Can I use GPT-o1-preview to “read” a book?
I saw tylercowen today had a link to the book pdf for "The Line: AI and the Future of Personhood" by James Boyle. Seems super interesting! Full of interesting questions I think about a lot. BUT it's 335 pages end to end, I want to get the main points and I don't read THAT fast. What am I to do?
Let's use ChatGPT o1-preview.
Stepping back, what are the main things I'd like to get out of the book overall? I want the main arguments, the interesting ideas, and an understanding of the main sources.
Great. However, o1 doesn't allow uploading files...
No problem - I had GPT-4o write me a python script so I could take the pdf and make a text file for each chapter. I also had to modify the script to split chapters if >65K characters to fit o1. I also pre-pended the prompt onto each file to make copy-pasting easier.
(NB: I now use Cursor basically every day for this type of projects. Not having AI-assisted coding feels like pre-history. And not having AI as a thought-partner feels like working with one hand behind my back…)
With that in hand, I then had o1 process the following prompt:
"For the following text, i want you to pull the following:
1. Main arguments for each chapter and for the book overall
2. Main "catchphrases" / novel ideas that he uses (e.g., new terms or analogies coined and then referenced a few times)
3. Novel ideas (especially counter-intuitive ones)
4. Sources mentioned most often
For each of these I also want key passages that capture these points."
The outcome: not bad!
Each section took about 20-50 seconds for o1 to process. I also ended up making new chats for each chapter since using the same chat ended up having o1 merge the summaries vs. just giving me the summaries for that particular chapter.
The insights - super valuable! The prompt helps me quickly understand which chapters I'd like to dive into in more details... I can also probe each chapter in the chat itself to dive deeper & connect to all of o1's other context
As for the prompt the most valuable pieces were:
The Catchphrases/novel ideas - it helps me grok the arguments and gives a quick feeling for some of the novel ideas
"Source mentioned most often" - also super helpful since I can see the clear themes (e.g., "Samuel Butler’s Erewhon", "Blade Runner (Film) and Philip K. Dick’s Do Androids Dream of Electric Sheep? (Book)", "Adam Smith’s The Theory of Moral Sentiments", "Alan Turing’s Work on Machine Intelligence" etc.) (e.g., you can also get a vibe check since the sources include Eliezer, Hinton and Kurzweil for chapter 3)
"Key passages": having it pull quotes helps show *the feeling* behind the arguments
(A full example of the output is included below to give you a taste - it’s quite robust and far from a bland summary).
As o1 gets a larger context window / allows uploads, this will be way easier and more valuable.
Could NotebookLM have been better? I tried uploading the files and using the "out of the box" summaries / study guides - they were OK but way higher level than o1. What about NotebookLM Deep dive audio? Also very high-level. Not even close to the same level of detail as o1.
Was it better than reading the book? Definitely not since this book actually looks quite thorough and thought-provoking. Consider this more a way of quickly groking what parts of the book are worth diving into deeper.
For the average book though... this might be good enough…
As for The Line, a lot of it looks great - in particular the chapters on Corporations and Nonhuman animals… lots to reflect on.
Example output for the corporations chapter section:
1. Main Arguments for Each Chapter and the Book Overall
Chapter 3: Corporations
Main Arguments:
• Flawed Legal History of Corporate Personhood: The chapter discusses how corporations gained equal protection rights under the Fourteenth Amendment through a flawed legal process, notably without proper argument or authoritative decision, and how this history undermines confidence in corporate constitutional rights.
• Key Passage: “The decision to give corporations rights under the Fourteenth Amendment was a hugely consequential one. Regardless of one’s views on the merits, it deserved careful legal analysis and reasoned decision-making. To have that decision made with ‘neither argument nor discussion’ was a major failure of the legal process. To have it rest on a court reporter’s account of a discussion before oral argument was a travesty.” (Page 163)
• Implications for AI Rights: The flawed history of corporate personhood may influence how society approaches the question of granting rights to Artificial Intelligences (AIs). Liberals, who might be expected to support extending rights, could oppose AI rights due to parallels with corporate rights expansion.
• Key Passage: “One might have imagined that liberals would be the most ardent defenders of personhood for our new silicon brothers and sisters… But if liberals view the decision over AI rights through the lens of the history of corporate personhood—from Santa Clara to Citizens United—then they might be its keenest foes.” (Page 163-164)
• Debate Over Corporate Constitutional Rights: There are strong arguments on both sides regarding whether corporations should have equal protection rights. The chapter emphasizes that this issue remains unresolved and contentious.
• Key Passage: “Where does all of this leave us now, in the era of Citizens United, on the question of equal protection rights for corporations? Set aside the checkered history. What is the right answer? Even leaving the history out of it, the case for corporate equal protection rights is surprisingly shaky.” (Page 166)
• Need to Rethink Corporate Rights: The lack of a convincing history or moral/legal theory justifying corporate rights suggests that society might need to reconsider the extent of rights granted to corporations.
• Key Passage: “One can believe it useful to grant corporations legal personhood… Yet when one turns to the larger questions about the full suite of constitutional rights to which corporations should be entitled, one finds that we have neither a convincing history nor a convincing moral or legal theory that justifies where we have ended up.” (Page 168)
Chapter 4: Nonhuman Animals
Main Arguments:
• Legal Personhood for Animals: The chapter introduces the case of Tommy the chimpanzee and discusses efforts by the Nonhuman Rights Project to grant legal personhood to certain nonhuman animals based on their cognitive abilities and rights.
• Key Passage: “The claim was brought by the Nonhuman Rights Project, founded by Steven Wise. The project has an audacious mission statement. The first goal is ‘[t]o change the common law status of great apes, elephants, dolphins, and whales from mere “things,” which lack the capacity to possess any legal right, to “legal persons,” who possess such fundamental rights as bodily liberty and bodily integrity.’” (Page 162-163)
• Approaches to Arguing for Animal Rights: The chapter explores two main arguments used to advocate for animal personhood: the qualities-of-mind argument (based on cognitive abilities) and the legal-rightsholders argument (based on existing rights animals possess).
• Key Passage: “In this chapter, I will explore both of these approaches, which I will call the qualities-of-mind argument and the legal-rightsholders argument.” (Page 165)
• Parallels with AI Personhood Debate: The discussions around animal personhood provide insights and a preview into the upcoming debates about AI personhood and how society defines personhood.
• Key Passage: “My claim is that these debates are both extremely important in their own right and a fascinating, sometimes surprising preview of what our debates about AI and transgenic species will be.” (Page 165)
Book Overall:
Main Arguments:
• Reexamining Personhood: The book argues that as society confronts new entities like AI and transgenic species, we must critically reassess our definitions of personhood and the moral and legal lines we draw.
• Key Passage: “As I said earlier, grappling with the question of synthetic Others may bring about a reexamination of the nature of human identity and consciousness… We have to take a hard look at the current places where we draw the line, at the legitimacy of our distinctions and the force of our reasoning.” (Page 170)
• Lessons from Corporate and Animal Personhood: The history and debates over corporate and animal personhood highlight the complexities and inconsistencies in our understanding of personhood, which are crucial as we consider rights for AI.
• Key Passage: “When we come to debate the question of whether Artificial Intelligences should be given legal personality… we will turn inevitably to both the theory and the law of corporate personhood and corporate fundamental rights… In both cases, as I hope this chapter has shown, the answer is, at best, that we are still debating the issue.” (Page 169)
• Challenging Established Notions: The book emphasizes that established notions of personhood are neither inevitable nor logically dictated, and they may need to be rethought in light of new challenges.
• Key Passage: “It might also convince us that the decisions we have made so far about corporate persons are neither inevitable nor dictated by precedent or logic. It might lead us, in fact, to rethink them.” (Page 170)
2. Main “Catchphrases” / Novel Ideas
• “Fruit of a Poisonous Tree”: Refers to the idea that the entire jurisprudence of corporate rights under the Fourteenth Amendment is tainted due to its flawed origins.
• Key Passage: “These facts raise some obvious questions. Is our entire jurisprudence of equal protection for corporations the fruit of a poisonous tree and therefore worthless?” (Page 163)
• “Brazen Historical Forgery”: Describes Conkling’s exaggerated or fraudulent arguments about the framers’ intent regarding corporate rights.
• Key Passage: “Should we see the Fourteenth Amendment rights of corporations as being built on the ‘brazen historical forgery’ of a persuasive argument about the intent of the framers?” (Page 163)
• “Sedimentary Layers”: Used to illustrate how flawed legal precedents accumulate over time, solidifying into established law.
• Key Passage: “As for Davis’s influence, we now have 135 years of precedent accreted in sedimentary layers around that initial flawed process.” (Page 164)
• “Pearl Built Up Around a Grain of Sand”: An analogy for how legal precedent can develop from an initial flawed decision into a significant body of law.
• Key Passage: “(‘Like a pearl built up around a grain of sand,’ the enthusiast for corporate rights might say.)” (Page 164)
• “Station on the Kantian Railway Line of Rights”: Suggests the progressive extension of rights to more groups over time.
• Key Passage: “After all, are they not eager to add another station on the Kantian railway line of rights extending to ever larger groups—first white men, then all men, then women, and so on?” (Page 163-164)
• “Our Own Voight-Kampff Test”: References the test from “Blade Runner” to determine if an entity is sentient, drawing parallels to how society might assess AI.
• Key Passage: “When analogies are made between corporations and Artificial Intelligences, I would expect both of those points to be made with great force, and appropriately so.” (Page 163)
• “Personhood Exceptionalism”: The idea that humans consider themselves uniquely deserving of personhood status.
• Key Passage: “Nonhuman animals eat, sleep, procreate, feel pain, and die. What, if anything, justifies the stark moral distinctions that we draw between the human and nonhuman animal world, distinctions that form at least one basis for our personhood exceptionalism?” (Page 166)
• “Performative Contradiction”: Highlights the inconsistency in using abstract reasoning to argue against the importance of abstract thought.
• Key Passage: “Arguing, using abstract logic, about the irrelevance of abstract thought involves an obvious performative contradiction.” (Page 172)
• “Qualities-of-Mind Argument” and “Legal-Rightsholders Argument”: The two approaches used to argue for nonhuman animal personhood.
• Key Passage: “In this chapter, I will explore both of these approaches, which I will call the qualities-of-mind argument and the legal-rightsholders argument.” (Page 165)
3. Novel Ideas (Especially Counter-Intuitive Ones)
• Liberals Opposing AI Personhood Due to Corporate Rights History: The counter-intuitive notion that liberals might resist granting rights to AI because of the problematic history of corporate personhood.
• Key Passage: “But if liberals view the decision over AI rights through the lens of the history of corporate personhood—from Santa Clara to Citizens United—then they might be its keenest foes.” (Page 163-164)
• Reconsidering Established Corporate Rights: The idea that current corporate rights are not based on solid historical or logical foundations and may need to be reevaluated.
• Key Passage: “It might also convince us that the decisions we have made so far about corporate persons are neither inevitable nor dictated by precedent or logic. It might lead us, in fact, to rethink them.” (Page 170)
• Debates Over Animal Personhood as a Preview for AI Rights: Using the current legal and moral debates over animal rights as a way to understand and anticipate the complexities of AI rights discussions.
• Key Passage: “My claim is that these debates are both extremely important in their own right and a fascinating, sometimes surprising preview of what our debates about AI and transgenic species will be.” (Page 165)
• Flawed Legal Processes Affecting Modern Rights Discussions: Recognizing that historical legal mistakes continue to influence current jurisprudence and debates over rights for non-traditional entities.
• Key Passage: “Confident extrapolation is hardly in order despite our hundreds of years of experience with corporate personality.” (Page 169)
4. Sources Mentioned Most Often
• Santa Clara County v. Southern Pacific Railroad (1886): A pivotal Supreme Court case often cited regarding corporate personhood under the Fourteenth Amendment.
• Key Passage: “I am less convinced than Graham of the inevitability of a decision like Santa Clara, even if it could have been better reasoned, or actually reasoned at all.” (Page 165)
• Conkling’s Arguments and Court Reporter Davis’s Headnote: The flawed arguments and misrepresentations that contributed to granting corporations Fourteenth Amendment rights.
• Key Passage: “Perhaps the Supreme Court was moved by Conkling’s claims, some of which seem exaggerated, if not a ‘brazen historical forgery.’… As for Davis’s influence, we now have 135 years of precedent accreted in sedimentary layers around that initial flawed process.” (Page 164)
• Historian Graham: Provided analysis on the development of corporate personhood and the irrelevancy of the framers’ intentions.
• Key Passage: “Less frequently cited are Graham’s own conclusions about the reasons for protecting corporations under the Fourteenth Amendment…” (Page 165)
• Supreme Court Justices:
• William O. Douglas and Hugo Black: Both advanced arguments against expansive corporate constitutional rights.
• Key Passage: “Indeed, two influential Justices, William O. Douglas and Hugo Black, advanced versions of each of these arguments in dissent in a 1949 case called Wheeling Steel v. Glander.” (Page 167)
• William Rehnquist: Criticized corporate rights expansions.
• Key Passage: “Indeed, this is one reason why constitutional originalism seems unpersuasive to many… For example, in Obergefell v. Hodges… It seems unlikely that those who wish to use originalist arguments to constrain the power of corporations would agree to accept originalist understandings of the amendment for categories such as sex and sexual orientation.” (Page 168)
• John Paul Stevens: Dissented in Citizens United, questioning corporations as members of “We the People.”
• Key Passage: “Earlier, I quoted Stevens’s dissent in Citizens United. True, he was dealing with a different constitutional right, but the application to equal protection seems clear…” (Page 167)
• Supreme Court Cases:
• Citizens United v. FEC: A case that expanded corporate political rights, often discussed in debates over corporate personhood.
• Key Passage: “The Citizens United decision was presented by its critics as something new, as an awful and unprecedented expansion of rights to entities never thought to be ‘members of “We the People” by whom and for whom our Constitution was established.’” (Page 168)
• Bush v. Gore, Roe v. Wade, Obergefell v. Hodges: Cases mentioned in discussions of legal precedent and constitutional interpretation.
• Key Passage: “The Supreme Court’s decision to overturn Roe v. Wade, perhaps throwing many other constitutional rights into doubt, puts the obvious asterisk after this statement.” (Page 164)
• Nonhuman Rights Project and Steven Wise: Advocates for granting legal personhood to certain nonhuman animals.
• Key Passage: “The claim was brought by the Nonhuman Rights Project, founded by Steven Wise. The project has an audacious mission statement…” (Page 162-163)
• Philosophers and Scholars:
• Jeremy Bentham: Referenced regarding animal ethics and the capacity to suffer.
• Key Passage: “Jeremy Bentham’s famous line sums the point up nicely: ‘The question is not, Can they reason? nor, Can they talk? but, Can they suffer?’” (Page 166)
• Frans de Waal: Cited for his work on animal cognition and the similarities between human and animal minds.
• Key Passage: “As the great primatologist and ethologist Frans de Waal puts it in his accurately titled book Are We Smart Enough to Know How Smart Animals Are?” (Page 170)
This structured summary captures the main arguments, catchphrases, novel ideas, and frequently mentioned sources from the provided text, along with key passages that illustrate each point.
This is a game changer. The first thing I was taught at university was how to "gut a book". We were expected to read 15+ texts a week and write 3 essays. The only way that was anywhere near feasible was by doing something akin to what o1 is doing here but with a much higher probability of missing key arguments along the way. AI-assisted "book-gutting" is the future
(P.s. your process still sounds time consuming. How long do you think it took to process the chapters with o1 and then read the summaries end to end?)
This is brilliant. Thanks for sharing!