Interesting observation. Spoilers -> He does the same thing in Tower of Babel, where the topology of the universe is structured in such a manner that the tower can physically reach "heaven", which ends up being a surprise to the reader and the characters at the same time. Masterful stuff.

I want to nitpick two things.

On compatibalism, the first definition presented is the correct one, the framing that "you have to make peace with determinism" isn't quite right. For compatiablists, determinism is freedom, because if one's actions did not follow from prior causes then they would not align with one's internal states.

The other is sneaking in the characterization of Chiang's AI doomer skepticism as a "blindspot". This topic is being debated to death on HN every day so I'll leave that argument for another thread, but IMO it contradicts the tone of the article about a writer whose depth of thought the author was just heaping praise on. I'm not saying its necessary to adopt his views on all things, but I think it deserved more than a footnote dismissal.

▲

LinchZhang 3 days ago | parent | next [-]

I appreciate the nitpicks!

Re #1 It's been several years since I read up on that area of philosophy. I'll need to reread some stuff to decide whether I think the definition I used is a fine enough simplification for sci-fi readers (and, well, myself) vs whether it missed enough nuances that it's essentially misleading.

(Some academic philosophers follow me on substack so maybe they'll also end up correcting me at some point!)

Re #2 ah I don't think of it as "sneaking in". It's more like "this is a view I have, this is a view many of my readers likely also have, given that this is a widely debated topic (as you say) and I'm not going to change anybody's minds on the object level I'm just going to mention it and move on."

▲

the_af 3 days ago | parent | next [-]

For the record, I also found #2 jarring.

I understand you cannot write as if walking on egg shells; you have your position and maybe your readers do as well. But this is far from a settled matter, and Chiang's position (which was describing earlier rather than current LLMs, but I still think it arguably holds today) is arguably correct, or valid. I probably agree with Chiang more than I agree with you, which is why I find it odd to call it a blind or weak spot as if the matter was settled. Maybe "while I admire Chiang, I fundamentally disagree on some topics, such as LLMs" would have felt less jarring.

(Not saying you must write like this, and it's impossible to write in a way nobody will object to. I'm just explaining why I -- and presumably the person you're responding to -- found it jarring).

▲

auxbuss 3 days ago | parent [-]

I agree. And this together with the obvious misunderstanding of Exhalation re: thermodynamics led me to put down the article.

I don't think the article was written by an LLM, but I'm convinced it was LLM-enabled. Which is a pity, because the author seems to have some interesting things to say. But that's the problem with leaning on an LLM: you lose your own voice, and good writing is centred around voice.

	▲	the_af 2 days ago \| parent [-]
		It's actually even worse. I thought the author was talking about Chiang's famous statement about LLMs being "lossy compression", and was ready to admit LLMs progress so fast this may not be the full picture. However, this is not the author's actual criticism! TFA's states: > I won't belabor obvious points like his nonfictional views on current-generation LLMs being surprisingly shallow [footnote] The footnote then links to an alleged "rebuttal" to Chiang by Scott Alexander, link here: https://slatestarcodex.com/2018/01/15/maybe-the-real-superin... This alleged "rebuttal" is actually referencing this Buzzfeed article by Ted Chiang: https://www.buzzfeednews.com/article/tedchiang/the-real-dang.... Regardless of whether you agree or not with Ted Chiang, his article isn't about "current-generation LLMs"... it's about unchecked capitalism and the fears of Silicon Valley entrepreneurs (at the risk of misrepresenting Chiang, he's saying it's ironic that Silicon Valley's worst fears resemble a sort of unchecked, rampant capitalism). You don't need to agree with Chiang to realize he's article is sort of neutral on AI/LLM, and is actually a criticism of Silicon Valley entrepreneurs! TFA's author cannot critique his views on capitalism as "shallow" just because he disagrees with them, or misrepresent them as being about state-of-the-art AI/tech when they are actually about capitalism. How could the article's author (and Scott Alexander) completely miss this?

▲

mannykannot 2 days ago | parent | prev [-]

As far as I know, compatibilism takes no position on whether the universe is deterministic. It is, rather, an antithesis to the thesis that determinism is logically incompatible with free will (or, from a more deflationary perspective, it offers an explanation for how we could feel we have 'classical' free will even in a deterministic universe.)

▲

jebarker 3 days ago | parent | prev [-]

Something that helped me grok Compatibilism (I think…) is that there’s actually two layered ideas. The first layer is practical: there is a sense of freedom that is based on your actions and apparent choices being determined by your internal state and not just the state of the outside world. Similarly, that allows for a practical definition of responsibility. The second layer is metaphysical: because of the first layer these choices/actions have “meaning” and justify moral praise and blame. I agree with the first layer and not the second.

▲

gowld 3 days ago | parent [-]

Compatibilism is the belief we are only characters in a story, but at least we can enjoy the show. Put another way, we should pretend to have free will, because if we are pre-determined, we don't have a choice anyway.

▲

jebarker 3 days ago | parent [-]

Honestly, I don’t find that explanation helpful or accurate. I don’t think Compatibilism says we need to pretend we have free will. It says we do have free will, just not the metaphysical kind that most people want.

▲

root_axis 3 days ago | parent [-]

I'd add that generally, compatibalists don't believe that libertarian free will is the kind that most people want.

Most people believe in free will, most people also believe that their actions follow from prior causes, and most people also believe in moral desert, so most people are definitionally compatibalists.

The typical framing where people are asked to go back in time and imagine redoing past events without memory of the future pumps their intuition for time travel fiction where doing things differently is the entire point of the scenario. If you ask people "if everything happened exactly as it happened, could you alter the past to change the future?" most people would say no.

	▲	jebarker 3 days ago \| parent [-]
		Great point - I shouldn’t have suggested what most people want, that’s always dangerous