domain:city-journal.org
MaiqTheTrue
Renrijra Krin
12m ago
But then why aren’t the more upscale places and homes more colorful? If anything, they’re much more neutral toned than the middle and lower class based places.
My theory is that somehow color got associated with low class or cheap. In order to not look cheap, you do neutrals.
SnapDragon
13m ago
I'm also not allowed to use the best models for my job, so take my advice (and, well, anyone else's) with a grain of salt. Any advice you get might be outdated in 6 months anyway; the field is evolving rapidly.
I think getting AI help with a large code base is still an open problem. Context windows keep growing, but (IMO) the model isn't going to get a deep understanding of a large project just from pasting it into the prompt. Keep to smaller components; give it the relevant source files, and also lots of English context (like the headers/docs you mentioned). You can ask it design questions (like "what data structure should I use here?"), or for code reviews, or have it implement new features. (I'm not sure about large refactors - that seems risky to me, because the model's temperature could make it randomly change code that it shouldn't. Stick to output at a scale that you can personally review.)
The most important thing to remember is that an LLM's superpower is comprehension: describe what you want in the same way you would to a fellow employee, and it will always understand. It's not some weird new IDE with cryptic key commands you have to memorize. It's a tool you can (and should) talk to normally.
MollieTheMare
16m ago
see for example the Statue of Liberty poem about bringing America the poor and hungry and persecuted.
I do wish people would not truncate the stanza:
"Keep, ancient lands, your storied pomp!" cries she
With silent lips. "Give me your tired, your poor,
Your huddled masses yearning to breathe free,
The wretched refuse of your teeming shore.
Send these, the homeless, tempest-tost to me,
I lift my lamp beside the golden door!"
(Emphasis mine) Sometimes people even truncate the poem mid line "Your huddled masses." There's not even a comma breaking the sentence there! Critically she doesn't say send me all. Her command for who to send does not require nobility, but does require carrying an essential notion of liberty with you. It is inscribed on the Statue of Liberty after all, not the Statue of Unlimited Open Boarders.
There's not an entirely negligible portion of the population that is fine with even fairly generous immigration policy. They might prefer, though, if the plan is to vote for the same shit policies that you are fleeing from that you do not come to the US.
TheAntipopulist
Formerly Ben___Garrison
17m ago
and they simply are not good at programming
At @self_made_human's request, I'm answering this. I strongly believe LLMs to be a powerful force-multiplier for SWEs and programmers. I'm relatively new in my latest position, and most of the devs there were pessimistic about AI until I started showing them what I was doing with it, and how to use it properly. Some notes:
-
LLMs will be best where you know the least. If you're working on a 100k codebase that you've been dealing with for 10+ years in a language you've known for 20+ years, then the alpha on LLMs might be genuinely small. But if you have to deal with a new framework or language that's at least somewhat popular, then LLMs will speed you up massively. At the very least it will be able to rapidly generate discrete chunks of code to build a toolbelt like a Super StackOverflow.
-
Using LLMs are a skill, and if you don't prompt it correctly then it can veer towards garbage. You'll want to learn things like setting up a system prompt and initial messages, chaining queries from higher level design decisions down to smaller tasks, and especially managing context are all important. One of the devs at my workplace tried to raw-dog the LLM by dumping in a massive codebase with no further instruction while asking for like 10 different things simultaneously, and claimed AI was worthless when the result didn't compile after one attempt. Stuff like that is just a skill issue.
-
Use recent models, not stuff like 4o-mini. A lot of the devs at my current workplace tried experimenting with LLMs when they first blew up in early 2023, but those models were quite rudimentary compared to what we have today. Yet a lot of tools like Roo Cline or whatever have defaulted to old, crappy models to keep costs down, but that just results in bad code. You should be using one of 1) Claude Opus, 2) ChatGPT o3, or 3) Google Gemini 2.5 pro.
zoink
18m ago
fight the dandelion infestation on your front lawn again
I've never had a problem with broadleaf weeds. Are you against using herbicide? I find spraying the whole yard is a waist, I spot spray broadleaf's with 2,4-D. Hit the dandelions before they go to seed and I just have to walk the lawn two to three times.
LateMechanic
21m ago
conventional theory is that borrowing is less inflationary than money printing. Do you disagree?
Definitely, primarily because the separation is just incoherent.
In the conventional view, you call central bank reserve account balances "money" which is evidence of "printing", while you call treasury t-bills "debt" which are evidence of "borrowing". But they are both are just government liabilities which promise to pay nothing in redemption other than other government liabilities, and which pay the policy rate of interest. To make it funnier, these are literally both types of accounts that the central bank runs on their books, because the Fed does the banking on behalf of the Treasury. You can call them checking & savings accounts at the government bank, although they pay the same interest, and you can freely swap between these accounts (you're never 'stuck' holding them). To call only one of these accounts "money" takes a special kind of incoherence.
Physical paper/coin money cash is just a bearer receipt version of the electronic reserves. That's the only government money that doesn't pay interest now, as a tiny fraction of the money supply, held for some types of convenience.
The MMT people often try to be more precise and avoid the word 'money' because it can lead to confusion sometimes, but I just plow ahead and risk it. The clear definition of money from what we use now and even throughout history is "transferable credit". So we can just talk about money as IOUs. Someone issued a financial liability, which someone else gets to hold as their financial asset. A bank balance is your valuable asset because it's the bank's debt. A reserve account balance is the bank's valuable asset because it's the central bank's debt. A $5 bill is your asset because it's the central bank's liability. Always credit-debt relationships, the issuer owing the holder.
So I'm perfectly happy to call the outstanding reserve balances part of the "government debt", just like I call treasury bonds/bills/securities "money". These are all both money and debt.
So issuing/printing new reserves or issuing/printing new t-bills, which are nearly perfect equivalents, no, they would not have different inflationary impacts. The size of the government deficit is the size of the amount of money being printed, and it always has been.
Many mainstream economists like Summers & Krugman got to this a bit late, in the 2010's, when they reconsidered that 0% government debt instruments were essentially 'money'. So they started reconsidering what they thought they knew about 'monetizing the government debt' and QE, etc. I'm not sure if they ever caught up to the fact that in 2008, central banks switched to paying interest on reserves directly, making 'money' look just like securities even when we're not in a 0% interest environment.
So far, youve shown that just printing the deficit gets you the same effect as cutting expenditures to balance the budget, with some inflation along the way.
But in the previous model, there is no ongoing deficit in the equilibrium youre inflating towards.
I'm sorry I couldn't really parse what you were saying in these 2 paragraphs. That balancing the budget was the same as running a deficit, or somehow it had the same outcome in a particular way?
And in what way is the equilibrium of your strategy better than just spending enough for full employment, and raising taxes to balance the budget?
Well if anyone is ever saving money, then by identity someone else must be dis-saving an equivalent amount (running down prior savings, or just issuing debt). Because it's all zero sum. In aggregate, what we find is that people like to accumulate monetary savings over time (even with populations that aren't growing I think). So unless your hypothetical has some way to stop that, that saving is a leakage in aggregate demand which will not let you get to full employment without the government being in deficit to supply the desired savings. A government balanced budget means they are draining out exactly as much money as they're injecting in to the economy.
Corvos
26m ago
Thanks, that’s good to know! LG seems to be the best brand, although Sony have some nice ones.
Lykurg
We're all living in Amerika
27m ago
None of my pseudonyms tries to project a unique personality. They speak with exactly the same personality—just on different topics.
It might actually be a good sign if they did. If even different fake online faces are only shaped by their own incentives and not each other, then your real life face is propably safe.
5434a
30m ago
I've got to say I find your all-encompassing scepticism towards Reddit a smidgen excessive - they were only saying the same thing that you said here in response to posts that echo what you've heard people say to your face, only at scale.
Practically everyone here is or was a Redditor of relatively long standing. Reddit is not a high quality forum populated exclusively by intelligent and thoughtful people, but it's not completely devoid of them either. Eternal September began a long time ago.
Lykurg
We're all living in Amerika
36m ago
I did wonder if the name derived from "face". I didnt think of a letter being alphabetical numbering without any indication its not part of the word.
atelier
Jimmys: unrustled
38m ago
I sometimes wonder if I'm the last user who still goes to each user's personal page a specific subreddit when I want to see something.
MaiqTheTrue
Renrijra Krin
40m ago
I don’t see this as all bad, to some degree everyone is acting. You don’t curse in front of grandma even if you do in other places. You don’t dress the same for work as you do to just hang out. As long as the character you play is something of a decent human being, it’s probably not harmful.
atelier
Jimmys: unrustled
40m ago
Yeah I always categorized it as a name like Louis Freeh, a former FBI director
TheAntipopulist
Formerly Ben___Garrison
43m ago
What do you do to get AI help with a large code base rather than a toy problem?
Two things mainly:
-
Have a good prompt that has the nuances of the crappy, antiquated setup my work is using for their legacy systems. I have to refine this when it runs into the same sorts of errors over and over (e.g. thinking we're using a more updated version of SQL when we're actually using one that was deprecated in 2005).
-
Play context manager, and break up problems into smaller chunks. The larger the problem that you're getting AI to do, the greater the chance that it will break down at some point. Each LLM has a certain max output length, and if you got even close to that then it can stop doing chain-of-though to budget its output tokens, which makes its intelligence tank. The recent Apple paper on the Tower of Hanoi demonstrated that pretty clearly.
TheAntipopulist
Formerly Ben___Garrison
57m ago
Twitter and Reddit both allow you to sort chronologically. I've just naturally stopped using most of the ones that don't have an option like that, such as Facebook and TikTok (I never got into TikTok in the first place, I bounced off hard). I also don't think "the algorithm" is necessarily always bad -- Youtube's recommended videos have exposed me to some truly excellent creators like Montemayor over the years. Sometimes I'll watch lower quality stuff like whatifalthist and my recommended will be populated by garbage for a bit, but that resolves itself after a week or so, and I could probably speed it up by marking those videos as things I don't want.
Ublock Origin blocks basically all ads, and is quite effective. I haven't noticed shills posing as users to be that much of a problem outside of stuff like porn.
thejdizzler
59m ago
Can you link the video? Sounds like something I need to hear.
WhiningCoil
1hr ago
fight the dandelion infestation on your front lawn again (fuuuuuuuuuuuck)
I'm waiting for it to stop raining long enough to put down something for the crabgrass currently threatening to destroy the overseeding I did last fall. I feel your pain.
hydroacetylene
1hr ago
My point was that any identity which can totalise will absorb literally all common symbols, of which the rainbow is one(literally every culture has given it a special meaning- the pagans thought it was a bridge for the gods to access the world, Christians think it’s a sign of God’s mercy, LGBT thinks it’s about diversity of deviant sexualities, South Africans think it’s about multiple races working together in harmony, etc etc), and this is unconnected to general cultural design trends.
gattsuru
1hr ago
Fair. My impression is that LASD has a lot of the same issues, but I'll admit I've got less current evidence on that.
gattsuru
1hr ago
I've tried turning off visibility of things like individual post scores, but that does just risk you changing to focus on notifications, instead. And given the extent twitter has driven people completely bonkers, that might be worse than the karma farming. There's always been worries about the masks we wear molding the face -- and even some theories about using that to improve ourselves -- but having the masks get molded in turn is Not Great Bob. And then what exactly it seems to be driving even the boring people toward is kinda disturbing.
You can do some efforts to de-algorithmify yourself, but that's only going to get the worst of it, and maybe not even that. And it's pretty incompatible with having a career or even a renumerative hobby online. Even some offline small business work is becoming increasingly hard to kick off without it. I'd like to advocate some level of in vino veritas, but a) I don't drink, and b) that doesn't seem to work great for those who pick it up. Trying to actively avoid collecting enough of a following maybe helps? But I dunno if that's just because I wouldn't notice the microscale examples of the trend, either.
The one bright spot is that Flanderization does, at least in part, reflect another trait specific to media, not people qua people. Ted Flanders didn't turn from slightly-religious neighbor into a fundie just because time's arrow flew, but also because the shows writers needed something new for each episode. "Simpsons Did It" is a problem for South Park, but it's also an issue for The Simpsons itself; even if most viewers won't recognize the psuedorerun, the show's staff and a lot of the commentariat will. If you have to get a column out for your tech column the weekend and three videos M/W/F, you start diving into this sorta A/B-to-death-testing because you don't have anything else, and the content doesn't have that much to start with.
For normal people, that doesn't quite work that way. Yes, history rhymes, and I'm probably one of the worst people on this site when it comes to bringing up ancient history from the long-ago days of two years ago. But anyone that hasn't let the mask embed into their skull can and probably will find something new because the world is filled with new stuff. Get a hobby, touch grass, fight the dandelion infestation on your front lawn again (fuuuuuuuuuuuck), talk about cooking.
There are plenty of tasks (e.g. speaking multiple languages) where ChatGPT exceeds the top human, too. Given how much cherrypicking the "AI is overhyped" people do, it really seems like we've actually redefined AGI to "can exceed the top human at EVERY task", which is kind of ridiculous. There's a reasonable argument that even lowly ChatGPT 3.0 was our first encounter with "general" AI, after all. You can have "general" intelligence and still, you know, fail at things. See: humans.
More options
Context Copy link