Paras Chopra: Program synthesis via naive search and genetic algorithms, with/without leveraging abstractions

Adithya: Personal AI server (cloud-in-a-box)

Me: Implementing Eric Hehner's unified algebra

Abhinav Tushar: Dictation mode for Emacs , including support for blending two streams (voice and text) at the same time (more thoughts)

Ruturaj Mohite: Circuitverse: Web-based Digital Logic Circuit Simulator

Nemo: Investigating if Arazzo specification is Turing Complete

Ishan Bhanuka: Semantic Code Search using Data flows

Advait: Remote-compute for video editors

Tagging @unnati patel@Prabhanshu Gupta @Atharva Raykar @KS Sreeram about their demos.

Shubhojit Basu: Agent-based swarm simulation with Love2D and Lua (and its use for EV battery swapping network optimization)

1. the flow is a simple one: an entry screen where you pick a colour, icon, a few tags for yourself / the "room" where everyone hangs out / and an overlay to show you anyone's avatar/name/tags 2. this is the nodegraph for the "room" view: here you'll see the api calls to the service returning your profile and everyone else who's joined, some decision logic to show a different profile card, and navigation actions, etc. a. everything is "data" so it can flow from node to node (live of course :-)) b. yellow lines are data, blue lines are cards, pink lines are action messages c. there are lots of node primitives, like object, filter, lookup, join, database queries, card, component, decision, state machine, various actions like save, navigate, API calls, etc. etc. etc. 3. the "cloud agent" is a simple one, no params even in this case, it just queries the data store it's sitting in and returns all the records to the caller (which is in the view in #2 above)