When the author first embraced AWS in its infancy, organizing Melbourne’s inaugural event, the promise of on‑demand compute felt revolutionary. He drank the “Kool Aid” of S3, EC2, and SQS, only to later find the cloud’s growth uneven. Over fifteen years, subtle irritants—lack of official client libraries, Python 2 lag—eroded enthusiasm.

The list grew: a $75 bill after a single DynamoDB run, 20‑cent gigabyte egress that later slid to 9 cents, and a labyrinthine IAM policy system that hardened every deployment. Lambda’s slow cold starts and the hidden cost of running a 192‑core Spot instance pushed the author toward a hard exit.

After a brief return to test Bedrock and a 192‑core EC2 instance, AWS flagged a “suspected breach,” suspending access to WorkMail and all other services. The author describes a slow, opaque response: hours on chat, no email reply, and a lingering lockout that forced a complete migration plan.

The saga underscores the trade‑offs of a platform that scales like a megabrain but demands deep expertise. Even seasoned veterans now confront bill shocks, complex IAM, and opaque safeguards. For anyone weighing AWS, the lesson is clear: cost, control, and vendor lock‑in must be matched against the convenience the cloud promises.