Loading activity…
@stevehou
Head of Research @Silicon_Data. @UMICH/@ETH_en/@UVA alum. Perpetually curious, but also "incredibly unsophisticated" (according to Chamath Palihapitiya)
AI
Anthropic reports 17 of 211 lab-tested designs met binding criteria, but none were tested for toxin neutralization.
AI
A quote-post jokes that there’s finally a use for the empty office buildings its author expects to see.
AI
An essay argues that falling prices and surging token use favor platforms that control distribution over benchmark-leading labs.
AI
A user says ChatGPT proposed space outside the building, while Grok mistook a 10-by-10-foot area for one measuring 10 by 5 feet 4 inches. Claude Opus 5.5, they say, produced a floor plan that respected the dimensions and later made a 3D rendering.
AI
DealBook says the round came almost exactly a month after Instinct raised $250 million at a $2.5 billion valuation.
AI
SpaceXAI calls it an improvement over Grok 4.6 at the same price and speed. Artificial Analysis reports better coding results, but more than double the output-token use in a separate evaluation.
AI
A user argues that a difficult proof’s value can lie in the reasoning and insights behind it—not just the result—so a machine-generated solution may not be as valuable.
AI
Steve Hou questions how advocates should rebut claims tying AI infrastructure to inflation, higher rates, and job losses.
