This means giving up all the advantages of central banking and is generally only done informally by countries experiencing hyperinflation without capital controls. It's a marker of a failed state.
I remember in live TV debates for the 2014 Scottish independence referendum, the Yes leader insisted that Scotland could not be prevented from using the Pound sterling. It's technically true, but a very, very bad idea.
> This means giving up all the advantages of central banking and is generally only done informally by countries experiencing hyperinflation without capital controls. It's a marker of a failed state.
The US did not have a central bank until 1914. And there was zero net inflation from 1800-1914. The central bank introduced endemic inflation, which appeared immediately.
By design. Having a slightly positive inflation rate gives you a cushion against deflation, which tends to cause market crashes. A currency I can be confident will lose a couple percent yearly is more useful than one that might be worth wildly different amounts year to year.
If you scroll down a bit on this page, you can see the massive inflation/deflation spikes in ~10yr cycles that existed prior to central banking.
You're completely missing the point. The Fed did a fantastic job, because they could have given a pretty good estimate of that number in 1914. If you had asked someone in 1814 what a dollar would be worth in 1914, hell, 1824, they would have been guessing, and been wildly wrong.
Making sure that the nitwits stuffing their mattresses with dollar bills maintain their net worth is not the goal of our monetary policy, nor is a good goal. The goal is to ensure predictability.
To this end, we can model the price level at some point in time in a way similar to how we model compounding interest: P(t) = P(0) * (1.0 + r)^t. Take P(0) = 1.0, r = 0.032, and t = 2026-1914 = 112. Then, P(112) = 1.0 * (1.032)^112 = 34.05. We've had an average 3.2% rate of inflation in the price level over the past 112 years, give or take.
We can make certain assumptions that the rate of inflation won't be far off from this when we evaluate certain financial risks.
Just as in modeling adjustable interest rates for compounding interest, we can make r depend on t, and at that point, it becomes an ODE problem: dP/dt = r(t) * P(t).
This betrays a lack of understanding of nominal versus real debt dynamics. (And also taxes, but I'll admit that my upbringing has probably given me a unique perspective on Caesar and what it means for us to be able to use his money.)
Of particular note, debt instruments are denominated in nominal dollars, and they're paid back in nominal dollars, but what concerns the creditor is the real value of those nominal payments. Economic growth has this pernicious habit of pushing nominal prices upward, and if the money supply and credit system don't grow commensurate with the resulting increased demand for liquidity, the real burden of existing nominal debts can rise sharply and unpredictably.
This means that borrowers can find themselves underwater on, e.g., mortgages while the nominal obligations remain fixed, and banks will swiftly foreclose on them and tighten credit when considering their balance sheets. Many of the panics of the 1800's included a lot of this very dynamic.
Montenegro unilaterally uses the euro and it doesn't seem to otherwise be a failed state. Besides this being a "marker", what do you think are the actual problems caused? Like why does a small country need its own capital controls when there is a very stable currency available nearby?
The reason for your own currency is you can set your own interest rates to fit your local economy. If you are in a local recession while others are doing well you might want lower interest rates, while others need them higher. Both places are trying to make the same balance of inflation vs stimulating the economy - but they need different answers.
I'm not convinced it is worth it. Generally world economies are tightly tied anyway and so what is right for large currencies is close enough for everybody. The less coupled you are to the world the more important it is that you can be different.
Not significantly. Your point that only failed states decide to use another nation's currency is automatically a failed state is simply untrue. It can be a marker, but it's not an automatic mark. Montenegro is not a failed state.
They don't give up the advantages of central banking at all.
What they give up is political control of their bank. There are advantages to having political control, but often political control is abused - which is why failed states have given up on it as part of their efforts to rebound. The US and EU both have controls in place to limit the power of politicians from making changes for political reasons.
It would, and this has been an issue even with bigger eurozone countries, like Greece.
Basically the same pressure that would have adjusted your exchange rates instead adjusts how much of the fixed-rate currency exists in your country. With fluctuating rates the pain of a financial outflow is more evenly spread than with a government running out of money, unless the government adjusts taxes to compensate.
For context, 15 was some brony tf2 player who cracked deep voice synthesis in order to make My Little Pony characters say funny things. He then offered this for free on his site. It was never a clever branding thing, it was just his username.
It's funny seeing HN types being genuinely confused as to why someone would just make interesting things for the sake of making them, and then offer them freely for the good of others. He used to be on HN, there's good insights in his posts: https://news.ycombinator.com/user?id=15ai
Thanks for sharing. It is a classic. HN has only seemingly changed in that there seems to be more of those ‘types’ than ever. The TF2 community has always been a creative bastion too.
Oh man, this is a blast from the past. Played TF2 with him a few times, he hasn't been online on Steam for a long time, but apparently he just left AI behind before it turned into the dystopian techbro circlejerk it is now, and still continues to work on other things he likes today, which is great to hear.
It does feel like the "hacker" part of Hacker News is all but gone at this point, the only thing people are trying to "hack" anymore is their bank account balance. I will remember 15.ai as a landmark from the tail-end of the "people doing cool things just for the sake of doing them and showing it to others" era of the internet.
I have friends that still play with him, he's still active on the same servers with a different username. He's currently working on an AI-free artist marketplace:
I think the point with those supermarket cameras is to influence behaviour, not to surveil, I doubt they are even recording. When I was a kid (maybe in a worse area than you) there were CRT TVs by entrances of supermarkets showing a live feed of you walking in. People behave differently when they feel they are being watched, it's why they have creepy CCTV posters with big eyes on them. Probably the same reason for evil eye charms too.
Yes, this has been under consideration when thinking just how much raw footage would have to be stored for 4 or more self-checkout cameras for every shop. Probably multiple terabytes per day, per chain, just for London. There is also the classic 'plastic camera with blinking LED' CCTV installation, intended to deter theft in a more lo-fi manner.
As an aside, the artist John Baldessari made an installation in the Tate Modern where you were filmed walking into a room and the image was projected on the wall, but with about a 10 second delay, so that the feedback loop between seeing yourself 'on screen' and being able to influence it was completely broken. It was quite interesting to experience.
You only need 3-5 fps SD video and with motion sensing you only record when there is a customer at the self-checkout. Technically, it is absolutely doable.
About 8 or 9 years ago I saw a self checkout pos provider demo a camera/screen system that would do basic images processing constantly to try and take pictures when people are clearly visible. A couple of low res, low colour palette, cropped images per user are much much easier to store. This was at a security conference.
Fair enough, sounds like a reasonable compromise, but I still question the entire premise of having these cameras in the first place, especially since they do not film the area where items are being handled. Tallying a loss of stock to a specific person would seem to be really quite difficult to do with only a set of photographs of the faces of everyone who utilised the self-checkouts within a certain period.
Which is again, a fair question with a reasonable answer:
Stores have loss 'prevention' (more like tracking) software which the security team can use. They are able to attach images of people stealing, but its hard to identify a person from a camera 15ft away when they are bent over a shelf; so they attach images of them stealing and of them passing through choke points - e.g. checkouts and entrances.
Tying them together is laborious, and is currently in its own little model rush - it's a technique called re-identification.
However, the next question is why doesn't it happen in practice? Well the stores have found police to be uniquely annoying on that they won't arrest and the state won't prosecute despite clear evidence.
One of the most annoying quirks of LLMs for me is the insistence on, after being corrected, loudly noting what wasn't done. After getting it to change something stupid in code, it will leave a comment in the code boasting about how it doesn't do the stupid thing, which to future readers reads like an 'asbestos free!' label on a cereal box.
Lets assume one of two worlds - one where AI is constantly improving and one where the tooling is a dead end. In the first world the AI generated comments are no better than the commentary that would be generated in the next six months - in the second world the AI generated comments will obscure the human written comments.
In what world are AI generated comments actually value adds?
If the AI (or the person) ran into a non-obvious constraint, or some externality, or there’s a plan that just isn’t implemented yet- then the comment wouldn’t be able to be divined by the AI. But yeah just pointing the AI at something and saying “write me docs” will get you something at best equally good as what’s going to be available.
One of the reasons this is there is the "psychology" of the LLM performs better with these kinds of affirmations. If you've made a mistake once you're likely to make it again, so leave a note.
This indeed pollutes the code with a bunch of nonsense. One of the things I have it maintain instead is a "findings" document with lessons learned and such.
> If you've made a mistake once you're likely to make it again, so leave a note.
In my experience the comments behave as a form of prompt injection. The llm makes a conceptual mistake, writes it in a comment, and now subsequent agents make the same mistake!
I have a contrarian take on this: if most people working on the codebase are using the same model, then these comments serve a very useful purpose: they tell the model's future self that despite its tendency towards this solution, it is incorrect and here's why. It encodes useful context right when the model needs it next to avoid making the same error.
I imagine this style of commenting is partially responsible for model improvement in SWE work, a trained trait through feedback, since it makes the comments actually useful. Not to you or me, but to an LLM.
Some of these errors are obvious to a human, that's why we find the comments unnecessary. But if everyone's using LLMs and said model tends to make this error, then they are very useful. It's like a comment you would leave for another human to help them avoid a footgun. It's just a LLM footgun, not a human one.
I agree, but that information should be in the commit, which the model also has access to. Of course it's more token intensive to look through commits, but "good" models might know when some more research about a particular bit of code they find weird would be a good idea. Thoughts?
The problem is that the list of things one shouldn't do is infinite. The things one should do are narrow and get narrower along two independent axes of clarity/wisdom: system requirements and engineer seniority. LLMs overindex on any words given, so you only want to give them negative guidance around truly repeated, almost common-sense pitfalls. But in a large distributed system often the changes come from all different angles, and each agent will find it's own unique failure modes. Those comments will dilute critical context for diverse agents far more than they will help on average, at least for the systems I'm working on, given current claude code chattiness.
One way to clear some of the low-hanging slop is to just have a separate agent code reviewing and pruning any comments that don't stand on their own purely in the context of the diff, but it still doesn't catch it all.
If LLMs are any good, like many of you claim, they won’t need these comments. They will ingest the code and know what was done. They won’t need comments about what wasn’t done.
In Scotland in the early 2000s, there was a moral panic over a tonic wine called Buckfast, made in Buckfast Abbey by an order of Benedictine monks.
It's the kind of thing meant for your granny to have when sick. It contains caffeine and alcohol, and neds started drinking it because caffeine and alcohol both feel good (humans have known this for thousands of years).
It quickly became part of the local youth culture, and was quickly blamed for all the deep-seated anti-social behaviour issues prevalent in the region. It was known as 'wreck the hoose juice'. Breathless tabloid headlines were written, schlocky documentaries were made (I was made to watch one in school), and politicians started laying the blame at supermarkets for selling the stuff.
The big supermarkets stopped selling it of their own accord, and still don't to this day. They were hurt by the bad PR and like to get ahead of regulation, which is not toothless in the UK. Buckie, with it's now notorious reputation, started to be drank by artsy pretentious students instead of violent neds. And would you believe it, when it was drank by this different demographic, the problems magically disappeared!
My point being, observational studies are useless here because only a certain kind of person is consuming a drink called 'FCKED UP'. Caffeine is one of the most well-studied drugs in medicine, and we know it's safe. Evidence to the contrary would have to be strong, not be anecdotal or observational.
I'm personally a much bigger fan of Vin Mariani. It's a wine made with cocaine alkaloids and endorsed by the pope. Unfortunately not as popular as it was in 1863...
> made in Buckfast Abbey by an order of Benedictine monks.
Historically, yes, but after they lost their license to sell wine (a century ago) they licensed the name to a drinks business that scaled production up and changed the recipe.
reply