bitshift's recent activity
-
Comment on Paying for AI: Have you found it to be worth it? in ~tech
-
Comment on The future of forums is lies, I guess in ~tech
bitshift Link ParentDid you directly email one of the admins an application or something? I got my account basically from posting "me too" or something equally low effort on one of the regularly-posted "who wants an...Did you directly email one of the admins an application or something? I got my account basically from posting "me too" or something equally low effort on one of the regularly-posted "who wants an account" Reddit threads.
-
Comment on The future of forums is lies, I guess in ~tech
bitshift Link ParentI presume MMOs are trying these things to some degree. If I were running an MMO, I'd be concerned not just about catching bad actors at account creation time, but also catching legit users who get...I presume MMOs are trying these things to some degree. If I were running an MMO, I'd be concerned not just about catching bad actors at account creation time, but also catching legit users who get tired of grinding and install a bot/cheat/etc.
the more realistic the bots are made the less lucrative they become to run
Thanks for putting that into words so concisely. I hope it stays true for a while!
-
Comment on The future of forums is lies, I guess in ~tech
bitshift Link ParentI like the idea of fighting LLMs with LLMs. It's not a perfect solution, but as the technology gets cheaper, it benefits the cat too, not just the mouse. In general, I think spam filters and other...I like the idea of fighting LLMs with LLMs. It's not a perfect solution, but as the technology gets cheaper, it benefits the cat too, not just the mouse.
In general, I think spam filters and other behavior-observing approaches are underutilized. Look at the behavior that aphyr documented: just enough fakery to get in the door, and then the first post is something an LLM could easily flag. And actually, an LLM would be overkill. Even a simple Bayesian filter would know most users on that site do not write all about batteries in their first post.
Granted, a determined spammer could continue the fakery for several posts and try to build a persona. But my point is that by observing behavior over time, you've successfully raised the bar for spammers in a way that mostly doesn't impact users. A spammer can automate some of the work, but they have to make sure their bot never slips up in any of its interactions. A user just has to be themselves.
-
Comment on ASCII Moon: View and cycle through the Moon's phases, rendered in ASCII art in ~space
bitshift LinkI am not the first person to complain about this, but the Moon is drawn in the wrong shape. It has a big circular cutout right now, like somebody took a bite out of it; no spherical object is...I am not the first person to complain about this, but the Moon is drawn in the wrong shape. It has a big circular cutout right now, like somebody took a bite out of it; no spherical object is shaded like that.
Also, it's hardly even ASCII art. If you zoom in on the light/shadow boundary, you'll see ASCII characters that are multiple colors: e.g., a forward slash
/whose top one-third is light blue, with the bottom two-thirds dark blue. It evokes an ASCII art aesthetic. But you can't retype it. You can't copy/paste those colors, not even as rich text. You can't display it as text on a retro computer, or in an email, or in a Word doc.I know there are more serious problems in the world, but right now I feel most strongly about this one.
-
Comment on South Korea banned dog meat. So what happens to the dogs? in ~food
bitshift (edited )Link ParentYou could even legislate that! No new business licenses after such-and-such date. Add additional requirements to prevent acquisition of new dogs (such as mandatory spaying/neutering, like you...I doubt anyone new will be entering the business?
You could even legislate that! No new business licenses after such-and-such date.
Add additional requirements to prevent acquisition of new dogs (such as mandatory spaying/neutering, like you said), and you're done. The businesses are now guaranteed to shut down. Don't outlaw the sale or consumption of dog meat because that causes a sudden shock that leaves a bunch of people in the lurch. Just cut off the supply and let the farmers wind down their own businesses in an orderly fashion; it would save the same number of dogs' lives, just with less turmoil and ill-will.
(It goes without saying that this approach requires adequate enforcement of the law—but so does the existing solution with the hard cutoff of the entire industry.)
-
Comment on Beware tech career advice from old heads in ~comp
bitshift Link ParentBefore every interview I conduct, I tell the candidate, "This is open book, so if you want to look stuff up or ask me questions, go for it." I do have candidates take me up on this, but it happens...Before every interview I conduct, I tell the candidate, "This is open book, so if you want to look stuff up or ask me questions, go for it."
I do have candidates take me up on this, but it happens less often than you'd think. Frequently I'm the one who's nudging them to look stuff up: "Hey, so you're absolutely right that this depends on what kind of data structure Python lists are under the hood… how would you find that out?" I wonder if all their past interviews have conditioned candidates to think they're admitting weakness by saying, "I don't know, let me Google it."
I guess what I'm trying to say is, if you ever have an open book interview, please—for my sake—use the gargantuan living encyclopedia your computer is plugged into.
-
Comment on Why Dua Lipa is so good at asking questions in ~books
bitshift Link ParentYour thought experiment reminded me of actress Hedy Lamarr, inventor of frequency hopping: It hits differently because decades have passed. But still… be a Hollywood bombshell, get increasingly...Your thought experiment reminded me of actress Hedy Lamarr, inventor of frequency hopping:
At the beginning of World War II, along with George Antheil, Lamarr co-invented a radio guidance system for Allied torpedoes that used spread spectrum and frequency hopping technology to defeat the threat of radio jamming by the Axis powers. However, the technology was not used in operational systems until after World War II, and then independently of their patent.
It hits differently because decades have passed. But still… be a Hollywood bombshell, get increasingly bored with all the shallow roles, start some engineering projects on the side? Your thought experiment wasn't too far off.
-
Comment on Russian losses hit two grim milestones ahead of Ukraine war anniversary in ~news
bitshift Link ParentI wonder about that, though. Russian commanders are lying to their higher-ups so much, even the Russians might not know the precise extent of their own losses. I'm sure some new info will come out...I wonder about that, though. Russian commanders are lying to their higher-ups so much, even the Russians might not know the precise extent of their own losses.
I'm sure some new info will come out after the war. But I'm guessing the existing third-party assessments won't change much.
-
Comment on reCAPTCHA: 819 million hours of wasted human time and billions of dollars in Google profits in ~tech
bitshift Link ParentAlso, in retrospect I realize my words were ambiguous: I hope it was clear that the "they" was BoingBoing, not you! If not, my apologies.Also, in retrospect I realize my words were ambiguous:
Because they couldn't be bothered to link to the source material
I hope it was clear that the "they" was BoingBoing, not you! If not, my apologies.
-
Comment on reCAPTCHA: 819 million hours of wasted human time and billions of dollars in Google profits in ~tech
bitshift LinkBecause they couldn't be bothered to link to the source material: here's the paper on arXiv they're talking about. My (perhaps unfair) summary: We surveyed some people and discovered that CAPTCHAs...Because they couldn't be bothered to link to the source material: here's the paper on arXiv they're talking about. My (perhaps unfair) summary:
- We surveyed some people and discovered that CAPTCHAs are frustrating.
- We did some napkin math and realized Google doesn't offer reCAPTCHA out of the goodness of their hearts.
- We reviewed the literature and found some good state of the art attacks on CAPTCHAs.
- Therefore, reCAPTCHA should be abolished.
To me, the evidence in support of their conclusion feels… underwhelming?
And I say that as someone who wants their claims to be true: I would love for user-tracking CAPTCHAs to be strictly worse than the alternatives. But reCAPTCHA still adds friction for bots (not all bots are going to use state of the art attacks), and I don't feel they've made a compelling argument that reCAPTCHA is worthless.
(Perhaps my ideal study would be, "We replaced reCAPTCHA with X for a year, and saw no statistically significant increase in bot traffic.")
-
Comment on European Union orders X to hand over algorithm documents in ~tech
bitshift Link ParentAlternatively: even if they're honest and make a reasonable effort to comply, what's to guarantee the result will be useful? "Here's a folder full of Word docs from when we designed the...Alternatively: even if they're honest and make a reasonable effort to comply, what's to guarantee the result will be useful?
"Here's a folder full of Word docs from when we designed the thing—though we did change a few things at implementation time. Here's snapshots of a few Git repos. Here's PDFs of some internal wiki articles from last year. And here's the AI model weights (we didn't hold on to all the training data, sorry)."
I don't know the exact scope of what X is legally required to provide. But whatever they hand over, it's going to be messy. It won't be a self-contained box with little brass gears on the inside and "THE ALGORITHM" emblazoned on the outside.
-
Comment on Proton CEO tweets support for Donald Trump's Department of Justice pick and the US Republican Party in ~society
bitshift Link ParentMy two cents: boycotts have a cost (can be money, time, energy, etc), and problems occur when people act like they're free. If it were just money, it would be more concrete. Let's say the local...My two cents: boycotts have a cost (can be money, time, energy, etc), and problems occur when people act like they're free.
If it were just money, it would be more concrete. Let's say the local food bank has run out of food and is having to turn down hungry people, so they put out an urgent call for donations. The local community is chattering things like "I just gave $100!" and "Hunger is such an important issue." Now for you, maybe money is tight, and you personally decide against donating. That ought to be okay! It would be out of line for someone to criticize you for "supporting hunger" or "this is a matter of right and wrong" or "if you actually cared…"
If the money isn't there, the money isn't there, and nobody should be making anybody feel bad. Same thing goes if the time/energy/spoons aren't there.
-
Comment on Matt Mullenweg deactivates WordPress accounts of contributors planning a fork in ~tech
bitshift Link ParentThank you for respectfully laying out your viewpoint, even though it's not the "popular" one. The world's complicated. Matt can simultaneously have been a net positive for OSS and Internet...Thank you for respectfully laying out your viewpoint, even though it's not the "popular" one.
The world's complicated. Matt can simultaneously have been a net positive for OSS and Internet publishing over the last two decades—while also being the wrong steward of the project for the next few decades.
Same goes for WPE. Are they saints? No. Do they have a financial interest? Yeah. But that doesn't mean they're incorrect in criticizing Matt's recent swings in leadership.
-
Comment on <deleted topic> in ~society
bitshift Link ParentI hope you're right! And I say that not in a sarcastic, pessimistic way, but with genuine optimism. May he waste both his own time, as well as that of everyone around him who thirsts for power.I hope you're right! And I say that not in a sarcastic, pessimistic way, but with genuine optimism. May he waste both his own time, as well as that of everyone around him who thirsts for power.
-
Comment on What’s your “I didn’t know I needed that” item? in ~life
bitshift Link ParentConsumables are great gifts. You don't have to worry, "Do they already have it?" or, "What if Aunt So-and-so gives them the same thing?" It will get used up. If two different people gifted me nice...Consumables are great gifts. You don't have to worry, "Do they already have it?" or, "What if Aunt So-and-so gives them the same thing?" It will get used up. If two different people gifted me nice coffee, I'll be twice as happy.
-
Comment on Are ‘ghost engineers’ real? Seeking Silicon Valley’s least productive coders. in ~tech
bitshift Link ParentMildly related, but I'm reminded of this story of a naive attempt to measure productivity by code changes: -2000 Lines Of CodeMildly related, but I'm reminded of this story of a naive attempt to measure productivity by code changes: -2000 Lines Of Code
-
Comment on Bomb threats made against US President-Elect Donald Trump cabinet nominees in ~society
bitshift Link ParentPolitical violence is not new, but I 100% agree that each additional instance sucks for the reason you mentioned: violence normalizes violence. Somewhere I read a really good analogy of how our...Political violence is not new, but I 100% agree that each additional instance sucks for the reason you mentioned: violence normalizes violence.
Somewhere I read a really good analogy of how our minds go wrong with this. When we think of violence, we're wired to think of good versus evil, knights slaying dragons, stuff like that. When a dragon kills one of the townsfolk, that's because the dragon is evil. And when a knight kills one of the dragons, well, the dragon had it coming. Killing the dragon solves the problem.
But in reality, violence is less like a dragon and more like Godzilla. It's not glamorous. Every day he shows up and tramples several city blocks. Some days he tramples more, some days less. Sometimes a different giant monster shows up and fights Godzilla, and even more of the city gets destroyed than usual. Your life is now all about surviving between monster attacks. That's the new normal.
And the problem is whenever people commit violence, they think they're killing an evil dragon, when they're actually waking Godzilla.
-
Comment on Recruiting help for election day posters in ~creative
bitshift Link ParentThis is good, though there might be a clearer way to phrase it. The phrase "can't find out" could be misinterpreted as "You can't risk letting your spouse find out," as opposed to the presumably...Just remember, your spouse can't find out how you voted
This is good, though there might be a clearer way to phrase it. The phrase "can't find out" could be misinterpreted as "You can't risk letting your spouse find out," as opposed to the presumably intended (and much more calming) "Your spouse is unable to find out."
Other than that issue, I love this statement. It's simple, and (unlike the provocative alternatives), it's hard to dispute. Heck, it's very nearly a nonpartisan statement! Obviously there's a partisan intent given the context of this thread, but it's not printed on the sign; I could hold that sign while claiming to be a Trump supporter, and you'd have no way of knowing for sure.
-
Comment on Trans activists release 6,000 crickets on transphobic LGB Alliance conference in ~lgbt
bitshift Link ParentNot the person you're replying to, but I know what my reaction would be: to ask, "How does this help good overcome evil?" Personally, I'd be afraid that pulling a stunt to disrupt the meeting has...Not the person you're replying to, but I know what my reaction would be: to ask, "How does this help good overcome evil?"
Personally, I'd be afraid that pulling a stunt to disrupt the meeting has a lot of potential downsides without a whole lot of upside. I'm not saying they were wrong to protest, or that every action in life has to have an attached cost/benefit analysis. What I'm saying is that I'm trying to do a cost/benefit analysis because that's my habit, and I'm scratching my head trying to come up with a long-term benefit.
Maybe I'm underestimating the value of raising social awareness? Or maybe I'm overly pessimistic about propagandists twisting the stunt into an argument for the other side?
The strongest argument I can come up with for the stunt is that it sends a message—"We will not be silenced"—where doing something that could get you in trouble is an integral part of the message.
I will second Claude projects. They sounded like a gimmick to me at first, but they've turned out to be super helpful so far: they eliminate the overhead of specifying the same context over and over with each new chat. You can definitely work around this manually—just ask really detailed questions. But it's so nice to upload (for example) a .txt of some software's help file and have it automatically just know all that.
You can also use AI to help generate context documents. Start a new chat: "Let's pretend you're a new person on my team. I'll start describing the project, and you ask questions about what you don't understand." Then have a back-and-forth conversation. After a while, you ask it to write an "onboarding doc" in Markdown, you do some editing on your own, and then you upload the doc to the project. It still takes effort; the AI can't magically know your situation, and you still have to type all the details. But you can do it in a more casual manner, and you don't waste time describing things the AI already knows.