Just this guy, you know?
Dagon
Karma: 13,998
Rephrase in active voice, please. WHO are you worried about discouraging you from voting however you like?
The only official prohibition I know of is mass-voting in order to manipulate someone’s karma (upvoting or downvoting many posts/comments in a short timeframe, based on the poster, not on the contents). And I know of no metric for whether your opinion is more or less valid than anyone else’s, nor what even “average ability-to-evaluate” actually means.
I tend to vote up things I’d like to see more of, and vote down things I’d like to see less of, and abstain if I’m near-indifferent. This is based on gut feel more than any specific list of dimensions—there’s much complexity in what I find a “good post or comment”, though there’s less complexity in what I hate. I find myself upvoting about 5:1 more than downvoting.
Private thoughts are not punished. You can feel and believe literally any crazy or harmful thing, without repercussion. As long as you don’t express it, call for normalization or acceptance of it, or act on it. Private = undetectable, harmless, and no basis for judgement of others.
Those actions (including acts of communication) CAN carry judgement and reprisal, as society tries to protect itself from variations in dimensions it considers harmful.
Absolutely, and thanks for demonstrating!
One CAN habitually or based on some knowledge of the audience be somewhat more precise (never perfectly, of course, and there’s an efficiency/precision tradeoff that’s REALLY unpleasant). But some aspects and details of an idea to communicate can best be identified with 2-way communication.multi-turn discussions allow reflection and refinement, to see how party A’s words impact party B and to add more over time.
Note: “truly precisely” is important in my statement. One can get closer and closer with more words/bits, up to some asymptote.
Standard theory around here is that you should reject offers where you’d get less than $5.
...
standard theory around here is also that nobody should make or give in to threats
You’re massively and incorrectly overgeneralizing the standard theory. In ideal rational-actor perfect predictor scenarios, one should be known as the type of person who doesn’t take bad deals, and the only way to do that against perfect predictors is to BE the type of person who doesn’t take bad deals.
The two things which don’t apply in your post:
1) you have no reason to believe the other person knows you or that your intent has any impact on their offer. They’re not a perfect predictor, so turning down their $2 just costs you $2, without changing the probability that they offer more in other universes or future interactions.2) It only applies to certain types of threat (where it’s not a direct causal clear action). Nobody’s advocating you withhold your wallet from an armed mugger.
Have you noticed a correlation here with other personality traits (status-seeking, humility, curiosity, etc.), or across domains (often on tech questions, rarely on politics)?
I have long observed this divergence, and it does feel somewhat bimodal (but is more continuous when I look for examples more carefully, for me at least). And to me, it seems more often topic/context sensitive than innate traits of a person. Though there are a LOT of people who I only interact with in limited domains, so it’s hard to tease apart the reasons.
Language is always ambiguous. Experience and beliefs cannot be encoded truly precisely, no matter how big your vector. The human advice is “use more words”. Instead of “green apples”, say “unripe apples” or “ripe apples like granny smith, which are green”.
Advice for LLMs is probably “use a more precise shared embedding space”, which kind of boils down to the same thing. More bits of transmission = more fidelity of concept.
One way of looking at things is that it puts a LOT more pressure on the fundamental conceit behind Democracy: one person one vote implies an equality on dimensions that nobody wants to look too closely at, because it’s very difficult to reconcile with our experience of other humans. Moral weight and dignity, everyone IS (or should be) equal. But decision-making and directing group behavior by coercion (aka voting)? hard to see it.
AI just makes the power/influence/decision-making gap much much wider. It doesn’t change my strong opinion that Democracy is the worst form of government, except all the others that have been tried. But it does make it even more obvious that Democracy is a choice of which untrue beliefs about human behaviors we’re committing to.
I don’t understand how human-piloted drones (the ones in use today, and almost certainly for early terrorist use) have anything to do with AI.
If you assume pure subjectivity (it’s unmeasurable and incomparable), there’s not much to study or say about it. That’d explain by a selection effect why most researchers choose to focus/redefine to problems they can make progress on. Camp #1 has something to publish.
Camp #2 may well be correct, but it’s a very uncomfortable position for moderns. If you can’t measure or define consciousness, what are you basing all this talk of human rights and respect-for-others on? If you say this in the wrong crowd, you’ll be booed or assumed to be a eugenicist or something.
This just moves the interesting bit up a level. First contact between superintelligences of different origin stars.
[thanks for the interesting discussion, I’m likely to bow out after this—I’ll continue to read, but probably not respond.]
I’m curious what change or behavior you or your parents would undertake if LEV is somewhat likely, which wouldn’t benefit you if it’s very unlikely. My model of most people (including myself) is that “unknown remaining lifespan between 20-60 years” is unlikely to differ much from “unknown remaining lifespan between 50-200 years, or longer”.
For individual decisions, most long-term horizons get rather flat before 20 years.
If your parents are old enough that they’re retired and probably have < 25 years remaining, it could matter, but that also massively reduces the likelihood they’ll be able to take advantage of any breakthroughs.
Life extension is an unusual technology in that it matters much more whether it will be available in your lifetime than whether it’s available right now
Well, no. It matters if there are choices/interventions that you need to make right now in order to gain the benefits. It matters a little whether it’ll be available decades from now (presuming you still live) only to the extent it changes your retirement/savings decisions, but for the VAST majority of people “being healthy for 20+ years and living another 15+” is very similar in terms of financial planning to “being healthy for 110 years and living another 50″. It’s just not possible for reasonable long-term plans to diverge that much.
Glad to see it being explored fully. Also pretty sure that the identification problem will remain a critical limiter—if we want random killing of all/most civilians in an area, gas, biological, neutron bomb, or just plain conventional siege (starve them, bomb any truck getting near) is likely to work.
It’s when you want to kill/disable ONLY combatants or insurgents, and they’re mixed in with civilians you want to protect that you’re restricted from using these very effective indiscriminate weapons. Which is almost always the case.The area-denial uses (perch or hunter-seeker) are horrific, and likely end up with similar hassles as landmines do: have to send in decoys or attractors to let the drones destroy themselves. In a bloody enough conflict, victims might be tempted to use prisoners from the side who planted the mines to find/detonate the mines.
Fully autonomous weapons have roughly the same alignment problems we’re always debating around here: how to make them target only the things we want targeted right now, and nothing else?
I think trying to separate belief/expectation this completely from actual observable results is a mistake. The single biggest mover of expected lifespan is MEASURED (or at least reported) life extension (or significant health extension into later years).
So a lot depends on WHO you’re talking about. The general populace is likely to lag in their beliefs vs what is possible. Early-adopters and pundits are likely to be ahead of what’s possible, but won’t track truth very closely (they’ll believe it’s sooner or easier than it turns out to be).
A useful conceptual dimension in this kind of exploration is “legibility”. You can only optimize what you can measure. As human culture has gotten more focused on efficiency, it’s been forced to concentrate on measurable outcomes (because you can’t define “efficiency” using unmeasured values).
The protection of slack (inefficient space for un-measured values) is critical for my flourishing, at least. And it’s exactly counter to what most rationalists and EAs prefer to do: measure and optimize.
Many posts would benefit from including those things. But most authors would do it poorly if forced/strongly encouraged, so I’d rather have it be intentional and just part of YOUR habits.
Just start your posts (or at least those which you think benefit) with a TLDR and Epistemic Status header in the text.
upvoted, and agreed it’s not new, but is well worth emphasizing and exploring.
There are (at least) two views of language and communication:Encoding of concepts, distinct from actors and recipients. This is the view of ideal communication, mutual knowledge (or progress toward mutual knowledge).
Behavior intended to influence others. Speech is a targeted manipulation of models and beliefs in other people’s minds, intended to change how they think and act.
Neither option is correct, and both are at a level of abstraction that “correct” may not even apply. but switching between them makes communication about communication very difficult.
I disagree pretty fully with the premise, and I think all the details fall apart because it tries to use formal definitions for something that doesn’t work that way.
Detecting something is necessary for minding it, protecting it, honoring it.
This is simply false. Historically and currently, MOST humans mind, protect, and honor things they imagine or models they infer. Minding and honoring are not behaviors, they’re beliefs and attitudes. Protecting could be either, but I don’t think you’re talking about physical protection from damage (which doesn’t have anything to do with life or consciousness—you can protect a rock).
You’ve snuck in a very false analogy by using “you”, rather than “a brands-new entity each iteration which is like you but without continuity to the future, or knowledge/memory about the iterations “
it’s not at all guaranteed that a model-in-training experiences anything like you imagine, or that it reflects on itself as a human individual would.
Economic theory without any mention of exchange or transfer of property and assignment of labor is unlikely to make predictions that fit the real world. Come back when you can include Coase’s theorem (or an alternate formulation of trade and comparative advantage).