In the last year, new generative AI platforms have made headlines. Large Language Models (LLMs) like ChatGPT, Bard, LLaMA and claude have been compared on their abilities to perform
- knowledge and reasoning tasks (answering questions, solving problems with external tools),
- language manipulation (generating summaries, translating text, “reversioning” stories), and
- common data analyses (structuring data, tagging and clustering data, writing and fixing code).
Unlike AI applications of the past, these platforms performed these tasks reasonably well “out of the box,” without a lot of extra training. This means prototyping a new AI task can be as simple as typing a prompt into ChatGPT.
Do you have a computing idea you would like to try, perhaps related to some data you’re looking at? Or some documents or a web site? Do you have an idea to engage audiences around some AI application? What about tools for local newsrooms?
The second Wednesday of every month, join us at the Brown Institute from 6-9pm for a dine-and-design event — open time to explore, to ask questions, to socialize with other students and practicing journalists, all interested in generative AI. A light dinner will be served.
We’ll talk about good design approaches, and about the strengths and weaknesses of these models. What works and what doesn’t? How do we make sure the platform is performing as we expect? What are the ethical concerns?
A light dinner will be served and perhaps a 10-minute presentation will kick off the evening with an inspiring example.
Students with all technical backgrounds are welcome. Come, build!