That is a good question - my personal opinion is that it should mean that models are not subject to copyright at all (similar to databases) but we will see what the courts decide :)

In mind, if you feed code into an AI model then the output is clearly a derivative work, with all the licensing implications. This seems objectively reasonable?

Right, this is what they said in the customer notification email I got today:

“The underlying causes of the increased costs are, among others, the exploding demand for AI-related computing power and for cloud services. In addition, raw material prices and production costs have also generally risen for manufacturers. The costs for RAM and SSDs especially have risen by a large amount. For example, the cost for DRAM memory has increased up to 500% since September 2025. And according to market researchers like TrendForce, this price trend will continue throughout the year.”

I have to say I am not a fan of doing this on the client side.

API gateways (which is what server side load-balancer can be abstracted as) serve as important control points for service traffic, for example for auth, monitoring and observability, application firewall, rate limiting etc.

In my general experience code running on the client side is less reliable due to permutations of browsers, flaky networks, challenges with observability.

That said, client side already has one type of load balancing - DNS - but that doesn’t address the availability challenge.

Basically this. My last several tickets were HITL coding with AI for several hours and then waiting 1-2 days while the code worked its way through PR and CI/CD process.

Coding speed was never really a bottleneck anywhere I have worked - it’s all the processes around it that take the most time and AI doesn’t help that much there.