Our New Paper on Distributed Agent Reasoning
We have published a new paper presenting a novel architecture for distributed agent reasoning across independent systems, all under strict data locality.
The work shows how agents on separate Orpius deployments can cooperate through natural-language messages, without shared identifiers, schemas, or any centralised data store.
A promising direction for organisations where data cannot move or be unified.
Read the paper on arXiv:
Distributed Agent Reasoning Across Independent Systems With Strict Data Locality
Orpius can see!
Orpius now supports image retrieval and analysis! Orpius can download images, analyze them, perform activities based on what it sees. Combined with the scheduling capabilities of Orpius, it works even when you’re sleeping, making for one heck of a powerful tool. I’ve also added an exciting feature revealed near the end of this post (hint: it changes how you’ll see things entirely). Btw., in-case you’re reading this as a repost, my name’s Daniel, and my...
Find Tables with Keyword Names
I was working with an SQL Server schema and discovered that one of my tables was named with a keyword; SELECT u.*, a.* FROM [Bpm].User u INNER JOIN [Bpm].Account a ON u.Id = a.UserId WHERE u.Username = 'foo@exammple.com' Msg 156, Level 15, State 1, Line 2 Incorrect syntax near the keyword ‘User’. Turns out ‘User’ is a keyword and I needed to enclose it in square brackets: FROM [Bpm].[User] .... It’s probably best just to...
Option Templates in UWP
Introduction Adding Options Sample Overview Exploring the .NET Standard Library Rendering Options in UWP Conclusion Introduction Just about every app needs a settings screen. A lot of developers choose to simply build-out static UI; hard-wiring buttons and text fields to a setting backing store. If one does this, however, eventually, as the number of settings grows, technical debt increases; making refactoring your settings screen into categories, or changing how the settings are stored or displayed,...
Asynchronous Commanding with Codon FX
Introduction Have you ever created a view-model for your app that contains an ICommand that needs to perform some asynchronous activity? Such as calling a web API or saving data to a file? If you have, you’ll know that the synchronous ICommand interface doesn’t lend itself easily to asynchronous operations. You end up having to build a mini-state-machine to disable and re-enable the command target when the command completes. Wouldn’t it be nice if commands...