MuffinButton
2mo ago
Thanks! I did briefly consider loading it up as a prompt. I have been allergic to shelling out for tokens for personal use, but maybe it's worth it in this case. I suppose under the theory that humans are LLMs on a biological substrate, I already asked an agentic swarm of LLMs. I'll try it on the lower-tier models, see where I get, and hopefully remember to respond.
MuffinButton
2mo ago
Does anyone have a good resource (or set of resources) for getting started using AI in coding that emphasizes building understanding? I've pretty much eschewed AI to this point, but it seems it's only getting more integrated and capable, so I might as well develop the skill. My experience is mostly in python in a data science context, with a little bit of C and the rare instances of assembly thrown in to run isolated or low-bandwidth sensors with extremely limited onboard compute. I have generally preferred VSCode, so bonus points if the resource integrates it.
I think the best guide for me is going to read something like, "A harness performs functions X, Y, and Z. Let's walk through setting up a basic harness so you can develop intuitions around what it's doing and how you might modify an existing harness to suit your future needs." Conversely, a terrible guide for me might read, "Download (someone else's) harness from (website); it's the best one for coding (no explanation as to why). Point your IDE at it and start typing your prompt (no elaboration on what makes a useful or effective prompt)."
MuffinButton
2mo ago
I believe it was a very wise man:
No one can serve two masters. Either you will hate the one and love the other, or you will be devoted to the one and despise the other.
While the context (from Matthew 6:24) has God and money as the two masters, it applies just as well to any two non-identical selections: they will somewhere and somehow conflict and a choice must be made.
MuffinButton
2mo ago
Part of the reason 'adult human female' rubs me the wrong way is that using 'female' as a noun in reference to humans is associated with misogyny.
Is it associated with misogyny, or is it associated with accusations of misogyny?
Who has produced and maintained this association? As far as I can tell, it's the same people who would find it convenient if those using the phrase "adult human female" were tainted by association with misogyny and whose claims could be safely ignored on that basis.
MuffinButton
2mo ago
bimodal
three tiers
ಠ_ಠ
MuffinButton
2mo ago
God help me, I will beat it on Commander Ironman mode.
If you have the discipline to not savescum, I'd suggest running it in normal mode and just not reloading previous saves. People have had their IM runs bricked because the game bugs out at an inopportune time. If you absolutely must use the in-game IM feature, keep a backup save after each mission and only ever use it in the event said bug bricks your run. It may not hit you, but if it does and you're months deep, it hurts.
MuffinButton
3mo ago
It might help to think of the net spending tally not as "progress", but as "a (rough) estimate of progress". There are enough significant one-off events that the error bars are likely larger than the expected savings on timescales shorter than multiple years. As I see it, you're likely to be in one of two scenarios (which one is TBD):
- This year's expenses are anomalously high -> Spending carefully will reveal its fruits in future years, and it's just by chance that they're not visible yet.
- This year's expenses will be typical going forward -> Spending carefully allowed you to weather them without tapping savings, which under this scenario would be ruinous in the long term.
Either way, you're better-positioned than you would've been in the counterfactual world where you hadn't adjusted behavior, and I would concur with your gut that it is quite the win.
MuffinButton
3mo ago
I think you're correct that the poll numbers are the direct lever involved, but my understanding is that just kicks what gets @WhiningCoil about the story up one level, i.e., Platner got a pass (in the polls) on all sorts of other behavior, but what mattered (in the polls) was a rape allegation. It's still a worthwhile distinction, since the relevance and valence of some issue to political higher-ups and to the voting population can differ significantly (cf. immigration), but that might only magnify the frustration.
The closest analogue I recall was a lot of people up in arms that Andrew Cuomo was pushed to resign after a series of sexual harassment allegations rather than his disastrous decision to put a bunch of COVID patients into nursing homes.
Edit: grammar
MuffinButton
3mo ago
My final paragraph was being kind. It was a description of my own (conditional) internal state, so I can verify it was true. Its necessity is left to the judgment of the reader—though I note that none of what anyone posts here is strictly necessary in the barest sense (and yet...). I do appreciate you linking the infographic, but I cannot help but point out that if it was taken down from its original source and you had to instead link the graphic as part of an article written by a radical feminist saying that it tripped her BS detector in the direction opposite her biases and is probably crap, then it's probably not worth including it at all.
Casually linking to two studies and inviting people (in this case, me) to look at them and see which source is more reasonable in its methodology doesn't engender any particular conclusion other than that you didn't seem to have read them particularly carefully. That I had to click through to find that one of the (as-described) studies isn't a study, but instead yet another innumerate, made-up social media graph which added negative value to the discussion was, in a nutshell, the point of my post. I know the graph was made up, because Enliven did show their homework. I award them no points, and may God have mercy on their souls.
Touching briefly on the object level: I think we are in agreement on due process and presumption of innocence.
I'm not sure what the back half of your response (PBS? Upvotes? Your GitHub?) has to do with anything I said, but after reading through, I think you might be misinterpreting others' disagreement and skepticism as emotional reactions. While I can't speak to others' feelings or lack thereof, you did impute an intense emotional response to me, and I invite you to not do that again.
MuffinButton
3mo ago
I have no dog in this particular fight, as I neither live in Maine nor have any plans to do so in the next six years. I am also going to leave aside the object-level concerns, as I don't think estimating general population false accusation rates clarifies much due to the numerous selection effects and atypical—at least in scale—incentives involved here. However, I do strongly object to the blithe implication of equivalent credibility you make:
About that “believe all women” streak, it’s an open question what percent of rape accusations are false rape accusations. I will list two sources with different conclusions:
- Source 1 claims 0.2% of rape accusations are false.
- Source 2 claims 28.8% of rape accusations are false.
Both sources show their homework, so people can look at them and see which source is more reasonable in its methodology.
Source 1 shows their homework for an infographic they made (and which doesn't render for me on the archived page—possibly a skill issue), not for any of the surveys or studies providing the numbers which are used in said homework. They arrive at a 0.2% rate by multiplying a 10% reporting rate by a 2% false accusation rate. Yes, it really is just "0.1*0.02 = 0.002 = 0.2%". Where did they get those two factors?
- The 10% reporting rate is assumed based on two studies, one of which (UK, 2006) estimated a reporting rate between 5% and 25% and the other which covered two years in the US and estimated 49% (2010) and 27% (2011). I don't know how one bases an assumption of "Let's go with 10%" on those numbers, none of the linked studies are accessible via either the archive or copy-pasting the URL, and since the project which hosts the page is "a campaign to bring sexual violence out of the closet and lift survivors to their full potential", this has "just trust me bro" levels of credibility. Regardless, if God were to descend from the heavens and reveal exactly what percentage of rapes result in an accusation, it would be irrelevant to the false accusation rate, because a false accusation requires there to be an accusation in the first place.
- The 2% false accusation rate is the lower bound of an estimate (2%–8%) from what I believe is a US study from 2009 (a different year and/or country than the studies which were mostly ignored to, charitably, "estimate" the 10% reporting rate figure) hosted here by a different, but aligned, activist organization.
This is certainly one of the methodologies of all time.
Source 2 shows their homework by skewering the very survey which Source 1 used for its 2% false accusation rate, in much the same way that Scott Alexander did in 2013. Both (correctly, as we claim to grant the presumption of innocence in the relevant jurisdictions) bother accounting for inconclusive claims in the numerator and denominator, but I recommend Scott's analysis since it also grapples with selection effects regarding the non-reporting factor.
Regardless of how believable you find the analyses (Source 2 or Scott's; I wouldn't dignify Source 1 with such a characterization), they are at least operating on the level of evidence relevant to assessing accusations, whereas Source 1 is not.
If I am being overwhelmingly charitable, I would say that taking claims made by activist organizations (of any stripe) at face value is unwise, and that if you want to use them in your reasoning, the fine people at Morton have a variety of products suited to this purpose.
If I am being less- but still far from un-charitable, I would say that I suspect you read neither of those sources (particularly Source 1) and just wanted to slap up a couple numbers to make a "but who could say, really" sort of point. One of the reasons I might suspect this is that multiple others still bother to read things as well. I can't imagine that throwing out links without reading them would improve one's credibility, and while I'm just some rando who broke a decade-long streak of lurking during a coffee break, more established members like @phosphorus2 and @FCfromSSC expressing parallel sentiments should motivate exercising more care and deliberation when posting—it certainly would in me.
"My cousin is cool; you should meet her." is the "normal" way to say this sentence and conveys the gender; straight substitution of "woman" sounds similarly odd in place of "female".
"I met a nice woman at work the other day" does sound more natural to me when "woman" is substituted, and I take your point that it sounds at least a bit odd with "female" used here as a noun. I'm not sure why I should assume misogyny, though; none was conveyed.
"I like to hang out at that cafe because of all the females" is creepy for reasons entirely unrelated to diction, and I suspect you'd side-eye someone who told you, "I like to hang out at that cafe because of all the women" in much the same way. In fact, the value judgment you make in the next sentence accrues here with "women" in place just as strongly!
I think we're on the same page as far as how the flow of things sound when "female" is used as a noun (vs. an adjective), and this seems like an apt place to point out that "woman" has usurped (!) the use of "female" as an adjective in quite a few places despite (to my ears) sounding far clunkier. "Woman doctor" and "woman engineer" don't exactly roll off the tongue, and my brain goes "noun -> noun" internally every time I hear them...and yet, I have never once heard misogyny attributed to their speakers! I think what I'm trying to say is that both words are used validly as nouns and adjectives when referring to humans, we have aesthetic tastes for when a sentence flows and when it doesn't, and it doesn't make much sense to me to assume anything about the speaker's intent (edit: or beliefs).
I mean, I agree; I'm just not sure how this applies.
More options
Context Copy link