Anthropic's AI needle in a haystack test, MicrosoftClippy, Microsoft Bob, Dune and Matrix inspired user experiences and much more

Photo by Tamanna Rumee on Unsplash

Published 11 March 2024


Andy and Michael R get together to talk tech – focusing on LLMs, the return of Clippy, Steam games, physical and virtual user experiences and Cyber 205.


Starting off with an article from Venture Beat where the Claude 3 Opus LLM was subjected to what’s called a “needle-in-a-haystack” test, and reacted with the accurate recognition of the needle as well as a statement that it seemed that the source material seemed so out of place, that it must either be a joke or a deliberate test.  The example source material – the needle – was material about pizza toppings, and was included in a corpus consisting of a random set of other documents.  Given the discussions from e451 about Nightshade and other tools could be used to disrupt scraping of content, it could be inferred that reactions such as the Opus response could defeat such disruption attempts by recognizing how out of place the disruptions are.


Clippy returns with an interesting article from the Wall Street Journal, and has surfaced many times in the Games at Work podcast over the years.  Examples include e329 and e237 among others.  The team also brings back Microsoft Bob as part of the discussion.


Turning to games and gaming, Michael R and Andy focus on several Steam games, including Sid Meier’s Alpha Centauri, Populous and Dungeon Keeper among others.  And with all the focus on the next chapter of Dune, the trailer for the Dune Awakening game certainly captures attention with unkillable sandworms.


Next up is all things UX, starting with the recommendation that physical buttons be returned to automobiles (hooray!).  Then, an Apple Vision Pro example of the Matrix with Magic Room.  Andy and Michael R wrap things up with the Cyber 205.


What would you want to ask Clippy?  Should Clippy be in the Matrix?  Have your bots 🤖 drop our bots 🤖 a line at @[email protected] (our home for now) and let us know! 


These show notes were lovingly hand crafted by a real human, and not by a bot.  All rights reserved.  That’s our story and we’re sticking to it.


Selected Article Links
AI

Venture Beat article: Anthropic’s Claude 3 knew when researchers were testing it


Games at Work e451: Fahrenheit


Wall Street Journal article: The Demoted Microsoft Worker Getting His Revenge


Wikipedia article: Microsoft Bob


Gaming

The Verge article: EA just added classics like Dungeon Keeper, SimCity 3000, and Populous on Steam


SlashFilm article: Try To Survive The World Of Arrakis In The Dune: Awakening Video Game Trailer


Dune Awakening


User Experience

Ars Technica article: European crash tester says carmakers must bring back physical controls


Matrix experience – https://www.threads.net/@nathievr/post/C4D1WZZvSsI 


space.com article: The Matrix movies in order


Apple Vision App Store: Magic Room: Retheme Your Space


Raytracing in Vision Pro: https://www.threads.net/@dreamwieber/post/C3_aWRZvv3u/ 


IT History Society: Cyber 205


Wikipedia article: CDC Cyber