Date: Saturday, May 18th, 2024
Time: 1:00 pm to 4:00 pm
Place: IN PERSON at the Squirrel Hill Carnegie Library
Zoom option is available. [Register for the Zoom meeting]
ATTENTION!
For Parsec’s May 18th Meeting. Due to a glitch with our ZOOM account, once the regular meeting starts as usual, sometime shortly after 1:00 pm, we will ask participants to switch to a new ZOOM link.
That link is: https://bit.ly/NewParsec
That will get you right back into the meeting without waiting. We’ll explain why at the top of the meeting. (Please wait until we ask everyone to switch over during the meeting.) If needed, the meeting ID is 978 0473 3050 and passcode is 669161
If you’re joining the meeting after 1:10 pm or so, just use the new link to start: https://bit.ly/NewParsec
Our Program: An Interactive Discussion on A.I. and Large Language Models. Facilitated by Scot Noel.
A.I. and especially Large Language Models have broken into the public consciousness over the last couple of years. Some absolutely hate them as apocalyptic, while others tout them as the greatest thing since the harnessing of fire. Saying that the truth is somewhere in between, while likely true, can obscure the level at which LLMs are indeed highly disruptive, world-wide technological shift.
We’ll take a look at benefits, issues, and the future of A.I. and Large Language models, from medicine to the military, from baked in bias to accelerated research, and how everything from the computer chip industry to the power grid is already being transformed by a technology that is only getting started.
This meeting is designed not as a presentation, but a series of interactive questions where attendees are encouraged to express their own experience, hopes, fears, and feelings about A.I. Let’s engage with humility, courtesy, curiosity, and most importantly a willingness to hear one another’s points of view.
Some A.I. Benefits:
- Adaptive Personalized Learning (ex: personalized reading and math instruction for K-8 students)
- Improved Translation Software (ex: aiding in language translation in scientific publishing)
- Accelerated Industrial Research (ex: finding sustainable replacements for lithium in batteries)
- LLMs in Medicine (ex: prediction of cancer metastasis and devising clinical treatment responses.)
Some A.I. Issues:
- Copyright and Creative Concerns (ex: the uncertain future of content creation and graphic design)
- The A.I. Cold War (ex: the Loyal Wingman autonomous jet fighter squadron escorts)
- Bias in Large Language Models (ex: training models have biases and stereotypes baked in.)
- Model Collapse (ex: degrading results through A.I.’s own “success” and model pollution by “resisters.” )
Some Changes That Might Go Either Way:
- Macrochips (ex: the Cerebras WSE-3 chips features 4 million transistors and 900,000 cores.)
- Power Grid (ex: A.I. is restructuring the power grid, power hungry and energy efficient.)
- A.I. is a Platform Shift (i.e. a change in the dominant layer that applications are built on.)
- Just getting started (ex: ChatGPT 5 better reasoning, improved accuracy and video support.)
If you missed the March Parsec meeting with guest speaker and 2024 Confluence guest of honor, Richard Kadrey, you can now watch it on our YouTube channel!