Raphael Mu
0 followers · 247 views
on the atlas — 5
- Existential Pessimism vs. Accelerationism: Why Tech Needs a Rational, Humanistic "Third Way"2 savers
- AI and International Relations — a Whole New Minefield to Navigate | United Nations University1 savers
- AI Could Defeat All Of Us Combined — LessWrong1 savers
- What failure looks like - LessWrong12 savers
- Mission and Vision - Astera Institute1 savers
highlights — 14
when democracy becomes more thoroughly entrenched in the social state, men stop listening to a favored class of people, but instead are slavish to impersonal forces such as theories
Existential Pessimism vs. Accelerationism: Why Tech Needs a Rational, Humanistic "Third Way"there are three broad types of AI — prediction machines, clustering machines and generative machines
AI and International Relations — a Whole New Minefield to Navigate | United Nations Universityinfluence-seeking systems mostly acquire influence by making themselves useful and looking as innocuous as possible
What failure looks like - LessWrongAvoiding end-to-end optimization may help prevent the emergence of influence-seeking behaviors (by improving human understanding of and hence control over the kind of reasoning that emerges).
What failure looks like - LessWrongIf ML systems are more sophisticated than humans, immune systems must themselves be automated. And if ML plays a large role in that automation, then the immune system is subject to the same pressure towards influence-seeking.
What failure looks like - LessWrongIf you try to allocate more influence to systems that seem nice and straightforward, you just ensure that “seem nice and straightforward” is the best strategy for seeking influence.
What failure looks like - LessWrongany influence-seeking policies we stumble across would also score well according to our training objective, because performing well on the training objective is a good strategy for obtaining influence
What failure looks like - LessWrongHuman reasoning gradually stops being able to compete with sophisticated, systematized manipulation and deception which is continuously improving by trial and error; human control over levers of power gradually becomes less and less effective; we ultimately lose any real ability to influence our society’s trajectory.
What failure looks like - LessWrongfast takeoff can be a key risk factor, but I’m scared even if we have several years
What failure looks like - LessWrongopen-source hardware, software, and wetware
Mission and Vision - Astera Institutegeneration, collection, or collation of open datasets
Mission and Vision - Astera Institutepublished literature, and datasets derived from it, can paint a skewed picture of reality by overemphasizing certain findings and underrepresenting others
Mission and Vision - Astera Institutedesigning studies where negative results update our understanding as much as positive ones
Mission and Vision - Astera Instituteshift from narrative-centric to data-centric
Mission and Vision - Astera Institute