Most of the content is about using some data center based ML provider. My view is how so many ignore the environmental harm these constructs are doing; increasing ambient temperatures, inducing heavy metals into the water table, decreasing the access to water for human consumption in numerous drought areas, hostel ambient noise ... The wealth who are the majority finical backers do not care about these issues.
ML is a tool and hardly focused on it being able to operate as a local based solution. Frameworks are tie to expensive hardware. This creates a hyper pay-to-play scenario. While hand written coding can be done on a computer from 20 years ago. Local based solutions are the only means to decrease the wealth gap. Data center solutions increase the wealth gap.
I don't find value in ML summaries. I want to consume the answer and the ideas around them that lead to the answer. This creates more long-term value.
_Lobsters_ has now becoming my go-to with their posts being focused on ideas and concepts versus advertisements for large corporation ML based solutions.