Section
Thinking Machines
Thinking Machines tracks how artificial intelligence meets — or misses — the work of accelerated development
Thinking Machines
The Memory Supercycle, From the Buyer's Side
A 105 percent quarterly price jump, three suppliers sold out through 2027, a wafer crowd-out that turns 22 percent of capacity into 9 percent of output, and one small bank's 2027 technology budget written around all of it. What the file says about who pays for the AI buildout.
Xcelerator Systems
The Payment-Time Standard
Small-firm procurement only becomes development when invoices move on time: acceptance, approval, payment, subcontractor pass-through, and working-capital support are the hidden utility layer beneath every contract.
Xcelerator Systems
The Buyer Layer
Procurement is the missing accelerator layer: the public and private buyers that turn small-firm capability into revenue, records, jobs, supplier depth, and civic capacity.
Xcelerator Systems
The Accelerator Test
Startup hubs should be judged by whether they turn talent into jobs, suppliers, products, and public capacity, not by how polished the pitch day looks.
Thinking Machines
Agents for the Small-Firm Back Office
Bookkeeping, inventory, tax, grants, and customer follow-up are becoming the business infrastructure stack for firms too small to staff a modern back office.
Thinking Machines
West-Washing Chinese AI Dispersion on the eve of Trillion Dollar IPOs
As America's A.I. champions race towards the largest IPO in history, a New York Times article framing the competition between OpenAI and Anthropic maps the spike in Anthropic usage on an LLM routing platform, OpenRouter. This boring chart hides an open secret.
Thinking Machines
The Two Curves
Anthropic's revenue is bending upward faster than almost any company in history. We must choose a path that leads to accelerated socioeconomic development. On the distance between what the machines can do and who is using them.