I broadly agree with this argument and you’ve articulated it superbly. This is an excellent post. Thank you for writing this.
The only part I really strongly disagree with is about the intentions behind the policy recommendations of AI 2027 — I think the policy recommendations and the reasons stated for them are sincere and honest, not cynical and manipulative. Whether some people at some AI companies are invoking similar ideas for cynical, manipulative reasons, I can’t say. All I can say is that all the evidence points toward the conclusion that the vast majority of people who want regulations due to concerns about AGI safety/alignment are completely sincere “true believers” and don’t have a financial motive.
The Beginning of Infinity by David Deutsch is one of my favourite books. His arguments in that book are incredibly creative and feel really promising. But I still treat them as creative conjecture rather than settled truth. So, I’m inclined to agree with almost all your points, and I’m strongly inclined to agree with some of them, but I wouldn’t state them as being absolutely true. That’s the caveat.
However, the rationalist/LessWrong/MIRI view on AGI risk is also conjecture, and it’s much worse conjecture than the arguments Deutsch has made (or that you articulated in this post). I think it is highly constructive and highly important to challenge these conjectures with other conjectures like Deutsch’s (or yours). Deutsch’s view is more coherent and more appealing in other ways (e.g. seems to rest on better philosophical/scientific foundations), so it’s at least preferable to the rationalist/LessWrong/MIRI view, even if it’s not correct.
Thank you for this thoughtful and generous comment, Yarrow. I appreciate it.
On your first point, I think your criticism is well placed. I should not have psychologized the intentions behind the policy recommendations in AI 2027. The argument does not require that the authors be cynical, and I have no reliable way of knowing their motivations. Their recommendations are entirely consistent with sincere concern viewed through their framework. After reconsidering this, I agree that my original framing was uncharitable. I will revise that section of the post, and I should also be clear that on this particular issue I diverge from Brett Hall’s interpretation. These fatalistic views are ones I myself once held with full sincerity, so presuming insincerity in others would be unjustified.
On your second point, I fully agree that all knowledge is conjectural. Deutsch emphasises fallibilism strongly, and I would never claim any of these arguments as settled truth. They are conjectures offered because, at present, they seem to be better explanations than the rationalist alternatives.
Thank you again for engaging with the post so carefully. Your comment improved the argument and helped catch a place where my own framing fell short of proper charity.
Edit: I have added several clarifying notes to the post (marked as “Edit”). I hope these address your well placed criticism and correct the earlier lack of charity.
There is a sense in which all knowledge is conjectural. The Standard Model of particle physics is conjectural in some sense. But what I’m saying about Deutsch’s epistemology and philosophy of science — and especially his philosophy of mind and overall theory of progress — is much stronger than that. I have tried to search in the past for any academic philosophers who have seriously discussed The Beginning of Infinity. I haven’t found much so far.
I’m impressed with the book, but in philosophy, it’s always possible to be impressed with arguments and then not to notice important shortcomings until someone points them out. I would love to see The Beginning of Infinity get more scrutiny from academic philosophers and possibly other experts as well, like scientists and historians of science.
Right, exactly. Even seemingly self evident mathematical claims are not exempt from this. We can never rule out the possibility of error, so epistemological certainty is unattainable. People can feel certain, but whenever they do, they are mistaken about the status of that feeling.
This is often confused with relativism. We say true things all the time. “2 + 2 = 4” is true, yet there is no final and infallible way to confirm it. Nothing is immune to criticism.
The Beginning of Infinity is my favorite book. I have been active in the critical rationalist circles that have formed around David Deutsch’s work. If you want pointers, I know quite a few people in that community. If you are looking for sharp criticisms of Deutsch, asking his admirers may not be ideal. Even though we fully accept that Deutsch is fallible like anyone else, most of us agree with him on the major points.
It is good that you want to find strong criticisms of his ideas. I hope you succeed. To be frank, I have yet to encounter substantial critiques of The Beginning of Infinity, though I am sure there are errors in it.
My friend Logan founded the Conjecture Institute this year. It might interest you:
I am specifically interested in commentaries on The Beginning of Infinity from professional academic philosophers, as well as, to a lesser extent, commentaries from scientists or people with other relevant expertise, such as historians of science. I haven’t found much written by philosophers about the book or about Deutsch’s ideas in general.
The work of “fans” or “followers” is much less interesting to me because I get why Deutsch’s ideas are appealing and would inspire a zealous following, but this is not a critical appraisal.
Hi, I’ve been an independent, non-academic philosopher for over 20 years. I was mentored by David Deutsch for a decade and I helped him with The Beginning of Infinity. I’m no longer a fan of Deutsch and I have commentary, much of it critical, on Deutsch’s work:
I broadly agree with this argument and you’ve articulated it superbly. This is an excellent post. Thank you for writing this.
The only part I really strongly disagree with is about the intentions behind the policy recommendations of AI 2027 — I think the policy recommendations and the reasons stated for them are sincere and honest, not cynical and manipulative. Whether some people at some AI companies are invoking similar ideas for cynical, manipulative reasons, I can’t say. All I can say is that all the evidence points toward the conclusion that the vast majority of people who want regulations due to concerns about AGI safety/alignment are completely sincere “true believers” and don’t have a financial motive.
The Beginning of Infinity by David Deutsch is one of my favourite books. His arguments in that book are incredibly creative and feel really promising. But I still treat them as creative conjecture rather than settled truth. So, I’m inclined to agree with almost all your points, and I’m strongly inclined to agree with some of them, but I wouldn’t state them as being absolutely true. That’s the caveat.
However, the rationalist/LessWrong/MIRI view on AGI risk is also conjecture, and it’s much worse conjecture than the arguments Deutsch has made (or that you articulated in this post). I think it is highly constructive and highly important to challenge these conjectures with other conjectures like Deutsch’s (or yours). Deutsch’s view is more coherent and more appealing in other ways (e.g. seems to rest on better philosophical/scientific foundations), so it’s at least preferable to the rationalist/LessWrong/MIRI view, even if it’s not correct.
Thank you for this thoughtful and generous comment, Yarrow. I appreciate it.
On your first point, I think your criticism is well placed. I should not have psychologized the intentions behind the policy recommendations in AI 2027. The argument does not require that the authors be cynical, and I have no reliable way of knowing their motivations. Their recommendations are entirely consistent with sincere concern viewed through their framework. After reconsidering this, I agree that my original framing was uncharitable. I will revise that section of the post, and I should also be clear that on this particular issue I diverge from Brett Hall’s interpretation. These fatalistic views are ones I myself once held with full sincerity, so presuming insincerity in others would be unjustified.
On your second point, I fully agree that all knowledge is conjectural. Deutsch emphasises fallibilism strongly, and I would never claim any of these arguments as settled truth. They are conjectures offered because, at present, they seem to be better explanations than the rationalist alternatives.
Thank you again for engaging with the post so carefully. Your comment improved the argument and helped catch a place where my own framing fell short of proper charity.
Edit: I have added several clarifying notes to the post (marked as “Edit”). I hope these address your well placed criticism and correct the earlier lack of charity.
Great response! Thank you!
There is a sense in which all knowledge is conjectural. The Standard Model of particle physics is conjectural in some sense. But what I’m saying about Deutsch’s epistemology and philosophy of science — and especially his philosophy of mind and overall theory of progress — is much stronger than that. I have tried to search in the past for any academic philosophers who have seriously discussed The Beginning of Infinity. I haven’t found much so far.
I’m impressed with the book, but in philosophy, it’s always possible to be impressed with arguments and then not to notice important shortcomings until someone points them out. I would love to see The Beginning of Infinity get more scrutiny from academic philosophers and possibly other experts as well, like scientists and historians of science.
Right, exactly. Even seemingly self evident mathematical claims are not exempt from this. We can never rule out the possibility of error, so epistemological certainty is unattainable. People can feel certain, but whenever they do, they are mistaken about the status of that feeling.
This is often confused with relativism. We say true things all the time. “2 + 2 = 4” is true, yet there is no final and infallible way to confirm it. Nothing is immune to criticism.
The Beginning of Infinity is my favorite book. I have been active in the critical rationalist circles that have formed around David Deutsch’s work. If you want pointers, I know quite a few people in that community. If you are looking for sharp criticisms of Deutsch, asking his admirers may not be ideal. Even though we fully accept that Deutsch is fallible like anyone else, most of us agree with him on the major points.
It is good that you want to find strong criticisms of his ideas. I hope you succeed. To be frank, I have yet to encounter substantial critiques of The Beginning of Infinity, though I am sure there are errors in it.
My friend Logan founded the Conjecture Institute this year. It might interest you:
https://www.conjectureinstitute.org/
I am specifically interested in commentaries on The Beginning of Infinity from professional academic philosophers, as well as, to a lesser extent, commentaries from scientists or people with other relevant expertise, such as historians of science. I haven’t found much written by philosophers about the book or about Deutsch’s ideas in general.
The work of “fans” or “followers” is much less interesting to me because I get why Deutsch’s ideas are appealing and would inspire a zealous following, but this is not a critical appraisal.
Hi, I’ve been an independent, non-academic philosopher for over 20 years. I was mentored by David Deutsch for a decade and I helped him with The Beginning of Infinity. I’m no longer a fan of Deutsch and I have commentary, much of it critical, on Deutsch’s work:
Fundamental Philosophical Errors in Taking Children Seriously
The Historicism of David Deutsch
Analysis of David Deutsch’s The Final Prejudice
David Deutsch Smears Ayn Rand
Deutsch Misquoted Turing
Misquotes by David Deutsch
You can contact me at curi@curi.us if you want to discuss further.