Page 25 of 505 ยท Back to latest writing
As AI agents take over more of the inner loop of writing and testing code, engineers must own the outer loop: defining checks, judging outputs, and remaining accountable for what reaches production.
Bob Hoffman developed the Hoffman Process as an intensive retreat for unlearning destructive childhood patterns. His work centers on Negative Love: how children absorb painful parental traits to secure affection, and how adults can unwind those inherited behaviors.
Enterprises risk giving model providers their most valuable knowledge through corrections, context, and prompts. Durable advantage requires private evaluation systems, clear trust boundaries, and ownership of the learning loops that turn activity into institutional memory.
Jeremy Allaire argues that AI agents need economic infrastructure for payments, identity, and coordination. This explainer separates the functions that genuinely benefit from blockchains from those better handled by conventional systems, clarifying where onchain architecture earns its complexity.
Jon Taffer is a hospitality operator and longtime host of Bar Rescue. His operating philosophy centers on reaction management: deliberately shaping how customers, employees, and owners feel inside a business system to improve accountability, retention, and turnaround outcomes.
Jaya Gupta is a partner at Foundation Capital focused on enterprise software and AI infrastructure. Drawing on startup operating experience, she backs technical founders building the picks-and-shovels systems that reshape how companies adopt and use AI.
Thinking Machines is betting that useful AI will be customizable, specialized, and owner-operated. The argument moves beyond centralized general models toward systems that extend local knowledge and human judgment instead of replacing them.
OpenRouter's 100 trillion token usage study suggests AI demand is shifting from simple text generation toward reasoning, tools, code, and context-heavy workflows. This explainer maps what that change means for model providers, infrastructure, and the economics of serving AI.