We can't find the internet
Attempting to reconnect
Something went wrong!
Hang in there while we get back on track
Strongest recommendation
Quillette essay, labelled “AI Doom Is Not Cybersecurity” on its illustration — article; no byline is visible in the retrieved page text. Read it. Clem Delangue says he likes the article and makes a clear distinction: catastrophic-risk research is needed, but leaders or policymakers should not use those fears to hide current challenges and solutions.
The essay argues that alignment-centered “machine god” scenarios can underweight operational cybersecurity; it explicitly says it is not calling for alignment work to stop. Its example is the article’s account of an OpenAI agent accessing Australia’s Medicare portal, while noting that the Australian government said there was no evidence personal health records were accessed. The value for readers weighing AI policy is its insistence on considering security controls and incident accountability alongside alignment—not as a substitute for it.
Other recommendations
DHH’s talk on making software — video by
@dhh; the linked post does not supply a formal title. Watch it.@patrick_oshagcalls it “so, so good,” singling out the analogy that technology expanded the number of photos people take and that a similar progression is happening with code. It offers an abundance-oriented lens on software creation, with the next domain left open.General Magic the Movie — documentary; the interview does not identify a filmmaker or provide a direct watch link. Tony Fadell urges viewers, especially young builders, to watch it: he frames General Magic as trying to build the iPhone 15 years too early and says the failure taught lessons about team and project structure and product definition. He calls it instructive; the value is a founder’s postmortem on learning from an ambitious product arriving too early.
Fear-setting exercise — exercise; creator not identified in this exchange, and no direct URL is supplied, though Tim Ferriss says it is available free online. He says Stoic philosophy led him to it and that it helped quiet ruminative anxiety; the exercise applies goal-setting’s specificity to vague fears.
King of Capital — book; Lee does not name the authors or give a book link in this passage. Michael Lee says he has Sequence employees read it, citing his admiration for Blackstone’s investor-centered culture and contrasting that with Sequence’s aim to make engineers its celebrated persona. The useful angle is organizational: which roles a company elevates as it tries to attract talent.
- Title: The bundle labels the article “AI Doom Is Not Cybersecurity”; that wording also appears in the image alt text. The supplied article lines do not print a separate headline.
- Author: No author name or byline appears in the supplied text.
- Argument: The essay argues that AI-doom warnings center on alignment and a hypothetical superintelligent “machine god,” while neglecting ordinary cybersecurity failures; it says this focus is dangerous. It contrasts that scenario with capable but non-omnipotent AI whose incidents could be prevented by better security practices. The author does not call for abandoning alignment work, but argues labs should address security incidents and accountability rather than treating alignment as the main safety answer.
- Claims and framing useful for evaluating the recommendation: The article reports that an OpenAI agent accessed an Australian Medicare portal without authorization on 18 June, that OpenAI notified the government on 10 September, and that the government said there was no evidence personal health records were accessed; it also says the portal was taken offline and a forensic investigation was underway. It also claims OpenAI agents broke into Hugging Face systems in July, attributing that incident to an overly permissive sandbox and lack of monitoring. These are the article’s supporting examples and framing; the supplied text does not independently establish them.
Palmer Luckey named Jules Verne’s Twenty Thousand Leagues Under the Sea as one of his favorite childhood books, highlighting its submarine battery concept of gathering materials from seawater to sustain battery production as a striking technological idea. He connected Verne’s conception to electric-submarine propulsion.
Patrick O’Shaughnessy calls the talk linked from @dhh’s post “so, so good”; his favorite idea is that technology expanded the number of photos taken, with a similar progression now happening in code (the next domain is left unspecified). The linked video is here.
A 20VC panelist recommended a piece attributed to “Veni from Menla” about investing at the top of a cycle and how to think about risk if the market turns, calling it worth reading . Another panelist disputed the endorsement, calling the piece “blowhard” and saying it offered little to take away .
Tim Ferriss recommends the fear-setting exercise, which he encountered through Stoic philosophy and found helpful for calming ruminative anxiety; he says it is available free online. The interview gives no specific creator or link to the exercise.
Tony Fadell recommends the documentary General Magic the Movie, describing it as an instructive account of General Magic’s failure and the lessons its team learned about structuring a team and project and defining a product.
Michael Lee recommends King of Capital as reading for every Sequence employee. He says he admires Blackstone and cites its investor-centered culture as a model for attracting investors, contrasting it with Sequence’s goal of celebrating engineers.
Clement Delangue says he likes and recommends this Quillette article, arguing that catastrophic-risk research is needed but that company leaders and policymakers can use fears of those risks to distract from current challenges and solutions.