ZeroNoise Logo zeronoise
Post
Bostrom’s *Superintelligence* returns as a high-context AI-risk recommendation
20 hours ago
2 min read
134 docs
Elon Musk’s resurfaced endorsement and Martin Casado’s historical framing make Bostrom’s *Superintelligence* the period’s strongest reading lead; Patrick Collison adds a terse but unusually emphatic pointer to Ruskin’s *Nature of Gothic*.

Most compelling: Superintelligence

  • Content type: Book · Creator: Bostrom · Link:Musk’s original recommendation.
  • Recommended by: Elon Musk, whose current re-share points back to the recommendation, and Martin Casado, who added the historical framing.
  • Key takeaway: Musk’s endorsement is unusually direct: “Worth reading,” followed by a call to be “super careful with AI” because it is “potentially more dangerous than nukes.” Casado calls it “the book that framed people’s mental picture of the technology, three years before transformers was published” and “the legacy of Bostrom.”
  • Why it matters: This is the strongest high-context reading lead in the set: one recommender supplies a concrete reason to read it—AI-risk caution—while the other presents it as a pre-Transformer reference point for how the technology was understood.

A terse but high-confidence second pick: Nature of Gothic

  • Content type: Book/essay · Creator: Ruskin · Link: No direct work URL is supplied; Collison’s recommendation.
  • Recommended by: Patrick Collison.
  • Key takeaway: Collison says it is “the best thing I’ve read so far this year.” The post gives no summary of the work’s argument or lesson.
  • Why it matters: The value here is the strength and specificity of the endorsement, not an extracted framework. Treat it as a reading lead to investigate directly rather than as a claim that Collison’s post has already explained the book.
Bostrom’s *Superintelligence* returns as a high-context AI-risk recommendation
Research extraction

Assessment: The article is substantive news reporting for an authenticity check, not a promotional resource. It reports Apple's official statement that Telegram was briefly removed for CSAM-availability and restored after Telegram removed content and banned the poster , and it adds prior-removal context ; a separate Amazon product-link block is promotional but peripheral . No author byline or Tim Sweeney commentary is present in the supplied lines .

  • Title: "Telegram briefly pulled from the App Store over child sexual abuse material availability" (from bundle metadata).
  • Author: not identified — the article text contains no byline .
  • Factual subject: Apple briefly removed Telegram from the App Store after "our review found content that violates our strict guidelines prohibiting child sexual abuse material," then restored it after the developer "promptly removed the content and banned the user who posted it."
  • Key takeaway: The removal was brief and CSAM-related per Apple; Telegram had not commented on the actual reason and had only posted "Reports of my demise are greatly exaggerated 🍎," while the original report had characterized the removal as apparent and unexplained.
  • Reporting vs commentary: The body is reported news — Apple's statement, 9to5Mac's outreach to Apple, and historical removals in 2018, Brazil in 2023, and China in 2024 . The supplied bundle contains no Tim Sweeney commentary .
  • Limitations: No byline; the original claim is hedged ("appears to have been removed," "for some users") ; Apple's statement came about 20 minutes after restoration ; Telegram itself had not explained the reason .
Telegram briefly pulled from the App Store over child sexual abuse material availability
Research extraction

Assessment: Substantive argument with promotional framing — a useful learning resource on open-source contribution models, but not non-promotional. The post argues that agent-built PRs shift judgment costs to maintainers and that valuable outside contribution moves to issues and design discussion (), then announces the goose project's adoption of that model ().

  • Title: Moving to issues as the new PRs ()
  • Author: None stated in the post; bundle metadata lists nil.
  • Main argument: Coding agents make plausible feature implementations cheap, but decisions about whether and how changes belong remain expensive; the maintainer ends up supplying the judgment, so open source should shift contribution from writing code to filing and discussing issues (, ).
  • Promotional flag: This is a project announcement on goose-docs.ai using first-person framing ('We're moving goose to this model') and linking to goose governance, contributing guidelines, and an issue board; it is advocacy for goose, not neutral analysis ().
  • Educational takeaway: The unit of contribution is 'taking a problem all the way to a verified solution'; the valuable place to contribute is good issues and the discussion that turns them into well-considered solutions, with contributors at each stage recognized as co-authors (, ).
Moving to issues as the new PRs | goose | Your open source AI agent
Eric Schmidt

Eric Schmidt endorsed Asari AI Labs' post and the video it links to, which describes self-improving agents optimizing the full vLLM inference stack, with up to 16% more throughput and interactivity for DeepSeek v4 Pro and GLM 5.2 on B200s (no MTP) . He wrote: "This Asari approach works... The power of agents that are incredibly smart," and highlighted potential reductions in data center serving costs . Asari reports every change was verified and its agents improved with each iteration . Video: Asari AI Labs video. Schmidt's post also links to the original Asari post .

Our self-improving agents optimized the full [@vLLM_project](https://x.com/vLLM_project) inference stack, with up to 16% more throughput … This Asari approach works, imagine the improvement in inference and the reduction of data center serving costs when people adopt this! Th…
Patrick Collison
  • John Ruskin, The Nature of Gothic (book/essay): Patrick Collison called it "the best thing I've read so far this year" in an aside to his reflections on aesthetics .
  • Solzhenitsyn's Nobel Lecture (lecture): Collison cited it in his aesthetics post, quoting "a true work of art carries its verification within itself" and noting beauty may serve as "a marker of deeper metaphysical coherence" .
As an aside and related to these topics: Ruskin (Nature of Gothic) is the best thing I've read so far this year. [https://x.com/patrickc/… Someone asked me recently why I’ve become interested in aesthetics after having spent most of my life more interested in STEM-adjacent to…
martin_casado

Martin Casado endorsed @newstart_2024's post with "Worth it." The post argues there is no safe level of alcohol, claiming every drink moves you closer to death a little faster and zero is the only safe number, citing the largest Lancet study ever done via Chris Williamson; it treats drinking as a known trade-off like eating a cheeseburger and praises the framing for "Clear data. Honest cost-benefit. No moralizing." The post includes a video.

Worth it. [https://x.com/newstart_2024/status/2083070062206779470](https://x.com/newstart_2024/status/2083070062206779470) There is no safe level of alcohol. Chris Williamson cited the largest Lancet study ever done on the subject. Every single drink moves you…
Tim Sweeney

Tim Sweeney shared with endorsement a 9to5mac report by @mvcmendes that Telegram was apparently pulled from the App Store worldwide (URL), calling Apple's move "business as usual" and arguing the App Store "desperately needs competition worldwide" .

Apple's App Store desperately needs competition worldwide. Apple capriciously removes a vital communication and privacy tool from availab… Telegram appears to have been pulled from the App Store worldwide [https://9to5mac.com/2026/08/03/telegram-appears-to-have-been-pulled-fr…
martin_casado

Elon Musk recommended Bostrom's Superintelligence, writing it's "worth reading" and warning AI is "potentially more dangerous than nukes" . Martin Casado echoed the book's impact, calling it "the book that framed people's mental picture of the technology, three years before transformers was published" and "the legacy of Bostrom," linking to Musk's original tweet .

Worth reading Superintelligence by Bostrom. We need to be super careful with AI. Potentially more dangerous than nukes. The book that framed people's mental picture of the technology, three years before transformers was published. The legacy of Bostrom. [ht…
Elon Musk

Elon Musk (re)shared his endorsement of Nick Bostrom's book Superintelligence, calling it "Worth reading" and warning that AI is "potentially more dangerous than nukes" . The recommendation is authentic, made in the context of urging caution with AI development.

Worth reading Superintelligence by Bostrom. We need to be super careful with AI. Potentially more dangerous than nukes. As I was saying … [https://x.com/elonmusk/status/495759307346952192](https://x.com/elonmusk/status/495759307346952192)