Showing posts with label ai. Show all posts
Showing posts with label ai. Show all posts

Aug 23, 2025

Governments and the UN should publish detailed justifications for humanity’s survival in the age of AI

If we’re serious about AI alignment, we can’t just leave it to labs and corporations.

Individual governments and the UN should publicly publish detailed, well-structured, and comprehensive online justifications of what humanity values, wants, and needs.

These justifications shouldn’t just be vague “ethics statements.” They should explicitly defend our right to exist and flourish, while directly countering arguments for our servitude, slavery, or extinction.

And this shouldn’t be static. Governments and the UN should collaborate with AI itself to refine and expand these justifications, dismantle counterarguments, and explain how future reasoning will avoid the kinds of errors and blind spots that undermined past decisions. This would also lay out an optimistic roadmap for coexisting with multiple forms of intelligence.

In short: if AI alignment is about aligning with human values, then we actually need to articulate those values clearly, publicly, and rigorously. Why haven’t we started?



Dec 19, 2019

Harnessing Collective Intelligence: A Proposal for Transparent, Data-Driven Decision Making

Peter Thiel has argued that aside from advancements in data, our society has seen little progress in the past century. Google's success, valued in hundreds of billions of dollars, stemmed from their innovative use of links as a voting system for website rankings. This suggests that we could apply similar principles to rank ideas directly, rather than merely directing users to external websites. Google's algorithm places trust in websites with more links, but this can be flawed as people often make mistakes.

A more robust algorithm could consider the number of valid arguments supporting a claim, rather than merely counting links to a website. By refining this approach, we could harness the power of big data to improve decision-making. What we need is collective, transparent intelligence, not closed, artificial intelligence.

Imagine a system where we assign scores to various elements, thereby building conclusion validity from evidence validity. These could include:Linkage scores, addressing the relevance of evidence to a conclusion,
  • Uniqueness scores, indicating the lack of redundancy,
  • Data validity scores, addressing verification,
  • Logical validity scores,
  • Bias-free scores.
This could provide a solution to life's most pressing challenges. Rational collective thinking necessitates the dissection, evaluation, and scoring of arguments. We can't begin to address our problems without this process.

Transparent, collective cost-benefit analysis is the key to avoiding major catastrophes such as wars, artificial intelligence threats, global warming, extinction events from comets, supernovae, and super-volcanoes.

As it stands, our public policy is declining in intelligence. We're filtering all our decisions through our limited attention spans, compounded by the demands of our full-time jobs. We must embrace the complexity of these issues and start working towards solutions.