Modelling
A 2,400-seat VDI migration end-to-end — the full cost build-up, both the lift-and-shift and replatform paths, priced to the dual-run tail. The conclusion is already public; the underlying model goes public next, so the numbers can be argued with.
Building
An SA skills suite on Claude for scoping and effort estimation — the discipline that turns a padded first draft into a buffer you can defend line by line. Field notes from the first real use are here.
Benchmarking
Graviton4 against a Kafka-heavy workload. Early p99 looks good, but I won't publish a number until both sides are right-sized — anything else just credits one change for two.
Reading / chewing on
- How MAP funding mechanics shift when the migrate credit is modelled against ARR instead of project spend.
- Where a fixed-price SoW should stop and a T&M change-control should start — the trigger, not the argument.
This is a /now page in the nownownow sense — what has my attention at the moment, kept honest with a date.