Unknown's avatar

Tripping on AI

James Tanner (FSC teacher) is sold on using AI to assist in genealogical research. Katherine Borges, Director of ISOGG, has lost her soul to AI. Respected members of MoCoGenSo are beginning to use AI.
So what should I do?

There was 1966 and ELIZA, the first chatbot.
In 1968 the film 2001: Space Odyssey is released
March 2016 Google DeepMind’s AlphaGo defeats Go champion Lee Sedol. Then Nov 30 2022 – the first public release of ChatGPT (GPT-3.5)

2 excellent YouTube presentations for your viewing pleasure:

==) David Chalmers, NYU professor of Philosophy and Neural Science

3 mn old presentation: “When we talk to AI, what are we talking to?”

https://youtu.be/uUjJOMcNU9w?is=E9ggLSrl8e_czkvt

==) Adam Brown, of Blueshift, Google Deepmind

4mn old presentation: “Training Sand to Think: Artificial General Intelligence & Future of Physics”

https://youtu.be/Mw60FH5iflI?is=Yokjhxqnbl3omCQ2

==) Some ways to improve AI results:

1. scale (bigger neural net)
2. more and better data
3. ask nicely = chain of thought = add “think step by step” to your prompt 4. thinking for a long time
5. reinforcement learning
6. conversation between multiple LLMs

imho. AI now is like early airplanes. Eventually airplanes were used heavily by military and little bits by wealthy experimenters. Then business use was developed, mainly for employee movement. Finally, use by middle to upper class for pleasure. How will AI progress within human culture?

Leave a Reply