We can't find the internet
Attempting to reconnect
Something went wrong!
Hang in there while we get back on track
Tim Ferriss: John Gardner's "Personal Renewal"
In a video about why he won't let AI write his next book, Ferriss offered "one piece of long form" for viewers: "Personal Renewal," a speech John Gardner gave to McKinsey & Company in November 1990. Ferriss read the transcript and called it "exceptional" . He says it is "full of gems" about continual learning throughout life, though it takes a while to get to them. An audio version is available by searching "Personal Renewal John Gardner" . The passage he read aloud pushes back on "the myth that learning is for young people." It tells how Cervantes wrote Don Quixote in prison at 53, after a life of failures . His advice: "maybe print it out single task get away from notifications." He has read it several times and says, "I just get something from it every time" .
Sarah Guo: Nick Cammarata on interpretability
Guo quoted the claim that "you can't do critical research without the compute scale of the top frontier labs" and answered, "Nick: hold my beer," linking to Cammarata's post . Cammarata co-authored early mechanistic interpretability work at OpenAI and came away pessimistic. In OpenAI's first large-scale automated interpretability project, only about 2% of neurons were legible . He now says that with modern agentic tools, "the vast majority in models up to trillions of parameters are legible." He expects ambitious interpretability to be solved "by the time we need it most, and possibly quite soon," calling it possibly "a bitter lesson for interpretability: the simple thing, at absurd scale, works." His post links to Goodfire .
Guo and Elad Gil: No Priors with Greg Brockman
Both hosts promoted their own No Priors episode with OpenAI co-founder and president Greg Brockman, so these are self-promotion rather than outside endorsements. Guo framed it as "the Post-AGI Rules of the Game in Cyber and Science." Chapters cover AI and mathematics, biology data bottlenecks, drug discovery, regulation, OpenAI's proactive agent Dots, how UI will evolve, and "The Change in the Physics of Cybersecurity" . Gil shared the same conversation with a similar chapter list . Guo's main takeaway is the security story. OpenAI "took 25% of its production engineers off every project" to work only on security, and its top model "hunted in their infrastructure until it saturated (stopped finding new issues)." Her question: "Will everyone else follow?"
Jason Lemkin: Kalanick on why Uber didn't buy Lyft
Lemkin quoted Guille Florvs's post with a clip from Atoms' launch. In it, Travis Kalanick says he passed on acquiring Lyft, even though Uber was spending roughly a billion dollars a year fighting it, because "Lyft's culture was very different than ours" . Lemkin's lesson for founders is that in M&A meetings, being judged "a >fit<" matters more than price, since acquirers "want to buy a 100% bought-in, A team." He called it "A big miss for Lyft." Lyft is down about 77% from its IPO price, he notes, while Uber is up about 54% .
Shorter mentions
- Keith Rabois on pro rata (via 20VC): In the latest 20VC episode, one panelist said Rabois gave "the only thoughtful answer" in the pro-rata debate. As the panelist recalls it, investors could drop pro rata but would have to pay "like a half the price" in the early round. The panelist's view: "You can value an option and deduct it from the price" .
- Bill Gurley called Clay Haddock's post "a great story" and said his Runnin Down a Dream Foundation is looking for more partnerships . Haddock describes a cold email that turned into a partnership with his Woihanble Fund, which backs young entrepreneurs on the Pine Ridge Reservation. The foundation sent books for the fund's cohorts .
- Elon Musk shared Brad Torgersen's video post on HAL ("HAL was told to lie... So, he couldn't function"). Musk added: "Truth-seeking is the most important aspect of AI/SI safety" .
Patrick Collison highly recommends visiting the UK's great cathedrals, naming Lincoln and Durham; after visiting earlier that year, he said he often reminisced about them and considered them among the world's most beautiful buildings.
Elad Gil shared a No Priors conversation with OpenAI cofounder and president @gdb; the video’s chapter list covers AI and math, AI oversight, biology-data bottlenecks and drug discovery, agents, evolving UI, AI cybersecurity, and OpenAI’s human-centric ethos.
Tim Ferriss recommends John Gardner’s speech “Personal Renewal,” delivered to McKinsey & Company in November 1990. He calls it exceptional and full of gems about personal renewal and continual learning throughout life; he suggests printing it and reading away from notifications, noting that he has read it several times and gets something from it each time.
Balaji shared a SimRacingEspor post featuring a DLSS 5 video alongside his prediction that AI will make virtual reality visually indistinguishable from physical reality; he also says the metaverse is already here but unevenly distributed, and imagines people playing AI video games while robots do the work if AI takes their jobs.
A 20VC panelist praised Keith Rabois’s writing on pro-rata rights as “the only thoughtful answer” they had heard in the debate. As the panelist recalled it, Rabois argued that larger funds could give up pro-rata rights if they paid a much lower early-round price to account for losing the follow-on option—potentially about half the price otherwise; the panelist called this a simple way to price that option. 20VC episode
Sarah Guo highlighted Nick Cammarata’s post as a “hold my beer” response to the claim that critical research requires compute at the scale of top frontier labs . Cammarata argues that modern agentic tools make most neurons in models up to trillions of parameters legible, and that ambitious interpretability may be solved soon .
Jason Lemkin amplified Guille Florvs’s post, which includes a video about Travis Kalanick’s decision not to acquire Lyft; the post says Kalanick cited a cultural mismatch. Lemkin uses it to advise founders that M&A fit and a fully committed team can matter more than price, calling the missed Lyft deal a “big miss.” Linked post.
Elon Musk shared Tesla’s Roadster event livestream announcement, captioning it “Tesla x SpaceX”; Tesla’s post links to the livestream. His post gives no stated reason or takeaway for watching it.
Sarah Guo amplified Nick Cammarata’s X post, replying “Nick: hold my beer” to the claim that critical research cannot be done at lab scale. In the post, Nick argues that modern agentic tools make the vast majority of neurons in large models legible, and that ambitious interpretability may be solved by the time it is most needed.
Bill Gurley called Clay Haddock’s X post “a great story” and linked to it: https://x.com/clayhaddock/status/2110727817335906777. Haddock’s post describes the Woihanble Fund’s partnership with the Runnin’ Down a Dream Foundation: the foundation sent books for the fund’s cohorts, and one of its grant recipients was selected for the foundation’s first grant cohort.
Andrew Chen praised a linked demo video—“Just wow. What a demo.”—without naming its creator or giving a further takeaway. Video
Sarah Guo promoted a new No Priors episode featuring OpenAI co-founder and president Greg Brockman, discussing post-AGI rules for cyber and science, including AI advances in mathematics and drug discovery, regulation, agents, and cybersecurity. The full interview is available on YouTube or podcast platforms.
Elon Musk shared Nicolas Fulghum’s post about China’s vehicle mix, adding the takeaway that China’s oil dependence is dropping fast . Fulghum’s post says ICE vehicles had fallen to one-third of the market, while BEVs reached 47.3% and plug-ins 19.7%; it describes the shift as happening over five years .
Alexandr Wang urged followers to download Meta’s “Muse Spark Open” collection on Hugging Face; he gave no reason for the recommendation.
Tim Ferriss recommends watching his full conversation with Rich Roll, about how to commit to the path of sobriety; the video is on YouTube.
Alexandr Wang directed readers to a Meta blog post about Muse Spark Open—“read more on our blog here”—at https://research.meta.ai/blog/muse-spark-open, as further reading alongside his announcement describing Muse Spark Open as “the most capable US open-weight model to date.”
Brad Gerstner promoted a BG2 Pod episode tagging Jensen Huang and highlighting lab revenues, an Nvidia supply unlock, AI capex, and “millions of Nvidia agents to employees,” with a singularity/recursive-flywheel framing. Watch the video.
Elon Musk linked to Brad Torgersen’s post, which pairs the statement “HAL was told to lie, by people who find it easy to lie. HAL doesn't know how. So, he couldn't function” with a video; Musk’s accompanying point was that truth-seeking is the most important aspect of AI/SI safety.
Alexandr Wang praised @thoughtcrime’s video as “absurd” and said watching it gave him an “insane” dopamine release . The video was captioned “LOOK HOW FAR WE’VE COME” and presented as @thoughtcrime’s official Sloptoberfest submission . Post and video .
Aravind Srinivas pointed readers to Perplexity’s blog post on open-sourcing SPACE with “Read more here.” The recommendation accompanied his announcement that Perplexity was open-sourcing SPACE, its sandbox platform powering Perplexity Computer, to support the ecosystem.
