Making Mountains

[Making Mountains @pierreohyeah]

@Farbairn163238 @SandyofCthulhu @Nerdcognito I think you're right that it mainly goes that direction, so if you do this sorta thing, you ought to also allow low-fighting skill characters to bypass fighting checks if they come up with great in-universe fighting plans.

[Making Mountains @pierreohyeah]

@ZyMazza That makes sense. So, if the issue requires international coordination, I suppose those people would first be policy makers and then, failing that, people who influence their decisions (voters, lobbyists etc), and you should only reference large harms that are truly immanent.

[Making Mountains @pierreohyeah]

@ZyMazza Yes, that is one of the risks. Obviously another is that someone who IS in danger of imminent harm will be harmed if they don't know about the danger.

What do you think is a good way to talk about large harms to achieve a good balance of these risks?

[Making Mountains @pierreohyeah]

@RokoMijic @SamoBurja @apralky That assumes that the biases you're seeing were not primarily CAUSED by the type of RL chosen. I suspect background-internet-noise-bias is subtler and differently-flavoured from what we see with current chatbots, and likely harder to extract alpha from with simple heuristics.

[Making Mountains @pierreohyeah]

@valigo Casey is correct, but there's a subtlety: CSS is badly designed, meaning it's hard to know inside-out, meaning most people have a fuzzy understanding of a subset of it. Knowing assembly is harder than knowing a CSS subset fuzzily, but easier than the inside-out CSS knowledge.

[Making Mountains @pierreohyeah]

@rfleury @krystalball Hmm, that makes sense. Do you have any thoughts about whether the best way to restore freedom of speech is a) restore the old frame, b) adopt another existing frame that can sustain it or c) develop and popularise a new frame (or something else entirely)

[Making Mountains @pierreohyeah]

@rfleury @krystalball Not to mention that if you restrict your idea search-space too much in pursuit of pure power politics, you might find yourself outcompeted in power politics by an entity that explores a greater diversity of ideas more freely.

[Making Mountains @pierreohyeah]

@rfleury @krystalball To me, the ethical frame of free speech having high value is a useful one, though; it encourages more exploration of idea search-space. If everyone abandons the frame for pure power politics, it seems we lose something. Surely some balance is possible?

[Making Mountains @pierreohyeah]

@maplemapleart @TylerGlaiel Likely he meant somewhere between 4-14 hours of work, depending on which 2 days you pick. If you only have 30 mins a day, pick something you can get playable in an hour. Don't spend 28 days making a prototype only to find it sucks (assuming you don't lose motivation day 9).

[Making Mountains @pierreohyeah]

@TheGingerBill Why do you think reason is not an empirical entity? It seems like we sometimes infer reasoning from observations of behaviour. Are all such inferences misguided? I grant that some are, but it feels like if ALL are, then we cannot know if other minds are reasoning.

[Making Mountains @pierreohyeah]

@nothings If the person spending time fixing your bug is one of the handful of people willing and able to make something of the impact-scale of the stb libraries, you better be damn sure in your skill and judgement before you insist that they must be the one who is wrong.

[Making Mountains @pierreohyeah]

@FreyaHolmer "Lerp, or, *30 minute explanation of Lerp* for long-" is also valid, so long as your explanation is a single run-on sentence. Definitely a power move.

[Making Mountains @pierreohyeah]

(made you look!)

[Making Mountains @pierreohyeah]

Following my near-death experience with that monster, a lot of people have been asking how I escaped Minos' labyrinth, so gather round people, get comfortable, and let me tell you: 🧵

[Making Mountains @pierreohyeah]

@TodePond By gum, that just might do it!

[Making Mountains @pierreohyeah]

@TodePond Method 1: encode your entire biological system in a falling sand simulation, such that left/right decisions are made by each bit in the sequence. Swap consciousness with this simulation while no-one's looking.
Method 2: purchase 128 human-sized cactuses, plus 1 cactus suit, then-

[Making Mountains @pierreohyeah]

@rfleury @ccreikey @jpeskyy @SunsetLearn Perhaps some opinionated compiler flag that is "learning mode", which enables more heuristic checks or runtime logging, like "are you allocating a lot? Arenas exist". Even just a few of these could "educate away" a bunch of these memory bugs, without the cost to iteration speed.

[Making Mountains @pierreohyeah]

@rfleury @ccreikey @jpeskyy @SunsetLearn I like your way of thinking about this; It's good to solve certain bug classes, but the question is "where should we solve them?". Instead of trying to solve lack-of-understanding by disabling useful features, perhaps we could have learnability as a language design priority.