Grokipedia is a new contender in the encyclopedia space. The website represents an AI curated knowledge set that is currently accessed via a search input. Obviously I sought to see how it handled topics I hold near and dear: game programming.
The articles I browsed were as follows
I must declare that the Video Game Programming article is pretty comprehensive. I quite enjoy seeing topics like performance tuning & optimization, tools programming, multiplayer netcode, development workflow, fundamental algorithms and data structs among plenty of other topics. I came away from the article impressed with how extensively the topic is covered. I do have a desire to see more of the history explored and left some feedback via the highlighter tooling (highlight a block of text and you can leave feedback. I am surprised visual engineering by technical artists (think Acerola from youtube) isn’t discussed as that role offers some interesting insights. Maybe it will come in due time.
Next I wanted to see what is covered in the Behavior Tree article. I have read through chapters from the “Game Ai Pro” book series (4 books have been posted at the time of writing with the book 4 being an online exclusive). In Game Ai Pro Book 3 Chapter 9 it declares that a software programmer named Damian Isla is credited with introducing a pragmatic behavior tree model for the Halo 2 game agent AI. I wanted to know if this is captured in the Grokipedia article.
It isn’t. Granted this may be because behavior trees are not exclusive to game programming so not delving into this makes some sort of sense. However; as a history junkie I ever desire that media help tell the real story of how we all get to enjoy the things we do, behavior trees in video games being one of them. I left feedback, for consideration from the AI curator.
I sought an advanced behavior tree implementation on Grokipedia which brought me to Event Driven Finite State Machine. A behavior tree can be seen as a form of finite state machine (re: Behavior Trees in Robotics and AI: An Introduction). I enjoy seeing the C program. This article is sparer than I would have expected like this is a work in progress article.
The Event Driven Programming article is comprehensive. I would recommend this to someone wanting to understand event driven programming more. I wish it included instructions on how to formulate events in a system, such as eventstorming (a play on the word brainstorming) and making sure that events are discussed exclusively as past tense happenings in the system.
I thought the Godot (Game Engine) article is good too. It captures the uniqueness of Godot which is the tree based programming model.
All of this to say I’m excited for the Grokipedia project. If an AI is truly capable of distilling information down by maximizing its truth seeking ability and that results in articles that are extremely high quality like some of the ones we talked about today then everyone should rejoice.
I still see some information gaps I’d like to see shored up before it is declared fit for prime time (the v1.0 release). Hopefully we’ll all see improvement soon!