Could be worse
27,618 karma · joined November 20, 2013
Availability is itself a meaningful feature for curating the feed.
Sometimes the comment section rescues things, other times it's a decision between two competing submissions, where—all else being equal—the accessible version ought to win.
Right, remember how in ~2023 "prompt engineering" was ostensibly the critical thing everybody "needs to learn now or get left behind"?
I feel these tools—or at least the business-model behind them—have a pattern: Basic adoption is easy, while protecting yourself against their flaws is a shifting target where expertise goes stale.
A good analogy might be a very smart person with a gambling addiction.
Knowing they ought to quit doesn't mean they can quit. Their intelligence becomes invested in the act of gambling itself, and almost no amount of reasoning can help them exit from a situation they didn't reason themselves into.
This makes me think of user-avatar scenarios, like systems that automatically make an abstract design linked to an anonymous user's IP.
If the possibility space is big enough, old heraldry rules could be applied the same way—though old families might not appreciate seeing "their" design randomly appropriated.
Yet everyone assumes that, like before, there must be a reason.
Worse, we can't tell apart real decisions choices from the dream-machine side effects.
Except a little worse, since they were raised alone in a library, act mostly the same, and have harsh limits on personal growth.
Hold up, let's not oversell things here. GPL/Copyleft just means that you have to share source if you share binaries. However big companies exploit the digital commons differently now.
Instead they run the important pieces remotely in their data-centers, and release neither binary nor source. If they do a client binary (e.g. phone apps), they can just sprinkle in some naive encryption and some "copyrighted content" on the server. Then if you make a custom client--or even document how to do it--you might go to jail. [0]
My stance was that far too many of our support tickets involve someone taking a screenshot, and we'd want that information in any such screenshot... Especially if the network was so bad that it couldn't reach us any other way!
Everybody wants to be the ones building cheap slop, hoping somebody else makes protections and repairs...
The context is: "I'm using the premier web-search engine of the internet to search for things on the internet the same way that zillions of people have searched for things on the internet for 20 years."
It's not missing, Google simply chose to build something new that will ignore it because they're trying to alter the relationship.
I've always felt that "proper" and responsible LLM use for search [0] would be two boxes: You can describe what you want in the first, and it'll propose search terms in the second, and then those get executed normally.
Yes, the "average user" [1] might not usually care about the second box... Until they need to because the query/results are wrong.
Showing them in tandem means:
1. Users are at least capable of learning through exposure.
2. Users may realize a key term can be added which the model could never have guessed.
3. Users may recognize a term in there that doesn't make sense, allowing them to detect a translation error.
4. If good search-terms leads to a bad outcome, it is possible for someone to report and diagnose it, rather than a fully black-box mystery.
_____
[0] Not just for websites, but also things like internal business software, or SQL queries.
[1] The average that might not exist. ( https://www.thestar.com/news/insight/when-u-s-air-force-disc... ) There are some features that everybody needs, just at different times.
As for the other direction... I imagine that corrupt officials would prefer to have the deniability.
I lived there, Mister Confidently-Incorrect!
It sounds like you're saw Singapore Fun Facts once, and in your mind it permanently excluded any other place from having similar features.
* Like I said, Hong Kong does have a significant sovereign wealth fund that ranks in the top 10. [0]. It doesn't stop mattering just because Singapore runs a bigger one.
* Here's a article from 2021 with a nice stacked-graph showing 45% public housing [1], close to the half I mentioned.
Not sure how much more I can say without spoiling it for people who haven't played.
For example, is it possible that they were coerced into doing something state-actor-y for the federal government, but they're not allowed to mention it because their speech has been stifled with a National Security Letter? (I know some lower courts ruled them unconstitutional in the past, but I'm not sure what the current status-quo is.)
"Oh, we're so sorry, oh most-honored and wise and ethical Executive Branch, we tried to cooperate, but somehow that Judicial Branch just started telling us we can't..."
Not from any wrongdoing on Toyota's part, but a combination of it being cheap, reliable, repairable, etc.
I'm convinced that LLM-generated comments accelerate this problem, because they embed twists that a human didn't choose, and then those affect what gets generated next.
Much the same way LLM-characters don't do so well at answering "I don't know", there's a problem of them failing to cull context.
I'd like to emphasize that it's not just "don't trust it when it says to rm -fr", but even a fully constrained boolean output can be negligently-trusted if it's hooked up to something impactful.
People suck at "good enough". All major engineering disasters were something "good enough" for somebody. People burn down their houses with "good enough" electrical work.
Remember when everyone was talking about treating computers as cattle rather than pets?
This seems like a weird inversion. Programs as pets instead of cattle. Toxoplasmosis, ahoy...
If 1 guy with $16 billion turns into 16 grandkids with (inflation adjusted) $1 billion endowments each, it sounds to me like the family fortune was preserved extremely well, the family is simply bigger.
The puzzle game is adventure-adjacent in that it involves walking-around exploration, carried items/keys, locked doors with switches, exchanging items, and bridges.
Topologically, it is a network of 3D rooms connected by "doors" that are actually non-euclidean portals. You walk around, but you can also change the direction of gravity 6 ways, and--here's the kicker--every room tiles infinitely with itself.
So while I won't deny it's easy to get stuck in some areas of the room-network trying to remember how to get to another, individual rooms are 3D yet also havea a quirky way of constraining your random walk.
[0] https://store.steampowered.com/app/473950/Manifold_Garden/
This is, unfortunately, easy to ruin that with partial mods, breaking your immersion when two objects from different universes are in the same room.
Once you start looking at even more factors, you start reinventing a regular DNA-matching test.
[0] The "Birthday Paradox" makes the odds of confusion even higher with more siblings.
Even if we assume your body has a foolproof is_same_species() test, your neighbors' cells may:
1. Be healthy, but have the wrong priorities or behavior, so that they start doing the wrong tasks or sending the wrong messages in the wrong places. [0]
2. Be healthy, but they're just a creepy flesh-mask a parasite has constructed around itself.
3. Be individually hijacked by viruses, and your body won't have the same baseline information to tell that they've deviated from your neighbor's standard.
4. They may be fully human, but gone feral into a transmissible tumor [1].
____
[0] This is actually a risk from pregnancy, as overambitious fetal cells leak into the mother.
[1] https://en.wikipedia.org/wiki/Clonally_transmissible_cancer