meeting, addis ababa, ethiopia-83519.jpg

Presentations

“The future of AI and humanity: doom or boom?”

April 2, 2024 at Microsoft New England Research and Development Center (NERD)

Mind First Foundation participated in an event organized and moderated by Dan Elton and co-sponsored by AICamp and Boston Astral Codex Ten. Speakers were MFF founder and Chief Scientist Preston Estep, and Chief Philosophy Officer Brian Manning Delaney.

Dr. Preston Estep talk, 2008 Singularity Summit (short clip)

In 2008 Mind First Foundation founder and Chief Scientist Dr. Preston Estep gave a featured presentation at the Singularity Summit in San Jose, California. In the short clip from that presentation that is featured above, Dr. Estep proposes that future artificial intelligence (AI) will eventually achieve recursive self-improvement, which probably will be increasingly competitive with humans. He suggests that the best way to mitigate growing dangers of powerful AI will be to control it through human-AI merger. At that time, he suggested that brain-computer interface (BCI) technologies were our best bet, and that we should focus on designing and producing better BCI. Since that time, many others have similarly advocated for this approach, most notably Elon Musk, which he is attempting to achieve through his BCI company, Neuralink. However, Dr. Estep’s thinking has changed in two notable ways. First, his recent research (submitted for publication) indicates that advanced AIs might not be as directly competitive with humans as he previously assumed. Second, given this newfound perspective, he has come to more fully appreciate Irving Good’s suggestion that the first ultraintelligent (a.k.a. superintelligent) AI is the last invention that humanity “need ever make provided that the machine is docile enough to tell us how to keep it under control.”* Thus, humanity should retain the priority of creating better human-AI interface technologies, but rather than simply creating these technologies, we should focus on creating the best possible AIs for a range of purposes, including the design of increasingly powerful interface technologies.

*Good, I. J. (1966). Speculations concerning the first ultraintelligent machine. In Advances in computers (Vol. 6, pp. 31–88). Elsevier.