4
A sprawling, contested case that GPT-4 shows early signs of generality. Read it alongside its critics: it is as much a document about how hard AGI is to evaluate as it is a claim about the model.
AGIEvaluation arxiv.org