Why are we using homogenous equality in our proof assistants? What's the advantage over heterogenous equality?
Anthropic should make an even bigger model size: magnum opus
thomasporter522.github.io/cellular-can...
deadlines
deadliness
Yeah... I was originally thinking about provably conservative extensions. Do you know of a general technique for proving an extension conservative? One that's formulaic enough to be algorithmically checked?
Soon our meditative acts will be done by Artificial Introspection
2. Initially, the rules governing the cellular automaton are unspecified, so the trace is undefined except for the random initial state. But you can paint cells in, specifying their states and forcing the instantiation of propagation rules. Watch the trace grow.
Logic thought: proving that proposition A is a consistent extension to a consistent theory amounts to proving that falsity does not follow from A. To prove this internally is to prove ¬¬A. So double negation elimination can be seen as the license to add (internally provably) consistent extensions.
3. In "interpolation mode" you can click and drag a cell continuously between states. The whole trace fuzzily interpolates accordingly.
1. At each point in time you're seeing a full spacetime trace of a cellular automaton. The horizontal axis is space, and the vertical axis is time (flowing down). The color of a cell represents its state.
