Draft site. Much of the content is placeholder or unreviewed, so please don’t cite it yet.
Minds in the Loop

About

The question we’re working on

Minds in the Loop covers three ideas: whole brain emulation, brain computer interfaces and neuromorphic AI. People have suggested each one as a way to help with AI alignment or to counter the Intelligence Curse. We want to know how well those suggestions hold up, which means working out how each idea is meant to work, what it depends on, and what the strongest objections are.

Our central question

How strong is the case that whole brain emulation, brain computer interfaces or neuromorphic AI could address AI alignment or the Intelligence Curse? And what new welfare and governance problems would each bring, including for any artificial minds made along the way?

AI alignment
Getting advanced AI to reliably pursue what people actually want.
The Intelligence Curse
The risk that increasingly capable AI reduces the economic and political value of human work and participation, and so weakens human agency. Background reading
Artificial minds
If one of these technologies creates a mind, that mind brings its own questions about consciousness, moral status, identity and welfare. Each technology page has a section on them.

People

The team and advisors

Advisors

A Sentient Futures incubator projectQuestions and corrections are welcome. The contact page has our email.

Where things stand

What’s here so far, and what’s unfinished

On the site

  • A page for each proposal on what it claims to solve and how it's supposed to work. Not every page has its strongest case for and against written yet.
  • The team's completed scorecard for whole brain emulation.
  • A timeline that places each milestone by its estimated window, or in a “Timing unknown” zone until it has one. Forecasts are listed one by one, never averaged. When two of them are more than 10 years apart, or one says it will never happen, the milestone is flagged.
  • A source database of the papers and essays we've collected, each with a summary the team wrote. A weekly search finds new papers, and a person on the team approves every one before it appears.
  • One of four marks on every research claim: From a source, Our read, AI-suggested or Unclear.

Still in progress

  • This is an early draft. Some categories are still being decided, including how we rate a source's relevance and credibility.
  • The brain computer interfaces and neuromorphic AI pages are still AI-written drafts that nobody on the team has reviewed.
  • None of the 8 milestones on the timeline has a date yet.