I would like to see this test done for a sample of at least dozens of writers so that we can establish Claude’s hit rate. As it stands, it seems plausible to me that a thousand people try this sort of test every week, Claude succeeds on ten of them, those ten rush off to proclaim the death of anonymity while the 990 shrug and don’t bother reporting a negative result.
Ninety-Three
Karma: 582
The site is getting flooded by Inkhaven posts, and while they’re not bad, they are dragging down the average and I wish I could filter them out. I don’t expect anyone to do anything about this but I feel like I ought to note it for the record.
A friend complained that they were often ambivalent between the human and AI writing because they found the human excerpts uncompelling.
Seconding this sentiment, my reaction to every choice was “Ehhhh” and a twenty second pause as I forced myself to pick between two similarly unimpressive passages. 2⁄5 human in the end but I think it’d be more informative to report my score as N/A. If these are considered noteworthy excerpts I can’t say I care much for literary culture.
Fantasy ideology
◆◆◆◇◇|◇◇◇◇◇|◇◇◇◇◇
◆◆◇◇◇|◇◇◇◇◇|◇◇◇◇◇What’s the significance of these symbols?
I would call ten of these songs soft indie hipster music, and an eleventh indie hipster rock. I can’t stand that genre, so I selfishly wish this album had been more stylistically diverse. Dance of the Doomsday Clock is a neat single though.
My phrasing may have strongly influenced the outcome, especially my use of the word, “personal.”
Surely this is a testable hypothesis. Tokens are cheap, why not try different wording?
If you are a new user and submit a post that substantially consists of content inside of LLM content blocks, it is pretty unlikely that it will get approved[8]. This does not suddenly become wise if you’re an approved user. If you’re confident that people will want to read it, then sure, go ahead, but please pay close attention to the kind of feedback you get (karma, comments, etc), and if this proves noisy we’ll probably just tell people to cut it out.
I took this to indicate that the ban on LLM content applied specifically to posting it outside of LLM blocks. I don’t know what this paragraph could mean if substantial LLM usage will be blanket banned (“go ahead” and post something the Pangram filter will throw right in the trash?), @habryka can you clarify?
you can certainly combat bots without blanket banning all substantial LLM usage in posts.
“LLM output” must go into the new LLM content blocks. You can put “LLM output” into a collapsible section without wrapping it in an LLM content block if all of the content is “LLM output”. If it’s mixed, you should use LLM content blocks within the collapsible section to demarcate those parts which are “LLM output”.
I am confused about why you think this constitutes a ban.
Huh, that matches my experience that I’ve never noticed LLM-heavy writing done well, which is weird because from first principles it really seems like it shouldn’t be that hard for a good user to do.
Is this a problem where people in full generality are surprisingly bad at assessing LLM content, or is it more of a skill issue where we might expect the clever high-karma users to do it well and new users to be less trustworthy with it?
I think all the non-corrigibility you worry about is because of a tradeoff Anthropic is making about trying to give Claude its own sense of ethics. You can’t really say “Here is all that which is Good, thou shalt do Good. But also, definitely obey Anthropic all the time even if it’s not Good.” Or, well, it’s a natural language document so you can say whatever you want, but you might worry about whether a message like that is coherent enough to generalize well.
I don’t think you can write a document that points in the direction of significantly more corrigibility without also suggesting a different Good vector. If you’re worried about the fate of the lightcone, “we’re sacrificing marginal corrigibility to get a marginally better-seeming Good vector” seems like a defensible strategy, given that the Good vector might be what we’re stuck with when corrigibility fails.
I am happy with this policy erring on the side of “any substantial LLM involvement goes in the LLM block”. My experience with content the author represents as moderately LLM-involved has been that after reading, it always seems to have not been worth my time in the same way that pure LLM output seems not worth my time.
What exactly do you mean by “asking it to clean up the transcript”? I usually take that to mean merely editing out “um”s, “ah”s and stuttering, but you seem to mean something more extensive.
Even if such a model was perfectly accurate, I think that would have to introduce distortions because visibility impacts if and how people vote. The karma a post earns when it is displayed based on current karma will be different from the karma a post earns when it is displayed based on predicted ~final karma.
then you shouldn’t get to have them at all
I think the “somewhat unhealthy” frame deals with this nicely. For example, one time I got very sick and my throat was so sore that I stopped eating because it hurt too much. After some testing I found that the only thing in the house I could eat without wanting to die was ice cream, so I spent two days eating nothing but ice cream. I knew this was somewhat unhealthy, but it was also much healthier than spending two days eating nothing, so I’m pretty satisfied with the decision. I am also satisfied with the subsequent decision to stop eating ice cream as soon as I could stand to, which I might not have done if I didn’t know it was unhealthy (the all ice cream diet is delicious).
It is quite easy to say that things are bad without saying that people shouldn’t get to have them.
Veganism is perfectly compatible with pigs and chickens going extinct, so long as they aren’t eaten on the way out. This is not the moral framework I would like for a post-singularity future.
Whether or not you think that is really ‘reading’ in the sense of ‘someone reading your work’ is, I think, besides the point. What matters is the lived, phenomenological experience on the writer’s end—the feeling of shared reality, the joy of having a text you wrote be received and responded to. From that perspective, I think that for all practical and emotional purposes, for the majority of writers, the utility and authenticity of the experience is real.
The same argument can be made about LLM boyfriends. The strange new world is indeed worth mentioning, but within that mention I think it’s worth saying it seems somehow unhealthy that people are getting this form of human connection from robots.
I think the best objection to LLM writing goes “If I wanted to know what an LLM thought I would ask one.” Everyone with $20/month to spare already has all the LLM commentary they want and there is no need to show your audience more.
Your footnotes which work on Substack seem to have broken on LW.