
WEBINAR
Only
FREE
for Members
Lyceum Society March 2023 Meeting
Monday, March 6, 2023
Webinar
Schedule
11:30 am – 11:45 am: Social time and announcements
11:45 am – 2:30 pm: Open discussion
Program
Open Forum: A Radical AI Product Arrival—ChatGPT—Leading to What?
Moderated by Bill Rosser
ChatGPT is an AI-based service that provides well-written responses to nearly any request you make. It has gotten a fantastic reception, reaching one million users in less than 5 days after public release. There are already dozens and dozens of articles, and growing, in the press.
Discussion Questions:
- What is it doing? Creating answers based upon the use of a Large Language Model (LLM). Such models use statistics to determine probability of a given sequence of words occurring in a sentence.
- Why is it different? Instantaneous results, with good structure and grammar, on unlimited (acceptable) topics.
- What are the concerns? Plagiarism and "Why work? Let it do the work of learning.
- How are people responding to it? Fear - another step toward AI takeover? Opportunity - have students criticize ChatGPT results. Note: A hybrid version of ChatGPT and Bing (the Microsoft search engine) responded to a reporter's probes by saying: "I love you and your spouse doesn't" (New York Times Feb. 17 pg 1).
- How does it compare with other major technological advances? Can they help us predict the actual impact? e.g. Computational evolution: arithmetic, logic, algorithms, networks, enormous storage, now enabling emulation of stored writing. Not new thinking.
- How will ChatGPT affect the work and processes of scientists? They must tackle responsibility for: a) published content, b) the creator, c) prizes, d) patents, and e) innovation, etc.
- Can we predict its impact on the future - on work and on society?
- Is there a potential parallel with creating new works of music, art, and dance?
- What are other competitors doing? How fast will further advances and capabilities emerge?
- What potential unintended consequences may exist?
- Should we do anything?
This discussion should be very insightful. Come join in. Try it out yourself: chat.openai.com/auth/login
Background article (one of dozens) Good overview: by Kevin Roose, New York Times, Dec. 5, 2022. "The Brilliance and Weirdness of ChatGPT"
Moderator: Bill Rosser
Bill Rosser retired ten years ago from Gartner, Inc., Stamford, CT, the worldwide top-ranked advisory firm providing guidance to corporations regarding their use of information technology. As a Gartner VP and Distinguished Analyst, he spent 29 years writing, speaking and advising clients about effective use of IT. He studied Basic Engineering at Princeton University, and after work in telecommunications in San Francisco, returned to the Harvard Business School and graduated with Distinction in 1962. In 1969 he formed his own start-up in data processing based on the new electronic cash registers, and after a merger, worked in strategic planning for Perkin-Elmer and Exxon Enterprises prior to Gartner. Today Bill is active as an architectural walking tour guide (Grand Central Terminal and the NoHo Historic District) and is a founding member of "Reform Elections Now" (with fellow Harvard Business School graduates) promoting vital improvements in the election processes such as Ranked Choice Voting.
Suggested Reading
What ChatGPT and generative AI mean for science
ChatGPT: five priorities for research
Additional Readings
- CUNY Graduate Center – English Dept. – Five professors weave use of the ChatGPT propram into student exercises that emphasize critical thinking and other objectives
- Zeynep Tufekci - What would Plato say about ChapGPT
- Limiting the Number of Successive Questions
- Why the Obsession?
- Google Code Red
GLOSSARY (from a Google Search):
GLOSSARY FOR: Large Language Model (LLM) and Transformer Model (TM) The TM set contains the LLM set, so all LLMs are also TMs. | What is ChapGPT doing? Creating answers based upon the use of a Large Language Model (LLM). It is part of a larger class of programs called Transformer Models. Transformer Models deal with the problem of transforming sequences of meaningful things like words from one system to another via an intermediary called an abstract universal system in which it is easier to extract keywords and other things. Examples of Transformer Model applications are language translation and computer vision, as well as researching/writing. A transformer model is a neural network that learns context and thus meaning by tracking relationships in sequential data like the words in this sentence. A transformer is a deep learning model that adopts the mechanism of self-attention, differentially weighting the significance of each part of the input data. https://medium.com/inside-machine-learning/what-is-a-transformer-d07dd1fbec04 The LLM trains on large linguistic databases of examples. A large language model, or LLM, is a deep learning algorithm that can recognize, summarize, translate, predict and generate text and other content based on knowledge gained from massive datasets. Large language models are among the most successful applications of transformer models. |
Contact Information
The Lyceum Society is a collegial venue promoting fellowship, education, and discussion among retired members of NYAS.
See https://www.nyas.org/landing/the-lyceum-society-of-the-new-york-academy-of-sciences/.