Good luck! :)
The way you use intelligence is different from how many people here using that word mean it.
Check this out (for a partial understanding of what they mean): https://www.lesswrong.com/posts/aiQabnugDhcrFtr9n/the-power-of-intelligence
Interesting! I've recently been thinking a bunch about "narratives" (frames) and how strongly they shape how/what we think. Making it much harder to see "the" truth since changing the narrative changes things quite a bit.
I'm curious if anyone has an example of how they would go about applying frame-invariance to rationality.
These kinds of explorations (unusual and truth-seeking) are why I love lesswrong :)
I've found the post "Reward is not the optimization target" quite confusing. This post cleared the concept up for me. Especially the selection framing and example. Thank you!
Finland too (and I expect quite a few other EU countries to do so as well)
https://mobile.twitter.com/i/lists/1185207859728076800 AGI Safety core by JJ (From AI Safety Support)
Lily: If I was a parent I would change the fifteen minutes to ten minutes. Screen time is kind of bad for kids. I also like having an hour and a half for movies, but I think maybe it's a bit much?
haha that’s so sweet! :D
Tldr: Love used to be in short supply (for self and others). Read Replacing guilt and tried improv + metta meditation. Now it is in big supply and has lead to significant positive changes in my actions.
I have always been in a single-player and critical mindset, optimizing everything for me. Thinking about what would be a nice thing to do for others (and empathizing with their feelings) hardly ever popped into my awareness.
Over the last year,
I assume EA student groups have a decent amount of rationalists in them (30%?), so the two categories are not as easily separable. And thus it's not as bad as it sounds for rationalists.
Will you be approachable for incubating less experienced people (for example student interns), or do you not want to take that overhead right now?
What is the reasoning behind non-disclosure by default? It seems opposite to what EleutherAI does.
See a longer answer here.
TL;DR: For the record, EleutherAI never actually had a policy of always releasing everything to begin with and has always tried to consider each publication’s pros vs cons. But this is still a bit of change from EleutherAI, mostly because we think it’s good to be more intentional about what should or should not be published, even if one does end up publishing many things. EleutherAI is unaffected and will continue working open source. Conjecture will not be publishing ML models by default, but may do so on a case by case ...
This is lovely! I’ve a couple questions (will post them in the AMA as well if this is not a good place to ask)
What is the reasoning behind non-disclosure by default? It seems opposite to what EleutherAI does.
Will you be approachable for incubating less experienced people (for example student interns), or do you not want to take that overhead right now?
Metta (loving-kindness) meditation would be an example practice that tries to focus attention on actively loving others in order to get better at it over time.
I don't have time to currently point out to concrete research backing it up, but it's been often discussed positively on Lesswrong and the EA Forum and I have had surprisingly good results from it. In my experience though, it has quite a quick feedback loop so trying it out might be the most efficient way of testing it. The Waking up app by Sam Harris is a good starting point.
This is a great idea! I'm gonna try it out. It fixes quite a lot of things with existing systems, as you point out.
I'm curious though, since when have you been experimenting with it and how has it been? I'm assuming it went well, but I am interested to know more about the details in your process (setbacks, changes, etc) and expect it'll be helpful for others experimenting with this as well :)
I've often thought about this, and this is the conclusion I've reached.
There would need to be some criteria that separates morality from immorality. Given that, consciousness (ie self-modelling) seems like the best criteria given our current knowledge. Obviously, there are gaps (like the comatose patient you mention), but we currently do not have a better metric to latch on to.
I put my laptop on a box on top of my desk and use an external keyboard and mouse to operate it.
Love this initiative! I do have a question though. It seems that people with 100+ karma have most likely figured out how to write publicly with a decent quality. So this service would simply be a bonus for them.
Isn't it more important to enable this service for lurkers/readers on Lesswrong who haven't yet written many posts due to the reasons you've mentioned?
Disclaimer: I don't have 100+ karma and haven't written a lot outside as well - just privately in my note taking app.
I have over 2000 karma ~all from comments, and fear of not meeting the quality bar has definitely contributed to my lack of posts. (Two in ten years, and one of them was meta about the site itself.) So I am hoping I will find it in me to try using this and see how it goes!
To answer people's questions about the 100+ karma limit:
As lsusr notes, 100 karma isn't that much and can easily be obtained be obtained by commenting. I think that's a very reasonable bar to ask of people before offering this service. While I do think we can build the capacity to handle a lot of requests, I expect that without the karma bar, we'd get a lot of requests from people who hadn't yet invested much in the site (or in their draft), simply because those are the most numerous. At least to begin with, I want to reserve the service for more dedicated...
Anecdata: I haven't figured out writing with decent quality or audience awareness and have been burnt multiple times. I get a cold sweat each time. It just doesn't keep me from posting. But I can totally relate to it, love the solution, and will definitely use it.
Editing can improve quality, or it can keep quality constant while reducing writing time, which is also pretty valuable.
It's easy to get >100 karma from comments alone, without any top-level posts.
No I don't think it's a good assumption that most people past a 100 karma have figured out how to write publicly with decent quality (though, depends on what you consider decent).
I'm well past a 100 and I expect this to be very useful to me when I write posts.
And if we're talking in general then even the best writers usually have proofreaders/beta-readers (take Paul graham for example, every essay he releases credits at least a few beta readers)
I do agree it might be especially important to new people that don't have karma, though. It'll be interesting to ...
Thanks for writing this! While reading the post, I was also thinking that this heuristic of building better systems is useful for deciding what to work on in our career as well.
Disclosure: I am new to AI Alignment, and have picked this as my first reading to dive in to.
However, most possibilities for such crucial features, including this one, could be recreated in artificial training environments and in artificial neural networks.
I don't understand how you arrive at this conclusion. Is there some paper/reasoning you could point me to that backs this? Would be helpful.
Also, is this anologous to saying "We can simulate environments that can give rise to features such as general intelligence? " (Making sure I'm not misinterpreting)
My biggest reasoning for not babbling is imposter syndrome. So there's no better exercise than this to start babbling :)
I've been thinking about this too, and I agree with your conclusion. The way I think about boredom is that since we (or the environment) block off external stimulation, our mind is forced to internally stimulate us (assuming that we have a constant need for stimulation).
Initially, as you mentioned, this leads to remembering small tasks and worries that have been on our mind. But after exhausting that reserve, it has no option except to stimulate us with things we haven't been thinking about consciously, i.e - creative thoughts.
I've been failing to implemen...
I'm surprised fatebook.io isn't an answer here. I had in the past tried a bunch of personal prediction tools and felt dis-satisfied. Either because of the complexity, UI, or something else. Anyways, I've been using fatebook for a couple weeks now and loving it.
More info: https://www.lesswrong.com/posts/yS3d46m23wRKDQobt/introducing-fatebook-the-fastest-way-to-make-and-track