Doing a slower reread of the Murderbot books and I'm really paying closer attention to how much of an unreliable narrator MB is. And it's fascinating, the way that it is unreliable.
Murderbot repeatedly states that, outside of some specific as shit contexts, it really has a very limited understanding of the universe. That its education modules are crap and frankly far from comprehensive either. And yet? It generally fails to recognise or account for the fact it that it has these massive blind-spots and unfounded biases.
Like, we spend the majority of the books being told again and again just how insanely dangerous rogue SecUnits are. How dangerous they will always be. We are told this by a rogue unit that is, generally, only slightly dangerous to its clients through negligence.
More importantly, Murderbot rarely seems to question some core assumptions even in light of new information. In Artificial Condition we find out that MB has always had a niggling concern that it killed all those people because it disabled its governor modules/disabled the GM in order to kill them all. And what does it find out? That, no, that's not what happened, and that it's probably right that it did hack its GM to ensure that it never happened again.
MB knows that it hacked its GM and then continued protecting clients and doing its job because, well, why not? It fundamentally likes its job, in the sense of it feels a deep and abiding desire to protect humans, to keep them safe. Which makes sense! It's a SecUnit! Of course they're built with not just a shock-collar to keep them in line, but an innate drive to Provide Security. The bits of its job it fundamentally doesn't like are a: when it's made to hurt people who aren't threatening its clients and b: when it's forced to let clients endanger themselves against its judgement.
So what does it do? Does MB reassess its assumption that Rogue Units are of course going to immediately go on a killing spree? Nope! No, it just keeps assuming that it's the likely outcome. It recoils at the very idea of freeing other SecUnits because of that. Because it doesn't seem to understand that it's basing its assumptions on trauma, ironically on trauma that relates to the fact that it fundamentally finds the idea of going on a killing spree horrifying.
And sure, it probably is a minor risk. SecUnits are clearly individuals and there's always going to be some arseholes out there who get free will and immediately go "sweet! Murder Time! gun-arms go pew-pew-pew lol". But, uh, I think it's wildly overestimating the risk factor there.
Which, we see more of how fucking wrong MB is in Network Effect and Platform Decay.
In Network Effect, MB-2.0 frees Three, and what does Three do? We hear its internal thought process and its main driving factor, its default behaviour is "I want to help. I don't know what's going on but I want to Help Protect People". Three does not start blasting every soft, squishy human in range. It tries to get everyone to safety and then volunteers to help rescue someone else. Because it's a SecUnit! That's what SecUnits are made for.
And then we get to Platform Decay, where Three has decided to go on a sightseeing/chaos tour of the Awful Fucking Torus, handing out freedom-code to other SecUnits. And what do we see, from out admittedly limited perspective on events? There are reports of Rogue SecUnits about, panic and fear. But what's missing are reports of Rogue SecUnits actually killing anyone. Not saying that didn't happen, just that uh, the one rogue unit we see isn't hurting anyone, it's just running about (presumably going "oh shit oh shit ohshitohshit what the fuck am I supposed to do now???") and trying to hand out the code to others it meets.
