Martin Casado: Liebowitz & Margolis on network effects

Casado wrote a short intellectual history of "network effect." It runs from Rohlfs's 1974 Bell Labs model of communications networks, to Katz and Shapiro's 1985 generalization (which he finds "largely tautological and useless"), to Rochet & Tirole's 2000 work, which split out two-sided marketplaces as their own category. The paper he recommends is the 1994 one by Liebowitz & Margolis. In his words, it shows that Katz and Shapiro's work "is sorta braindead" and starts "to classify things better (this is a great paper, I recommend reading it)." @martin_casado His point is that the business press picked up the idea at different stages, which is why there are so many working definitions: "the business jargon tower of babel." @martin_casado

Bill Gurley replied that you "can't leave out Brian Arthur (and Cormac McCarthy)," and linked to a further source (link). @bgurley

Tim Ferriss: books and essays for being "the only"

On a recent YouTube conversation, Ferriss recommended several resources:

• Kevin Kelly, "1,000 True Fans" (at kk.org). He says it "was true 10 years ago and it is going to become ultra ultra true as AI starts to gobble everything." He paired it with the "law of categories" chapter from The 22 Immutable Laws of Marketing, which he says "had a huge impact," and with Blue Ocean Strategy. All three back his advice to aim to be "the only," not the best. Jay Shetty Podcast
• Peter Drucker, The Effective Executive. "Everybody should read that. Holy cow, what a great book." He took from it Drucker's distinction between effectiveness (doing the right things) and efficiency (doing things right). Jay Shetty Podcast
• Oliver Burkeman, 4,000 Weeks. He called it "phenomenal" and singled out its chapter "cosmic insignificance therapy," which he says makes problems "seem less life or death by zooming out." Jay Shetty Podcast
• Seneca, Letters from a Stoic (Penguin Classics). He says he has "probably given away 100 copies." Jay Shetty Podcast
• He also pointed to his own PDF of 17 reflection questions at tim.blog/17, which he says he still uses "all the time." Jay Shetty Podcast

Patrick O'Shaughnessy: the Colossus profile of Larry David

Yesterday's brief mentioned a teaser that didn't name its subject. The Colossus profile is now out, and it is about Larry David. O'Shaughnessy's endorsement was a Larry David line: "Pretty, pretty good." @patrick_oshag The author is Blake Harris, who spent six years talking to David for his book Emperor of Nothing: The Making of Larry David. Colossus presents David as "one of history's purest founders" and explains "why it took him so long to break through." @colossusmag Colossus is O'Shaughnessy's own outlet.

Satya Nadella: the papers behind Microsoft's AI bet

In a talk on security and superintelligence, Nadella said he still has Dario Amodei's paper "Scaling Laws and the Path to AGI." He says it got Microsoft to see emergent capabilities from scaling, and in 2019 to decide, "Let's go all in." Microsoft Security He then called "Jacob's paper" "a good one," quoting its line "We are not building AI, we are growing AI." He gave no full name or title. He tied it to his view that "alignment may be elusive, even while scaling works." Microsoft Security

Sameer Dholakia (Bessemer): Machines of Love and Grace

On 20VC, Dholakia endorsed Dario Amodei's thesis in Machines of Love and Grace. He expects automatable wet labs plus AI reasoning over large datasets to bring "in the next decade what would have taken us half a century" in drug discovery: "I really believe that." 20VC with Harry Stebbings

Shorter mention

Tim Sweeney called Jan Orszulik's thread "amazing." @TimSweeneyEpic Orszulik demos a game-dev tool he built as an example of using AI "to achieve YOUR vision while maintain[ing] absolute control over the result." @JanOrszulik