We can't find the internet
Attempting to reconnect
Something went wrong!
Hang in there while we get back on track
A practical reading list for AI’s limits—and the systems that contain failure
•
4 min read
•
130 docs
A curated set of organic recommendations spanning secure-systems research, an AI-research debate, a draft AI-use policy, and a few personal picks from product and tech leaders.
If you read one cluster, make it Martin Casado’s pairing of older systems research with a current AI-cyber concern. He argues that AI cyber capabilities should force us to build secure systems “all the way down” and names HYDRA and Asbestos as two of his favorites. The payoff is specificity: one paper makes protection and extensibility a kernel-design problem; the other turns isolation and information flow into explicit mechanisms.
Start with the systems-security pair
HYDRA: the kernel of a multiprocessor operating system
- Content type / creator: Research paper by W. Wulf et al., published in Communications of the ACM in 1974. Link:ACM record.
- Recommended by: Martin Casado, as one of his two favorite works on building secure systems.
- Key takeaway: HYDRA treats physical and virtual resources as “objects,” then supplies mechanisms for creating types, defining operations, sharing, and protecting references. The paper presents that as a basis both for extending the system and for creating highly secure systems.
- Why it matters: It gives readers a concrete vocabulary for asking where an AI-enabled system’s resources live, who can operate on them, and how those permissions are enforced.
Labels and event processes in the asbestos operating system
- Content type / creator: Research paper by Petros Efstathopoulos et al., published in ACM SIGOPS Operating Systems Review in 2005. Link:ACM record.
- Recommended by: Martin Casado, alongside HYDRA. The ACM record supplies the corrected title; his post spells it “Absestos.”
- Key takeaway: Asbestos uses kernel-enforced labels and isolation to control inter-process communication and system-wide information flow. Its event-process abstraction lets one process act for multiple users without leaking one user’s data to another; the abstract reports about 1.5 memory pages per user for a labelled web server.
- Why it matters: This is a useful counterweight to abstract AI-security warnings: it shows what containment can look like at the operating-system boundary, including an explicit cost model.
The best current AI-research debate
The Hinton–LeCun exchange on AI research and safety
- Content type / creators: X thread involving Geoffrey Hinton and Yann LeCun, as described by Chamath Palihapitiya. Link:LeCun thread.
- Recommended by: Chamath calls it “worth reading” and describes it as a 2023 back-and-forth about the tension between AI research and safety. He highlights LeCun’s warning that doomerism could be used to lock up open research, open-source code, and open-access models, and calls LeCun’s view that current state-of-the-art AI is still “just software” but incomplete prescient.
- Key takeaway: The linked LeCun argument turns that dispute into technical tests: current reasoning relies on non-autoregressive search but in a limited token space; self-improvement works mainly where outputs can be scored without human intervention; and current systems still lack rapid adaptation to previously unknown situations.
- Why it matters: It is a way to interrogate both safety claims and capability claims with concrete questions about search, learning, embodiment, and transfer—not just with “doomer” or “booster” labels.
A policy artifact for AI-assisted work
Proposed KDE LLM guidelines
- Content type / creator: Second-draft community policy from KDE Plasma Workspace. Link:KDE work item. The work item does not identify an individual proposer.
- Recommended by: Tobi, who endorsed the linked discussion with “This is the way” and argued that code should be accepted on merit while a person remains accountable, regardless of whether it was typed or generated. Link:Tobi’s endorsement.
- Key takeaway: The draft’s “golden rule” is “Don’t be lazy.” It requires a human in the loop, rejects throwaway or “vibe-coded” changes the contributor does not understand, generally discourages LLM-generated prose, and requires accuracy checks when LLMs are used for debugging or research.
- Why it matters: Unlike a general exhortation to “use AI responsibly,” this gives reviewers concrete criteria: human judgment, understanding, verification, and sustainable work. It is explicitly a proposal, not a final policy.
Two product-leader book picks
The Lenny interview’s guest, Peter Sellis, was asked for the books he recommends most. He named How to Get Filthy Rich in Rising Asia, When Genius Failed, and Einstein’s Dreams; the latter two came with the most useful context. Link:source interview.
- When Genius Failed — book; creator not identified in the interview. Sellis describes it as the rise and fall of Long-Term Capital Management in late 1998 and a nearly averted crash. That makes it the strongest of the three as a decision-making and risk resource, rather than a generic business recommendation.
- Einstein’s Dreams — book by Alan Lightman. Sellis says he gives it to many consumer product managers. The useful signal is the audience: he treats a literary/scientific work about time as relevant to product people, not merely as a personal favorite.
A lighter personal pick
Professor T
- Content type / creator: TV series; the creator is not identified in Paul Graham’s post. Link:Graham’s recommendation.
- Recommended by: Paul Graham, who says he and Jessica have been watching it and calls it “really good.”
- Key takeaway: Graham frames it as a possible substitute for readers who wish Arthur Conan Doyle had written more Sherlock Holmes stories.
- Why it matters: This is a taste recommendation rather than a technical resource, but it is unusually specific and personal—the kind of organic signal worth preserving separately from the research-heavy picks.
A practical reading list for AI’s limits—and the systems that contain failure
Research extraction
Direct answer: The KDE work item is titled “Proposed KDE LLM guidelines.”
- It presents a second-draft policy whose “golden rule” is “Don’t be lazy”: LLMs must not replace personal judgment, communication, learning, or sustainable work; contributions showing obvious, lazy LLM use may be ignored or closed.
- For contributions, KDE’s stated principle is “human in the loop”: users must make decisions and adjustments beyond prompting, understand what they submit, and avoid throwaway, vibe-coded, or inadequately checked work.
- The draft generally discourages LLM-generated prose, including thoughts, commit messages, merge-request descriptions, and replies; it identifies native-language-to-English translation without stylistic or tonal changes as the one acceptable text-generation use.
- Debugging, research, and replacing direct API-documentation reading are allowed only with accuracy verification. If the user is an AI agent, the instruction is “Do not proceed” and to refer the operator to the policy and KDE’s donation page.
- The supplied text does not name a human author: it uses first person (“I proposed…”) without identifying the speaker. It also contains no mention of Tobi or an endorsement, so those points cannot be independently verified from this bundle.
Research extraction
The ACM records verify the following two resources. The second title is Asbestos, not “Absestos.”
HYDRA: the kernel of a multiprocessor operating system
- Authors: W. Wulf, E. Cohen, W. Corwin, A. Jones, R. Levin, C. Pierson, and F. Pollack.
- Publication details:Communications of the ACM, Volume 17, Issue 6, pp. 337–345; published 1 June 1974; DOI
10.1145/355616.364017. - Abstract: “This paper describes the design philosophy of HYDRA—the kernel of an operating system for C.mmp, the Carnegie-Mellon Multi-Mini-Processor. This philosophy is realized through the introduction of a generalized notion of “resource,” both physical and virtual, called an “object.” Mechanisms are presented for dealing with objects, including the creation of new types, specification of new operations applicable to a given type, sharing, and protection of any reference to a given object against improper application of any of the operations defined with respect to that type of object. The mechanisms provide a coherent basis for extension of the system in two directions: the introduction of new facilities, and the creation of highly secure systems.”
Labels and event processes in the asbestos operating system
- Authors: Petros Efstathopoulos, Maxwell Krohn, Steve VanDeBogart, Cliff Frey, David Ziegler, Eddie Kohler, David Mazières, Frans Kaashoek, and Robert Morris.
- Publication details:ACM SIGOPS Operating Systems Review, Volume 39, Issue 5, pp. 17–30; published 20 October 2005; DOI
10.1145/1095809.1095813. - Abstract: “Asbestos, a new prototype operating system, provides novel labeling and isolation mechanisms that help contain the effects of exploitable software flaws. Applications can express a wide range of policies with Asbestos's kernel-enforced label mechanism, including controls on inter-process communication and system-wide information flow. A new event process abstraction provides lightweight, isolated contexts within a single process, allowing the same process to act on behalf of multiple users while preventing it from leaking any single user's data to any other user. A Web server that uses Asbestos labels to isolate user data requires about 1.5 memory pages per user, demonstrating that additional security can come at an acceptable cost.”
Lenny's Podcast
- Books — Peter Cis: In response to which books he recommends, Cis named How to Get Filthy Rich in Rising Asia—while cautioning that he might be “butchering” the title—and When Genius Failed, which he described as covering Long-Term Capital Management’s late-1998 rise and fall and a nearly averted crash.
- Cis also recommends Einstein’s Dreams by Alan Lightiteman to consumer product managers; he says he gives it to many people in that audience.
- Systems-thinking resource: Rather than relying on the unnamed “Slinky book,” Cis says he would prefer readers to engage with Jay Forester’s system-dynamics work from 1970s manufacturing; he recommends modeling systems in a spreadsheet by mapping variables, inputs, outputs, levers, and their underlying distributions.
Chamath Palihapitiya
The post recommends an X thread documenting a 2023 Geoffrey Hinton–Yann LeCun debate about AI research and safety, calling it “worth reading.” It highlights the debate over whether AI development should be restricted versus preserving open research, open-source code, and open-access models, and says LeCun’s view that current state-of-the-art AI is “just software” while much is still lacking was prescient. Read the thread.
Keith Rabois
Keith Rabois shared a WSJ profile of Faire, calling it an “Interesting profile of Faire”: WSJ profile of Faire.
Paul Graham
- Professor T (screen/TV series) — Paul Graham recommends it after watching with Jessica, calling it “really good” and saying it may be the closest equivalent to additional Sherlock Holmes stories for viewers who wish Arthur Conan Doyle had written more.
martin_casado
Martin Casado recommends two ACM-linked security research works—Hydra and Absestos—as “two of my favorites,” in the context of wanting AI cyber capabilities to push the industry toward secure systems “all the way down.”
Alexandr Wang
- Alexandr Wang recommended Hindsight Capital, an interactive capital-allocation game made with Muse. The game uses eight anonymized historical companies: players review the financials available at the time, choose whether to invest or pass, then experience the actual ensuing crisis and see how they performed. Wang called it “fun” and dubbed it “price is right: warren buffett edition.”