Most AI Work Can Wait

kept by eddie jul 02

Routing layer design, not model choice, drives cost efficiency by sending 70-80% of AI traffic to cheap local or async models.

Today I've made the difficult decision to reduce the size of Coinbase by ~14%
Coinbase CEO Brian Armstrong announced a 14% layoff, attributing the cuts to AI-driven productivity gains and a need for leaner operations.
also on Brian Armstrong, Coinbase, #developer-tools
This is an email I sent earlier today to all employees at Coinbase: Team, Today I’ve made the difficult decision to reduce the size of Coi
Coinbase CEO Brian Armstrong lays off 14% of staff, citing crypto market downturn and AI productivity gains, restructuring with no pure managers and AI-native pods.
also on Brian Armstrong, Coinbase
Is AI Doing Less & Less?
A six-month evolution from fully agentic AI to a hybrid architecture shows 65% of workflow nodes now run as deterministic code, improving reliability and cost efficiency.
also on Tomasz Tunguz, #developer-tools
never waste a token
Putting a durable buffer between AI agents and LLM providers prevents token waste and re-billing when a process crashes mid-stream, with resumable streaming and crash recovery using the same mechanism.
also on #cost-optimization, #ai-coding, #software-engineering, #developer-tools
The Pulse: a new trend, smart model routing - The Pragmatic Engineer
Several vendors now offer intelligent model routers that automatically select the cheapest sufficient LLM per task, promising 20-30% cost savings, a trend expected to become table stakes.
also on #cost-optimization, #ai-coding, #developer-tools