Roman LeventovinEngineering IdeasSociaLLM: proposal for a language model design for personalised apps, social science, and AI safety…I first proposed this model here, as a base model for a proposed app to improve global online discourse through personalised comment…4 min read·Dec 19, 2023----
Roman LeventovinEngineering IdeasProposal for improving the global online discourse through personalised comment ordering on all…This is a description of a project I’ve been thinking about making in the last few weeks. I’m looking for feedback and collaborators…7 min read·Dec 6, 2023----
Roman LeventovinEngineering IdeasAligning an H-JEPA agent via training on the outputs of an LLM-based “exemplary actor”This post includes the overview and the conclusion of the article that was posted on LessWrong.7 min read·Jun 1, 2023----
Roman LeventovinEngineering IdeasAn LLM-based “exemplary actor”Design and properties of a simple language model cognitive architecture14 min read·May 29, 2023----
Roman LeventovYou illustrate the mechanism of the Gaia network in the agriculture space and also argue that it…I think it makes sense. However, when you try to transfer this principle into the domain of AI, I think it becomes more challenging on two…3 min read·Apr 28, 2023----
Roman LeventovinEngineering IdeasFor alignment, we should simultaneously use multiple theories of cognition and valueThis post is a follow-up to “A multi-disciplinary view on AI safety research”. I elaborate on some arguments behind this view.6 min read·Apr 24, 2023----
Roman LeventovWill people be motivated to learn difficult disciplines and skills without economic incentive?In a recent interview with ABC news, Sam Altman emphasised multiple times that he sees individual AI tutoring as one of the greatest…6 min read·Mar 20, 2023----
Roman LeventovinEngineering IdeasMorphological intelligence, superhuman empathy, and ethical arbitrationMorphological intelligence (Levin, 2022) is the ability to solve problems in the morphological space, e.g., navigate to a target morphology…3 min read·Feb 13, 2023----
Roman LeventovinEngineering IdeasA multi-disciplinary view on AI safety researchThis post has been originally published on LessWrong. I didn’t copy the footnotes into Medium because there are too many of them. Please…29 min read·Feb 9, 2023----
Roman LeventovinEngineering IdeasCritique of some recent philosophy of LLMs’ mindsI structure this post as a critique of some recent papers on the philosophy of mind in application to LLMs, concretely, on whether we can…23 min read·Jan 20, 2023----