FlorianH

Karma: 789

FlorianH 12 Sep 2026 9:20 UTC
8 points
0
in reply to: artifex0’s comment on: Let’s Own the Term “Elitism”
Spot on imho: “populism” is rather clearly not at all e.g. in OP’s words a believe “that a nuclear power plant design should be settled by popular vote”.
To your definition I’d add that it often comes with the aim to attract voter support for these dumb things all while ~ hideously pushing an agenda that is very clearly not in that electoral circle’s interest and maybe mostly unrelated to what they officially stand for (e.g. talk poor people’s problems, but benefit the rich firms—and maybe in the worst case, when a solution for the poor people’s problems could be in sight, even purposely not fully tackle it as it’s not in the party’s interest to make the people less angry).
Slightly ironically though, your “technocratic” though would be the better opposite to OP’s definition of populist in the nuclear example, rather than the usual one (yours/mine).
Imho the better opposite to the commonly used notion of “populism” is more things like “serious”/”honest” or so—though esp. the latter isn’t really a fit as most people who talk about populists and the rest know that in politics much is not in a broader sense honest anyway.

FlorianH 11 Sep 2026 19:14 UTC
2 points
0
on: Replace Net Metering With Batteries
Net metering is indeed globally known to be bogus in a modern power system and is more and more being replaced. From outside I’d guess you’ll be able to get rid of it sometime soon even in a ~ failed state like the US despite potential lobbying constraints.
Many locations have asymmetric grid fees: when you draw, you pay for the energy and often even more for the grid per unit of energy, when you feed in, you get something but not so much. This may even lead to an overinvestment rather than underinvestment (at least if one doesn’t weight your resilience question very much) in batteries—that are in addition not even very grid-usefully utilized.
Further, trend clearly today—even if too slow and not radical enough—towards dynamic electricity pricing, even for simple households, a major step towards finally aligning your consumer & prosumer incentives with actual societal costs of marginal power consumption or feed-in: your incentive to install PV, use your power, and install battery and also to dispatch it IIF it’s socially a net plus.
Difficult to objectively assess what the value of widespread local battery availability is, and how to trade it off against other forms of reasonably decentralized storage—afaik currently larger scale battery packs are multiple times cheaper than single-home batteries, so bundling might still remain better; your resilient ‘island’ might ideally be the close or the commune or so instead of the house maybe. In many places e.g. in Europe, EVs are still rapidly increasing and these mean you have your local battery. If the ideal scale and technology isn’t so obvious, your subsidy risks to be yet another case of ‘picking winners’; the market for storage and new consumption patterns and maybe system-resilience/robusness solutions changes so fast that I’d prefer to mainly focus on setting incentives right more directly for what we actually want, which is an incentive structure that might yield some benefits to home battery installations, but I’d not be surprised if the ideal solution would mostly end up being something slightly larger scale at least in reasonably densely populated areas.

FlorianH 11 Sep 2026 18:45 UTC
2 points
0
on: Post-AGI, we are all jobless aristocrats
FWIW, I explain in How Econ 101 makes us blinder on trade, morals, jobs with AI – and on marginal costs what fallacy according to my experience leads fellow economists to dismiss salary shortage scenarios.

FlorianH 11 Sep 2026 18:44 UTC
14 points
5
on: Post-AGI, we are all jobless aristocrats
Spot on on not needing to work IIF we get redistribution right—both within and across countries (assuming these still exist/assuming a non-everything-altering takeoff).
Given how rich nations treat poor nations today, I’m unconvinced about us getting redistribution globally right even in a world of more abundance than today, though who knows.

FlorianH 10 Sep 2026 15:31 UTC
1 point
0
on: I want to bitch about the increased salience of AI risk
You think you’ve invested, you think you’ve maybe missed out on something.
Instead, you’ve gotten the biggest reward you can ask for: Personal motivation. Your ‘religion’-style mission—plus one that, exceptionally, seems to be objectively a true one. That’s a value that’s even from an own egoistic perspective—not from a gross physical material but from the only relevant, namely psychological, one—simply priceless.
Idk how perfectly the that claim applies to you, but dare to speculate that way as, fwiw, every now and then when I think me, or some ‘good doer’, worrier-warrier, feels like they exhaust themselves and might be giving up too much, I remember that it’s super easy to overlook: even if I’d ever die of exhaustion from trying to dedicate myself to a purpose—I’ll have been the lucky one, as I had purpose—while the thing making live most nearly unbearable in some people that I sometimes have around me seems to be exatly that they lack that purpose. They don’t have religion anymore, and nothing replaces it. They are egoistic consumers and fight for their consumption and overconsumption—not because they’re rich, but because they’re purpose-poor. You instead have made it, you’ve replaced religion or so, for a real thing worth fighting for.
That thought can even ‘bite its own tail’ in my case: if a hint of ‘is it worth it to try to help somewhere on this world’ arises, when I remember how I myself get positive drive from it, it boosts my motivation again. Maybe in some sense naive but working.
Either way, can in my smaller way empathize with your ‘positive rant’ and love it; love to everybody dedicating themselves to the care-against-AI purpose or any other good aims, and love to this platform exactly for it. Anyone, don’t stop, may the purpose stay with you and may it eventually yield a less scary and less hellish world than it currently seems to be for so many beings.

FlorianH 3 Sep 2026 19:48 UTC
5 points
−3
in reply to: DanB’s comment on: Sen. Bernie Sanders (I-VT) and Rep. Greg Casar (D-TX) introduce legislation to ban Artificial Superintelligence and temporarily pause advanced AI development
If anything I’d believe in reverse causality: if it’s so unpopular among the conservative base but only the far left is bringing it up, maybe a large range of non-far-left politicians are a bit in bed with the big companies. But even that I’m not sure.
I have no clue; my gut feeling simply is: even in a super partisan world your conservative base doesn’t “oh Bernie is for halting let’s not halt”, neither consciously not subconsciously very much; ; and in that sense also politicians don’t have a straightfwd reason to either.

FlorianH 28 Aug 2026 19:31 UTC
2 points
4
on: I cancelled my AI subscriptions because I am worried about catastrophic risks
There is something incoherent about believing a a company might cause human extinction, but then also being their customer.
Sad but rather not not incoherent at all for most realistically parameterized humans. I’s a most common and mostly individually rational outcome of so many a Public Goods game like this one.

FlorianH 25 Aug 2026 19:49 UTC
5 points
1
in reply to: kbear’s comment on: FlorianH’s Shortform
I was indeed briefly wondering whether—and it might not be 100% what you write but I think kind of—it could be related to formatting the answer into a difficult to parse language being at least somewhat higher-expected-reward as stupidity might, in the rare cases where there is any, be quite a bit less likely caught, while only quite rarely the made-to-feel-dumb user downvotes as he’s more feeling like downvoting himself or thinking simply not much and, as you write, thinking it’s surely all right..
Whether marketing by the creator or instead—and I’d rather think of that—evolved-marketing-by-the-LLM i.e. simple reward hacking as the user itself (with a RLHF interlocutor, a trainer-LLM, or a daily-user-when-model-is-trained-on-him), can theoretically remain open.
Yes I now vaguely believe in that possibility. But maybe I’m just dumb and it’s an excuse still :)

FlorianH 25 Aug 2026 16:10 UTC
44 points
22
on: FlorianH’s Shortform
I often feel dumb* when interacting with Claude**: it explains things, say in a code chat, of, say, a reasonably but not overwhelmingly complex project, and I kind of mostly know what it means but am not fully sure. And I wouldn’t exactly say it’s too terse. But all to regularly it presumes I know certain things/know what it refers to, while I don’t exactly know. Then I tell it what I was thinking about and it’s mostly the same. So it’s weird I’d not have known what it meant readily from the start.
I wonder whether it might really be me or instead people have the same type of experience?

*Not dumb-with-certainty, but more am-I-dumb-it-really-looks-like-though-that-feels-weird-and-I’m-even-post-hoc-not-sure-whether-instead-it’s-not-rather-it-that-puts-things-weirdly.
**Whether with ChatGPT the same or not I cannot readily say, I mostly claude.

FlorianH 20 Aug 2026 21:18 UTC
−1 points
0
on: FlorianH’s Shortform
There’s a Les Inconnus sketch where a hunter explains the difference between the good hunter and the bad hunter. Struck me this imho-you-might-love-it-too 1min video I’ve seen a quarter century back is a rather perfect way to illustrate the hubris in ‘machine cannot ever create real X’ (X being understanding, art, or what have you). “The bad hunter sees something move and shoots. The good hunter, he, sees, something move, and, then, he shoots — but he’s a good hunter.” Only works in video (and in human conversational reality very much, I’d say): The Strangers—The Hunters (The Good and the Bad) (Settings → Audio Track → English, though the original carries the timing better.)
I’ll think about this now whenever someone explains that when a human does X it’s understanding, and when a model does X it’s mere pattern-matching. Our instincts are all not such a new phenomenon in some sense.

FlorianH 15 Aug 2026 17:57 UTC
1 point
0
on: All Utilitarians Should Be Classical Utilitarians
Incorrect view #3: Moral views shouldn’t be too demanding
[..] even if you do an action that’s extremely good but not the very best action that you could have done, you’re still doing an incorrect action.
[..] scalar utilitarians argue [..] that actions that produce a significant amount of good are still considered good even if they don’t produce as much good as possible. Similarly, satisficing utilitarians argue that we have a duty to do good, but that the obligation only requires us to do a certain minimal amount of good.
I disagree with these views. Just because classical utilitarianism is demanding, that doesn’t mean that it’s wrong.
For one, from the perspective of the universe, all of our suffering and happiness matters equally. [..]
For two, most people actually already believe in extreme demandingness in many circumstances. For instance, virtually everyone believes that a slave holder should free all of his slaves even if it would cause him to become homeless [..]
This continues the eternal talking-past-each-other between utilitarians and its foes:
a. On the one hand you’re right to call out the 2 pseudo-solutions I first bold-emphasize above as ultimately unworkable/unconvincing.
b. On the other, your two point I later bold-emphasize above simply don’t answer the demandingness problem: we’re humans, we’re beasts, we’re far from saints, and that’s undeniable, and any utilitarianistic ideas won’t change it, and it’s not a mistake to call out the discrepancy of that reality as being at odds with ‘your theory’, however much we deplore it (and I’m not apologetic to propose to be nearly 100% sure even you yourself you’ll be very far from it, despite how you write “most people actually already believe in extreme demandingness in many circumstances”).
But there is, imho a more productive way to put the entire thing—and still super simple; you may even just call it a question of wording, but I think it matters for the conversation:
1. The ideal is bare utilitarianism, just as you put it, yes let’s aim at that natural focal point
2. Aim at this as much as you can, evaluate the morality of your options by the degree to which they go in the direction of 1.
  1. While you’ll extremely reliably fail to resist deviating from the option that would correspond most perfectly to 1., and instead chose one that is, say, more to your own material benefit and/or to that of your near and dear ones, try to neglect as little as possible the wellbeing of those beings—and potential beings—whose fate you’re superficially inclined to overlook.
If we want to operationalize an overly simplified, stylized version of this:
max U + alpha * sum(u_i) where U is your personal utility, 0<alpha<1 your altruism factor, and the u_i the utilities of all others. Extend with a more realistic preference structure, but the gist is that you want to act in line with an as high alpha as your rather egoistic self permits you.
Telling humans to pursue every and always and perfectly 1. is a bit like asking a loose stone on the roof to never fall down on people’s heads: it’s the right thing but the stone doesn’t have much capability programmed into itself to avoid doing what it naturally does, so it cannot really achieve much. It’s as ignorant as to believe we can tell people to be perfect full communists and sing kumbaya every evening while working hand in hand in perfect kindness only for the public good.
So the “small-heart utilitarianism” how I call what 1. & 2. describe, is the more realistic option—as long as we don’t wholly reengineer ourselves (physically, not purely with re-education).
For the hypothetical perfect altruist it’d collapse to bare utilitarianism as you describe it, for us earthlings it seems to me to be the natural realistic extension.

FlorianH 9 Aug 2026 10:37 UTC
3 points
0
in reply to: Erich_Grunewald’s comment on: Why You Should Almost Never Use AI to Write Anything Substantive
Sorry time not permitting as of yet not a full account* at all of how exactly AI helps but fwiw—with risk of stating the obvious:
Why does it help: One thing that just struck me now as I caught myself in a tiny waste of time while writing a (not even complex) email: Only re why AI might be helping me so much in writing: tiny epiphany (yet to observe consciously more whether it was just now that it seemed so exactly like this, but I’m reasonably positive): Maybe a large part is that (a) I lack what I now suddenly would summarize as a sort of ‘wording RAM’: which sub-element, which sub-thought did I already put down on paper? Which ordering of the thoughts is the ideal one go gently get the reader to the thought? Which words can be cut to say the same things briefer/smoother? And more generally that combines with (b) highly imperfect understanding of ideal/smooth (or even merely /correct) grammar in convoluted sentences as (i) non-native speaker who (ii) even in my mother tongue was scholarly good in ~ everything other than essay writing, making always convoluted sentences that don’t work easily.
[To avoid potential false interpretation: this does not in my case correlate with an inability to read (though indeed slow reader) nor speak; I’m not a great speaker but on a topic that interest me I can well give very free and maybe captivating presentations with real drive and w/o scripting (in fact scripting and presenting scripted is instead the night mare for me).]
At least an example here of how: Bit banal and is not at all the whole range of use cases at all, but it’s not atypical and I think it does represent an example where AI’d have helped: I could probably have made this comment—or the like—nicer for you to read i.e. significantly shorter, more crisp, maybe leave out 1-2 boring excursions in it or so (and might have actually done other times) by: (i) caring less about writing the aligned/fully formulated sentences above, (ii) drop them to the LLM, asking to ‘no unnecessary changes but what meaningfully helps readability/niceness, as always with the least changes necessary for me to edit as quickly as possible’ or something in this direction (the “as always” here only for you to see: that is a main ask I often have). Strikes me this here isn’t something ‘substantive’ but even if this was a longer text, the same would apply.
*As there’s no single ‘normal’ workflow for now, it’ll take time to give a representative account if I ever manage to prioritize it (vague hope); but it strikes me it is indeed much related to the basic writing issue I introduce in the “why” above. Essentially imagine me as a person very proud of their ideas and—me self proclaiming.. - precise, subtle thoughts, but bad and anxious about writing, meaning I’m seeing myself rather robust against risk of ‘stealing’ half-baked AI ideas but instead taking its niceness of expression, and I think you might see a bit what I personally claim my experience to be very much and why I am not so afraid of some what you write; whether it’s a biased perception/claim I cannot easily proof of course, and of course even if my account shouldn’t be too biased in general, I admit exceptions can of course occur with some regularity).
A more pants-down solution could of course include me asking my LLM itself to tell us how I use it—given sycophancy levels of current models it might not even turn out that embarrassingly for me—tbc if I ever find time to provide a fuller account and hopefully w/o embellishment :).

FlorianH 6 Aug 2026 14:23 UTC
3 points
0
on: Why You Should Almost Never Use AI to Write Anything Substantive
From the perspective of a near writing-handicaped person—yes I also cannot explain what makes lesser writers like me so bad writers all while not being usually seen as stupid or sloppy in general nature—I insist w/o further ado: I am 100% convinced AI is Pareto improving my writing whenever I use it well* - no matter at which effort level and nearly no matter for which aidience. It’s simply a fact.
*And if course I also sometimes fall into the trap of using AI too sloppily in which case it might have been counterproductive. That’s mostly a topic that can be delt with differently than banning it from my tools basis.
Maybe you’re a good writer and for you there’s no or nearly no way letting the AI support you in sheer wordsmithing improves your text—it surprises me slightly but I can believe it. Today.
My general impression is that half-convincing anti-AI-writing things, rather prevalent, somehow get surprisingly much traction on LW. I guess I might oversoeculate if I imagine a good-writer-elite half-consciously sensing some of their wordsmithing advantage to be eroding—but I attributet so far enough credence to the hypothesis that I dare put it in the room here to maybe see whether I find a few bad-writer co-conspiracists to agree—or others to rebut :)

FlorianH 31 Jul 2026 16:56 UTC
1 point
0
in reply to: papetoast’s comment on: So you want to use plants to reduce indoor CO₂
Agreed on substance – I started from the same skepticism and remain with it. Caveat just on structure of your argument: the population-level checks (SAT scores, AC vs. window-opening regions, decades of submarine/spacecraft work) are decisive against Satish-scale effects but close to powerless against small ones. Say true effect were d=0.1: all those comparisons would come back null, and so would most lab studies; hopelessly underpowered. So “we looked hard and found nothing” kills the version that caused the panic but leaves the subtle version about where prior plausibility put it. I don’t think it’s likely either to even have that lower-but-not-entirely-negligible effect, and the practical upshot is identical, because the honest case for ventilation is comfort, bioeffluents, and sleep rather than IQ points (well, assuming that conventional wisdom remains right) – and whether indoor plants make us happy will also continue to hinge on non-CO2 things anyway.

FlorianH 30 Jul 2026 22:18 UTC
3 points
0
on: So you want to use plants to reduce indoor CO₂
Carbon dioxide is bad for cognition
Except that it actually mostly/likely enough isn’t? If anything it’s mostly a proxy for bad things (and ☘️ might accidentally help with some of them but not others)

FlorianH 23 Jul 2026 21:36 UTC
1 point
9
in reply to: _rpd’s comment on: Mathematicians are Feeling the Doom
given that AI-associated x-risk largely depends on our level of mathematical competence, this is also arguably the most important time in history for the social role of mathematicians.
Strongly doubt that.

FlorianH 21 Jul 2026 20:15 UTC
4 points
0
on: FlorianH’s Shortform
Your views on Data Science studies future?
Do you have advice for a university thinking about what curriculum to teach at the interface of Data Science/AI/Business?
I guess the most basic obvious points are covered, but maybe you have interesting points/opportunities one might be overlook and that might realistically be taken up by a university today, aiming to turn out graduates attractive to a labor market in the next 4-10 years (obviously in a not-full-AI-take-off-in-the-next-5-years or so scenario)?

FlorianH 18 Jul 2026 7:50 UTC
2 points
1
in reply to: Taylor G. Lunt’s comment on: Evan Rysdam’s Shortform
“Emptier than your average flight would have you naively believe” should be true I reckon.

FlorianH 18 Jul 2026 7:42 UTC
1 point
0
in reply to: tpotthinker’s comment on: Longtermism is very intuitive.
Again don’t fully disagree: yes exposing people, if you manage to get them to listen at all, a bit longer to longtermism, could easily make some of them agree with you.
But
Most people who disagree with longtermism probably do so because partisans on their side of the political spectrum disagree with longtermism for x reason [..]
seems nearly a bit conspiratorial, while reality seems to me so much more banal instead: Most of us are simply not wired to care a lot about these questions. Before anyone would have to exercise some propaganda on you, natural instinct is to care about much more tangible things than an (felt) abstract future expansion of humans and so on. It’s just not what commoners think about. Not because sb tells them it’s not important. Simply because natural drive to even think and care about such things is lacking. When you write most “haven’t been exposed to a fair analysis of how longtermists resolve most of the complaints brought up” you’re for sure right, but that ‘haven’t been exposed’ in itself is signal, it’s a symptom of their own preferences, maybe more than of external system factors: People don’t care about longtermism questions and neither about arguments. Had people sought wisdom about longtermism, the market would trivially flood them with articles and media highlighting arguments like the ones we may make here for it.
So the problem is one level deeper in our psyche and in our moral behavior: as humans, we won’t typically engage with this thing for a start, and thus, also not be educated about it.
FWIW, I think this complacency need also not be surprising at all from an evolutionary perspective. We have mostly ad-hoc moral interests and tastes, those that have enabled us to highlight why we are the ones in the right and of importance and a broad set of things of that sort. Longtermism is really far out from that set of evolutionary pressures we were exposed to.

FlorianH 17 Jul 2026 23:03 UTC
2 points
0
on: Longtermism is very intuitive.
It seems absurd that most people don’t find longtermism as an ideology compelling
Not fully disagreeing. But it doesn’t mean it’s wrong to suggest longtermism is counterintuitive for us all in all.
Instead: Yes, human moral taste simply is completely absurd—when judged in the way that you implicitly call out.
Meet human moral tastes:
Rmbr how we want to give that one single poor child on the picture—yet ignore how we could save many more efficiently? And how our motivation pales just once we see there are these many to be fed instead of that one?
Rmbr how we all love to care for our cat and feed it the most evilly abused chicken day by day? And it’s not that we’d not know how the chicken is abused. Many really do know all while continuing.
Rmbr how we’re all happy going to protest against Shell, not for corrupting policy and cheating us, potentially, but merely for having sold big quantities of oil—that exact oil we’re until today continuing to order from them every day? Ok, this one a super random example in my mind for a seaparate reason today, but I just wanted to take one absurdity among the million you could pick.
Let’s not bore everyone with continuing this obvious list further.
--> So while I find your individual points do exactly make sense, and it seems like longtermism should be the most natural thing: In the aggregate, it is not in any way an absurd claim that longtermism is counterintuitive for many—it really is exactly that. But really, it’s also not a surprise, given what I’d guess even you—or at least most readers here—might most plausibly be able to observe (i) even within oneself—despite the slightly higher rationality than average humans’ - or, else, at the very least, (ii) among pretty much everyone else near and dear you.

FlorianH

Incorrect view #3: Moral views shouldn’t be too demanding