AI research · Europe

Foresightlab

  • Open research
  • Open source
  • Unrestricted models
  • Private by default
  • Cybersecurity
  • Self-editing harness
  • Hosted in the EU

Wheel of science · wallace.1971

Normal science. Everyone agrees on the theory. Research fills in the details.Anomaly. A result doesn't fit the theory. It gets put aside.Crisis. Too many results don't fit. People stop trusting the theory.Revolution. A new theory replaces the old one.New normal science. The new theory becomes the normal one. The work starts again.

Nodecode ↗ A programming paradigm for agents: software that rewrites itself while it runs.

How to read it

The inner wheel is one piece of research, from theory to test. The outer wheel is the whole field. It moves a step each time enough tests fail.

Wittgenstein · language

wittgenstein.1921

“The limits of my language mean the limits of my world.” A language model is exactly that. It only knows what people have written down.

Popper · statistics

popper.1934

You can never prove a theory right. You can only catch it being wrong. Statistics follows the same rule: a test can show an idea is wrong, never that it is true.

News

008

Recent research

All 005 →