Doug Lenat does indeed belong in that Pantheon of greats in artificial intelligence, including Marvin Minsky, John McCarthy, Alan Newell, and Herb Simon. His work was seminal and hi death is a deep loss. Thanks for letting us know, Gary.
Sep 1, 2023·edited Sep 1, 2023Liked by Gary Marcus
Sorry to hear the passing of Doug Lenat. While I disagree with all symbolic approaches to common sense, including Cyc, I'm grateful to Lenat for his work in the field since it helped me form my own hypothesis. May he rest in peace.
Sad to hear this. We used to talk about Cyc quite a bit in the 1990s but it fell off the radar for most of us.
It's encouraging to hear that you are devoted to integrating neural network and symbolic approaches. I'm not an AI researcher but did study the philosophy of AI with a focus on nn vs. symbolic approaches. I don't believe that NN/LLMs will generalize well despite their astounding capabilities in some areas. We absolutely need the symbolic approach to advance.
This is very sad news. Lenat was a giant in the field, and one of the greats.
His bravery in forging ahead to catalog the human knowledge and reasoning is admirable. That's how we move forward, by going into the unknown at all odds. We may succeed or we may fail, but we try mightily hard.
Sep 2, 2023·edited Sep 2, 2023Liked by Gary Marcus
It's been said that the true impact of a scientist is measured not just by what they discover but also by what they enable others to discover. In that sense, Lenat's contributions are invaluable. Even if Cyc hasn't solved all the problems it aimed to, the discussions it sparked, the research it informed, and the debates it fueled have been integral to the AI community's understanding of what 'common sense' in machines could look like.
As we ponder the future of AI, particularly the intersect between neural networks and symbolic systems, we cannot ignore the foundation that Lenat laid. The question now arises: who will pick up the torch and continue his monumental work? Is the source code of Cyc something that could be made accessible to drive future innovations in AI?
May he rest in peace, and may his intellectual legacy continue to inspire future generations of AI researchers, thinkers, and enthusiasts.
I'm so sorry to learn of Lenat's passing -- a tremendous loss to the AI community! I'm a computer scientist at a national laboratory and I've been suggesting for a while that someone look into using OpenCyc as a way to mitigate LLM hallucinations and otherwise improve responses.
Condolences, Gary. Thanks for letting us know. I briefly new/met Doug in 1986/7. Some of my Kodak colleagues and I were early “Cyclists”. Our company sponsored several projects at MCC (Microelectronics and Computer Technology Consortium). I remember my manager insisting on one day in particular that we should stop by Doug’s office to meet someone important. There was Doug in conversation with Marvin Minsky. At 25 years old, I had no idea who they were then. Nor did they know me. :) I’ll remember fondly our time on the bleeding edge of AI. RIP Doug (and Marvin). 🚴🏾
Loved the paper that you two wrote. I agree that the next gen solution will need to incorporate some form of symbolic reasoning with the LLM models. Reminds me of Roger Schank and Bob Abelson's "Scripts, Plans, Goals and Understanding".
Oh, how sad. I recall the excitement when Lenat first started the Cyc project -- and the disappointment that followed. Yet, as you point out, it wasn't a complete failure, and still seems to hold a lot of potential value, that we haven't figured out how to unlock.
Sad to hear the news, but thanks for sharing it here. I met Doug Lenat once when we presented at a small workshop, was great to finally meet my first "AI Hero". I learned of his work when I was a kid and it got me extra-inspired about AI. I'm always reminding people of Cyc as well as Lenat's other work such as Automated Mathematician, etc. Thanks, Doug, for being such an inspiration.
OMG, so sorry :(
I had the honor, privilege, to be on Cyc for a brief while - I cherish those memories to this day.
RIP, dear Doug :(
Thank you for this beautiful tribute.
Doug Lenat does indeed belong in that Pantheon of greats in artificial intelligence, including Marvin Minsky, John McCarthy, Alan Newell, and Herb Simon. His work was seminal and hi death is a deep loss. Thanks for letting us know, Gary.
Sorry for your loss, Gary. I've been aware of Cyc since, it seems like forever. I urge everyone to read that short paper. It's sobering.
Sorry to hear the passing of Doug Lenat. While I disagree with all symbolic approaches to common sense, including Cyc, I'm grateful to Lenat for his work in the field since it helped me form my own hypothesis. May he rest in peace.
Mistaking code for common sense; data for decision making; networks for neurons. These fallacies need to fall. Here's hoping symbolism is Cyclical.
Sad to hear this. We used to talk about Cyc quite a bit in the 1990s but it fell off the radar for most of us.
It's encouraging to hear that you are devoted to integrating neural network and symbolic approaches. I'm not an AI researcher but did study the philosophy of AI with a focus on nn vs. symbolic approaches. I don't believe that NN/LLMs will generalize well despite their astounding capabilities in some areas. We absolutely need the symbolic approach to advance.
This is very sad news. Lenat was a giant in the field, and one of the greats.
His bravery in forging ahead to catalog the human knowledge and reasoning is admirable. That's how we move forward, by going into the unknown at all odds. We may succeed or we may fail, but we try mightily hard.
It's been said that the true impact of a scientist is measured not just by what they discover but also by what they enable others to discover. In that sense, Lenat's contributions are invaluable. Even if Cyc hasn't solved all the problems it aimed to, the discussions it sparked, the research it informed, and the debates it fueled have been integral to the AI community's understanding of what 'common sense' in machines could look like.
As we ponder the future of AI, particularly the intersect between neural networks and symbolic systems, we cannot ignore the foundation that Lenat laid. The question now arises: who will pick up the torch and continue his monumental work? Is the source code of Cyc something that could be made accessible to drive future innovations in AI?
May he rest in peace, and may his intellectual legacy continue to inspire future generations of AI researchers, thinkers, and enthusiasts.
I'm so sorry to learn of Lenat's passing -- a tremendous loss to the AI community! I'm a computer scientist at a national laboratory and I've been suggesting for a while that someone look into using OpenCyc as a way to mitigate LLM hallucinations and otherwise improve responses.
Condolences to his family, in and outside AI.
I hope Cyc become openly available so a larger community can inspire from and continue his path. May his memory be eternal.
Condolences, Gary. Thanks for letting us know. I briefly new/met Doug in 1986/7. Some of my Kodak colleagues and I were early “Cyclists”. Our company sponsored several projects at MCC (Microelectronics and Computer Technology Consortium). I remember my manager insisting on one day in particular that we should stop by Doug’s office to meet someone important. There was Doug in conversation with Marvin Minsky. At 25 years old, I had no idea who they were then. Nor did they know me. :) I’ll remember fondly our time on the bleeding edge of AI. RIP Doug (and Marvin). 🚴🏾
A beautiful tribute!
Loved the paper that you two wrote. I agree that the next gen solution will need to incorporate some form of symbolic reasoning with the LLM models. Reminds me of Roger Schank and Bob Abelson's "Scripts, Plans, Goals and Understanding".
Oh, how sad. I recall the excitement when Lenat first started the Cyc project -- and the disappointment that followed. Yet, as you point out, it wasn't a complete failure, and still seems to hold a lot of potential value, that we haven't figured out how to unlock.
We're still in the early days of AI, aren't we?
Sad to hear the news, but thanks for sharing it here. I met Doug Lenat once when we presented at a small workshop, was great to finally meet my first "AI Hero". I learned of his work when I was a kid and it got me extra-inspired about AI. I'm always reminding people of Cyc as well as Lenat's other work such as Automated Mathematician, etc. Thanks, Doug, for being such an inspiration.