Researchers Unveil NeuralOS Generating Screen Frames From User Actions End-to-End

TBench Science / @DillmannSteven, @ryanmart3n, @alexgshaw, @Mike_A_Merril, @AlexGDimakis, @sanmikoyejo, @lschmidt3 (@Stanford) A benchmark for evaluating AI agents on real computational workflows across the natural sciences, with tasks authored and verified by scientific domain experts.

NeuralOS was selected for Laude Slingshots!

Short term, we're building the computer equivalent of a driving simulator, so agents can safely practice using computers and the internet.

Long term, we want every pixel of the UI generated on the fly in response to user actions, rather than assembled from predefined components.

Grateful to @LaudeInstitute and the team: @RivardLuke, Sun Sun, @NRC_uOttawa, @WenhuChen

Congrats to Research Partner @bradenjhancock and Laude Institute co-founder @ChrisRytting on assembling another exceptional batch. Every project in Slingshots // THREE ships open source. Full announcement: http://laude.org/updates/slingshots-three

Video World Models / @soniajoseph (@Mila_Quebec, @McGillU) World models for planning, scientific simulation, and verification across physical and industrial systems.

Rules to Code / @marzieh_nabi, Sareh Nabi, Mike Genesereth, @RolandVogl (@Stanford) A framework for teaching LLMs to codify human expertise into formal logic.

@ryanmart3n @alexgshaw @AlexGDimakis @sanmikoyejo @lschmidt3 @Stanford @StevenDillmann @Mike_A_Merrill