We discovered a bug in AWS Bedrock that is double counting cache writes when thinking/reasoning is enabled for the Anthropic models. It’s not clear to me if this is limited to just AWS Bedrock or all providers. AWS Support is aware.
We’ve also observed a much higher cache miss rate in the past few weeks. Combine both together and your usage consumption can be greatly increased.
This is the way. As others mention opening window is your cheapest option, but you lose a lot of your heating and cooling efficiency. We installed a whole house ERV for about $10K and saw a significant drop in CO2 readings.
I wonder if there is a case to be made for a patch branch that gets all PRs. Then the maintainer can do any further tweaks needed before merging down to mainline.
Another way to look at it is you are expected to include business factors into your decision making. Sometimes the most optimal engineering solution is not the most optimal business solution. Politics can definitely be a factor, but being able explain clearly why doing something helps the business can be a big part of making the jump to more senior roles.
Have you considered looking at Amazon for their ARM offering (Graviton)? I'd be hesistant to use M1 minis for a production workflow as they are not really production grade (lacking ECC memory, not sure how long they are rated to run at high CPU, lack of user replaceable disks, no RAID, etc...).
It sounds like you just don’t enjoy frontend work. If you are trying to build a product on your own it’s impossible to avoid, but there are plenty of companies in the EU and US hiring for backend developers. Understanding how both sides get work done is important, but there is nothing wrong with focusing on just one area.
We’ve also observed a much higher cache miss rate in the past few weeks. Combine both together and your usage consumption can be greatly increased.