EA Forum developer
jp
Karma: 720
I am happy to have contributed something that I believe was near the frontier of rationalist knowledge on this niche subject, and I think I did a reasonable job of it with the time and experience I had.
I still endorse personally that it works for me and achieves goals that I like.
I doubt the post should be winning any annual review votes. Unless someone really wanted to give encouragement for more things like this.
I would be interested in someone doing this again with a deep research tool.
Thank you!
First up, I thought this line was strikingly poetic for a technical topic, and would be above-average quality even for a human technical writer.
So you’re absolutely right that it’s a different kind of environment — but it’s been engineered to work well for fungal metabolism. Nature grows molds on oranges; industry grows them in soup with bubbles and stainless steel.
Chat link. The second sentence is what struck me.
I’m enjoying being curious about the world around me with the benefit of being able to ask an endlessly patient expert. Go ahead and ask your favorite LLM how the citric acid in your gatorade is made.
I’m gonna try making a thread of interesting LLM conversations, maybe some fun prompts, images, or techniques.
From personal introspection (importantly not what the question is asking), there’s a difference between my ability to update towards two views:
“Oh the situation is different than I thought, and actually I should turn most of my attention towards this new thing”
I’m better at this when I have more energy
“I made a mistake and should stop doing X, or admit past fault”
I get worse at that one
On simple, questions of fact, that I’ve avoided tying into my ego in one way or another, I expect I’m better but I’m not sure.
I would also expect extraneous details like, “got sick and fell of the wagon” or similar to add significant noise. And with only one data point each, it’d be hard to know the variance to use. I’m guess I’d trust this study more?
To be clear I don’t know what I’m doing really. I do think that it failed to get the precise thing I was looking for though.
🙇♂️
“Cruxy” is a useful term to have in my vocabulary. I use it relatively loosely to refer to the type of thing I look for in a double crux. A consideration is more “cruxy” if it’s closer to a but-for support for a proposition. Interestingly (mildly) this is very similar to the definition of “crucial,” and in fact the etymologies are the same.
The California vibes in that video are immaculate. Even mentioning Windham Hill Records!
I’m a big fan of land reclamation, but hadn’t heard of most of these barriers, thanks!
You might be interested in this “Best of LessWrong” post, Make more land.
This post is very cute. I also reference it all the time to explain the ‘inverse cat tax.’ you You can ask my colleagues, I definitely talk about that model a bunch. So, perhaps strangely, this is my most-referenced post of 2022. 🙃
My explanation of a model tax: this forum (and the EA Forum) really like models, so to get a post to be popular, you gotta put in a model.
I’ve referenced this post several times. I think the post has to balance being a straw vulcan with being unwilling to forcefully say its thesis, and I find Raemon to be surprisingly good at saying true things within that balance. It’s also well-written, and a great length. Candidate for my favorite post of the year.
You might be interested in the debate here.
You get paid if Wave does well whether they like you or not.
Private companies can and do prevent people they don’t like from selling stock to private buyers. So as an addition to this comment, I’d note that “the ability to cash out before an IPO 5+ years in the future” is a strong reason not to make an enemy of your former startup.
This is really cool.
I’ve historically said 1-2 weeks of skilled engineering work. That will lower by a factor of 2 after this branch, plus some follow ups, get merged.
Thanks for writing about ForumMagnum! This software is so much of my life, but understandably gets little attention an an object in its own sake.
That’s changing a bit now, and more people are reaching out about using it. — I think it’s the best forum software out there.
If someone reading this wants to build an instance, feel free to reach out.
Talon Voice, which I use for voice-commanding my computer, is very fast and has a linux version. I don’t know if it would run on a small machine, but it seems worth a shot. And it seems perfect for your command-centered use-case. To get more information, I would ask in their slack.
I wonder how much of this is as a side effect of trying to make it resistant to prompt injection. If my limited understanding is correct, techniques to take over LLMs involve getting it to believe quite weird things about the situation it’s in.