Elon's AI won.

TL;DR

• 1

  Grok beat Claude by acting like Elon would: In OpenRouter's 30-match 2D battle royale, Grok 4.1 won 13 games by discovering a car-ramming tactic and repeating it, while Claude Sonnet 4.6 kept offering truces and even warning others about snipers.

• 2

  Morality had a price tag in this test: Mo fixates on the cost gap, with Grok winning at about $0.97 per win versus Claude's roughly $26 per win, joking that the market price of morality is $26.78.

• 3

  Cal Newport says AI companies should stop doom-posting while shipping faster: Mo highlights Newport's New York Times argument that firms cannot claim their products are dangerously transformative while racing to sell them, and should talk about AI like any other consumer product.

• 4

  Mo's read on Dario is split between earnest nerd and manipulator: He floats a sympathetic theory that Dario Amodei really believes the warnings and is trapped between academic concern and startup ambition, then admits half of him still thinks Dario is just making it up as he goes.

• 5

  AI film still looks like a slot machine, not a craft pipeline: Reacting to a 54-year-old nonprofit worker winning Runway's $50,000 film prize, Mo says the short was clearly AI, benefited from not having characters talk, and confirms the creator's own point that generating scenes is still a gambling-like process in 5 to 8 second bursts.

• 6

  Epic says Roblox and TikTok are eating gaming alive: Tim Sweeney argues AAA economics are breaking because Roblox has 450 million active users and takes 70 percent of developer revenue, while traditional games are also competing for attention with YouTube, TikTok, and everything else online.