Weekly Web Harvest for 2023-04-02

[2108.07258] On the Opportunities and Risks of Foundation ModelsThough foundation models are based on standard deep learning and transfer learning, their scale results in new emergent capabilities,and their effectiveness across so many tasks incentivizes homogenization. Homogenization provides powerful leverage but demands caution, as the defects of the foundation model are inherited by all the adapted […]

Read More… from Weekly Web Harvest for 2023-04-02

Weekly Web Harvest for 2023-03-26

nomic-ai/gpt4all: gpt4all: a chatbot trained on a massive collection of clean assistant data including code, stories and dialoguegpt4all: a chatbot trained on a massive collection of clean assistant data including code, stories and dialogue ‘Indirect prompt injection’ attacks could upend chatbotsIn response to a discussion of his team’s report on Hacker News Forum, Greshake said, […]

Read More… from Weekly Web Harvest for 2023-03-26

Brute Forcing Canvas Catalog

Canvas Catalog is a front-end Instructure sells so that institutions can create a storefront for courses held in Canvas. In our case, we’re starting to do some paid non-credit courses. While I had a variety of interactions with Catalog back at VCU, I wasn’t deeply involved and most of our goals were fairly linear. It’s […]

Read More… from Brute Forcing Canvas Catalog

Making it work: Using Gravity Forms and URL parameters with Pardot and Canvas

We’ve been working with MIIS to get a non-credit course going in Canvas through Catalog.1 It’s been interesting to see what works smoothly and what does not. We have a couple major systems that sort of work together but are requiring different paths for exceptions. We’ve got a nice path for non-Middlebury people to sign […]

Read More… from Making it work: Using Gravity Forms and URL parameters with Pardot and Canvas