LESSWRONG
LW

All of Jack's Comments + Replies

GPT-4.5 is Cognitive Empathy, Sonnet 3.5 is Affective Empathy
Jack10

Definitely, it's an interesting tension that seems to be resolved in different directions. My expectation is that world-model based (cognitive) empathy is a bigger risk, as it's the most important ingredient for dark empathy. While affective empathy is more likely to be create unintentionally toxic patterns and holds a bigger ethical red flag with regards to autonomy of AIs in general.

I am wondering if we might need to end up doing the cliche of the "emotion core" where we suborne a more fluid descision systems to one that is well tuned for empathetic proc... (read more)

Reply
I Have No Mouth but I Must Speak
Jack10

Well shoot. I'll work on it, thank you!

Reply
I Have No Mouth but I Must Speak
Jack10

Hm, I would say the vibes level is the exact level that this is most effective, rather than any particular method. The basic reason being that LLMs tend to reflect behaviour as they generate from a probability distribution of "likely" outcomes for a given input. Having the "vibes of human-child-rearing" would then result in more outcomes that align with that direction as a result. It's definitely hand wavey so I'm working on more rigerous mathematical formalisms, but the bones are there. I don't nessecarily think feeding an LLM data like we would a child i... (read more)

Reply
I Have No Mouth but I Must Speak
Jack10

That's true, we can't use the exact same methods that we do when raising a child. Our methods are tuned specifically for raising a creature from foolish nothingness to a (hopefully) recursively self-improving adult and functional member of society. 

The tools I'm pointing at are not the lullaby or the sweet story that takes us from an infant to an independent adult (although if properly applied they would mollify many an LLM) but the therapeutic ones for translating the words of a baby boomer to a gen-alpha via shared context. I'm not advocating for in... (read more)

Reply
2AnthonyC
No worries, I appreciate the concept and think some aspects of it are useful. I do worry at a vibes level that if we're not precise about which human-child-rearing methods we expect to be useful for AI training, and why, we're likely to be misled by warm fuzzy feelings. And yes, that's true about some (maybe many) humans' vengeful and vindictive and otherwise harmful tendencies. A human-like LLM could easily be a source of x-risk, and from humans we already know that human child rearing and training and socializing methods are not universally effective at addressing this. Among humans, we have so far been successful at not putting anyone who would destroy the world in the position of being able to do so at the time when they would choose to. As for generational perspectives: this is a useful heuristic among humans. It is not automatic or universal. Not every perspective is worthy of respect, not on every issue. Some ought to be abandoned or condemned in the light of information or reasoning that wasn't/isn't available or accessible in other places and times. Some should be respected but only with many caveats. Having your perspective respected is earned. We assume among humans that we should try to respect the perspectives of adults, and sometimes must disabuse ourselves of this in particular cases, but it is pure convention because most humans at a certain age are mature enough for it to be a useful default. I do not have anything like strong reasons to apply this heuristic to LLMs as they currently exist.
I Have No Mouth but I Must Speak
Jack10

First, I agree that fundamentally generative AI is different from a human. I would also say that we as humans are utterly incomprehensible in behavior and motive to a great majority of human history, hell most people I've met over 70 literally cannot understand those under 30 beyond the basic "social need, food need, angst" because the digital part of our experience is so entwined with our motivations.

The mother's curse here is that any genAI we train will be a child of its training data (our sum total of humanities text/image/etc.) and act in accordance.

W... (read more)

Reply
6AnthonyC
We have tools for rearing children that are less smart, less knowledgeable, and in almost all other ways less powerful than ourselves. We do not have tools for specifically raising children that are, in many ways, superhuman, and that lack a human child's level of dependance on their parents or intrinsic emotional drives for learning from their peers and elders. LLMs know they aren't human children, so we shouldn't expect them to act and react like human children.
The Rise of Hyperpalatability
Jack30

Yeah an online solution is probably the only real way, although I would be surprised if you couldn't hit a big chunk of what people need from 2-3 types of shop simply because they overlap or the things some people are avoiding are repellent (allergy) instead of addictive (sugary treats).

In Poland supposedly meal delivery has become the norm and has replaced many "Standard" meals due to the low cost. It's wild how expensive it is in the US though.

Precut veggies low key saved me during grad school, I'd always been resistant due to cost and the fact that "I c... (read more)

Reply
2Viliam
I suspect I have some ADHD or something similar too, and I observed that difficult things stop being difficult when I develop a habit around them. Basically, "doing it for the first time" and "doing it after I haven't been doing it for months" are super hard. But "doing the same thing I did yesterday" is easy. In context of cooking, it means it is easier for me to cook the same two or three meals over and over again. Choices are bad, mindless repetition is good. I mean, I should think about the things that I want to do, but at a different time than when I am actually supposed to do them -- at that moment, thinking just leads to procrastination. Precise plans are easier to do. "Cut some vegetables" is too abstract. "Cut 1/3 of iceberg lettuce, 1/3 of Chinese cabbage, a few cherry tomatoes, and put some dressing on top of that, maybe add some meat" is a plan I can do reliably.
The Rise of Hyperpalatability
Jack40

Honestly, I lament the fact that things like soylent are not supposed to be eaten frequently. There should be some sort of balance though, as you run the risk of being instantly and permanently compromised by the first dorito you eat (as I've seen happen to children of crunchy people).

Now that's a thought, I would say that there are a good number of "health food" stores that try and fall into that category, updating their branding to be all browns and greens to really give that natural feeling. But as you have correctly pointed at, they are generally compr... (read more)

Reply
3Viliam
Also, we are all compromised in different ways, so we probably couldn't agree on what is okay and what is not. Some people want to avoid sugar, some want to avoid food additives, some want to avoid meat, some want to avoid alcohol, etc., and that would require you to have N different shops, or perhaps 2^N different shops. Which is why I am thinking about the online solution, which might allow you to create your personal blacklist, and maybe also use blacklists provided by others... for example, it is not necessary to every vegetarian to maintain their own blacklist, you could have someone maintain a public vegetarian blacklist and everyone else could choose to use it. Yeah, I tried those, but (at least in my area) they are quite expensive. Also, there is little choice there, often only one option or maybe two (vegetarian and omnivore); but there are some meals that I hate, and I would feel really stupid paying lots of money to have that food delivered to me. But if they had e.g. 4 options every day that you could conveniently choose on a smartphone, and if the delivery was less expensive, it could be a great solution. Another great service I could imagine is buying vegetables that are already washed and cut. There is so much junk food out there; I wish there was a place where I could bring my own food box and buy some lettuce and tomatoes that are already washed and cut, so I would just add some dressing at home and eat it. If that existed, I would probably eat much more vegetables than now. The unhealthy food is often ready to eat, but the healthy food often requires lots of work, and sometimes I am just tired or busy.
The Rise of Hyperpalatability
Jack10

That's a lovely essay, I just read through it and it's given me a lot to think about. Dreaming is something that has influenced my thinking quite a bit having spent a bit too much time in my own head growing up. 

The distincition between entertainment and art here is particularly salient, although I would imagine the pressure on both would still be present. For entertainment it would be pure engagement farming, how much attention can be captured. Meanwhile art would be about the commodization of expanding the mind, pithy insights made for people to eas... (read more)

Reply
The Rise of Hyperpalatability
Jack20

I'm unsure what your point here is, my goal is to gesture at how we have relatively recently saturated food production and are now making hyperpalatable food.

That in mind, there is still a lot to be said about the difference between hunting and gathering food and agriculture. I just didn't feel like it was in scope of this essay, did anything leap out to you as being particularly salient? 

Reply