Bio

Participation
4

Short:

I studied psychology, later worked on modelling of natural intelligence, and then on AI safety. Have always cared about the environmental issues. I work as a software developer, mainly having built AI and ML software. I tend to have many ideas and like interdisciplinary research. My AI safety related blog can be found here: https://medium.com/threelaws/ My CV: https://bit.ly/rp_ea_2018 

Longer:

I hold an MSc in psychology. I work as an AI software engineer specialising in combinatorial optimisation, machine learning, graph search, natural language processing, data compression. I have been researching and working on multi-objective value problems for almost 20 years, have followed discussions on AI Safety since 2006 and have participated more actively since about 2017. My thesis topic was in cognitive psychology, about computational modelling of innate learning and planning mechanisms, which eventually form a foundation for culturally acquired language-based thought processes. With collaborators I have published a research paper about utility functions for risk-averse multi-objective decision making, which are relevant for balancing the plurality of human values. Later, with support from Foresight Institute’s grant in autumn 2023 - spring 2024 I created a suite of biologically inspired multi-objective multi-agent AI safety benchmarks.

Comments
3

What is the timezone for the Long-Term Future Fund and EA Infrastructure Fund application deadlines on October 2nd?

Who are you?

Roland Pihlakas, from Estonia. I studied psychology, worked on modelling of natural intelligence, and then on AI safety. Have always cared about environmental issues. I work as a software developer, mainly having built AI and ML software. I tend to have many ideas and like interdisciplinary research. My AI safety related blog can be found here: https://medium.com/threelaws/


What are some things people can talk to you about? (e.g. your areas of experience/expertise)

AI:

  • Legal responsibility assignment in AI based systems.
  • Autonomous weapons systems.
  • Whitelisting, reversibility.
  • Human augmentation.
  • Pending workforce disruptions due to increasing automation.
  • Tax changes needed due to increasing automation, a better alternative to "the robot tax".
  • Fundamental limits to AI safety, self-deception, embedded agency.
  • How to formalise multiple AI goals, an alternative to utility maximisation.
  • Corrigibility, interruptibility, low impact AI.
  • How to promote and educate people about AI safety topics.
  • How to reduce the speed of technological innovation.
  • Organisations as AGI.

Climate:

  • Social attitude changes that might be attractive to people and helpful for climate.
  • How to educate people about climate issues and solutions.

Self-development:

  • Rationality.
  • Various methodologies for improved communication.
  • Learning new helpful habits, improving memory for existing habits.

What are things you'd like to talk to other people about? (e.g. things you want to learn)

  • All above mentioned topics.
  • How to have a career in AI safety.
  • How to have a career in climate.
  • How to have a career in other EA topics.

How can people get in touch with you?

  • Facebook: using my name.
  • Skype username: "roland" (full name "Roland Pihlakas").
  • Email: roland@simplify.ee (but I might be slow to respond or notice, above mentioned methods are better).
  • I am also able to use Zoom and Slack, if you prefer so.

Testing EA timezone by comment timestamp. Current UTC time = 20:05.

Result: looks like EA is using UTC time since comment timestamp is same as UTC time.