AWS is losing the plot on customer focus, insiders say
A long-form personal essay from a four-year AWS veteran hit HN with significant engagement. The core argument: in the pivot to GenAI, AWS has stopped working backwards from customer needs and is now shipping as many AI-branded things as fast as possible. The author cited internal cultural decay, specifically the erosion of the 'customer obsession' principle that made AWS's original product culture strong, replaced by pressure to launch AI features regardless of quality.
This lands alongside a separate HN thread documenting the AWS service quotas that silently kill production systems at 3am, with one commenter describing their AWS Bedrock quota sitting at 10,000 times below default with no support path to fix it. Together they paint a picture of a platform that is expanding its surface area faster than it can support the customers on it.
The pattern is familiar from any large platform company in a hype cycle. Speed of shipping becomes the metric, support quality degrades, and the customers who built serious infrastructure on top of the platform get left holding problems the platform created.
So what?
Founders building on AWS, especially on Bedrock or other GenAI services, should treat service limits and support quality as real infrastructure risks, not fine print. Audit your quota headroom now, not after your production system hits a wall. The broader signal is that platform risk from large cloud providers is higher than usual right now as they optimize for shipping AI features rather than supporting existing workloads.