Unit Elastic: Cheaper AI Tokens Stopped Growing AI Spend
Since August, Ramp's data shows AI token volume up 65% and token spend up 1.5%. Why unit-elastic demand is a problem for the AI capex math.
Recent posts
View all posts →Latest posts on AI engineering, model development, and what we're learning
Found Five Times, Fixed Once: What PHP2Shell Says About AI Bug Hunting
PHP2Shell was flagged in 2004, 2020 and 2024, then proven twice with AI in three weeks. The fix took 71 days, and almost nobody has installed it.
Open Weights, Datacenter Bill: What It Costs to Run Kimi K3.1 Yourself
Kimi K3.1's weights land by Nov 4. Four Mac Studios lose to its $2/$10 API by about 2x, and a rented 8x B300 node needs ~2,100 tokens a second to win.