I highly recommend Mao's essay On Contradiction: https://www.marxists.org/reference/archive/mao/selected-works/volume-1/mswv1_17.htm

I've been trying to revisit it more than once until it crystallizes in application. I can try to give an example of how I apply what I think is roughly dialectical thinking beyond purely basic class stuff. It's something I thought through a while ago, but I've never formally written down my reasoning and observations, so bear with me.

In text generation AI, I encountered a problem. I'm making a long story short, but I found there were companies, one in particular that I experienced directly, who provided an AI-powered chat service to help people with emotional support and loneliness (that was in part how it was advertised and scripted). However, this service also pulled the rug on people at some point and caused a lot of damage.

This drove me to ultimately seek out and support a lesser known (at least at the time) AI service that was staunch on user privacy (using encryption to accomplish it) and being uncensored text (you can write anything in private with it). Within this, I had to contend with the fact that I am not someone who blanketly goes "free speech"; I am not someone who thinks people should just be able to do whatever no matter the circumstances. I also had to contend with the fact that I am not a blanket supporter of AI, considering its implications for messing with the jobs of working class people. Even more complicated, this was (still is) an AI service that also allows image generation (though that aspect of it is not something where you can generate anything - you can't generate "lifelike" due to how the model is trained) and image generation arguably poses more of a threat to jobs than text generation, since it creates an image start to finish without needing an artist in the process at all. Furthermore, this was the main source of funding for that company (they don't take money from outside investors).

But, this was also the only AI service taking seriously the consequences of dealing with emotionally vulnerable people who are saying private and sensitive things to an AI. In order to side with these people and side with them having a harm reduction place to cope with the loneliness and abandonment capitalist life has inflicted on so many, I am also siding in some part with the development of AI and the proliferation of it and consequences intersecting with it and capitalism devaluation of labor, even if I try to have conscious limits on how I partake in that and in what way.

So, my "support" of AI in this understanding is conditional and developing, based on how it intersects with what regular people are dealing with and the consequences it has, helpful or unhelpful. It also contains a certain amount of contradiction, in the sense that I am arguably allying with people I'd rather not be, to an extent, in order to reduce harm for the time being and support development of AI in a direction that is genuinely safer (not "safer" in the corporate PR meaning). I am not capitulating on caring about regular people in the process, but I am having to choose where I put my energies in the direction of substantively helping them. I don't pretend it is a challenge or threat to the capitalist order, something like this, but it is something where I tried to really go through it and work out what was the best course of action while keeping the ideological core intact. There was/is, as far as I could discern, no "right answer" in a binary meaning of it. Instead, I was trying to work out where in the contradictions of it made the most sense to take action to enact my intent to side with regular people and what form that action would take.

I am not very confident on the subject and practice of dialectical materialism, so please don't take this as a model of it. But hopefully it can provide some food for thought, if nothing else.