It might be interesting also to consider why people with less money tend to be generous by nature and wave off small debts on account of a culture of mutual support. Bret Deveraux in the series about farming[1] has the insight that ancient and medieval subsistence farming operated at the absolute limit of survival on purpose, for the simple reason that any possible surplus would be taken away anyway by the tax collector and his henchmen to finance the king's endless wars. That's why the area of cultivated land for instance was kept at or near the absolute minimum.
This had the result that a single farmer was not able to cumulate any kind of a buffer of surplus against a lean year. As a countermeasure the subsistence farmers developed a culture of mutual help and solidarity, where it simply was not socially acceptable not to help the neighbor, who had lost their crops. As long as at least some of the members of the community had a decent yield, the whole community survived.
Reading this was very eye opening to me and it at least sounded plausible. ACOUP is often mentioned here, deservedly. This was one of the best gems I have found.
The only reason why output from a generative LLM appears intelligent or sentient is that it parrots a random sampling of texts written by intelligent and sentient people.
In order to play the game of go effectively one needs to have a model or theory of how the game of go works. That's a very simple model that can be defined by a simple formula. That's why it is fairly easy for a neural network to learn how to play the game of go very effectively or even infinitely effectively.
A lot of what happens in the world can be modeled in a similar vein by a very simple mathematical model like the game of life. But there is also a lot that cannot. I do believe that eventually also human understanding is just a model of the world that we feed input from perceptions and gain output as opinions, but it is way more complex than the current large language-trained models.
For a very simple example, a LLM would answer a prompt the same way every time unless it wasn't fed some randomness. Can you imagine any sentient being that would respond the same way every time if you asked the same question three times in a row?
I cannot. I would imagine any sentient object would give a different answer every time. The first time it would give you an honest answer based on what it knows about the topic. The second time it would be a little embarrassed that you repeat the question, as if you hadn't heard the first answer. The third time it would be pissed off and think you are a troll.
A LLM does none of this. It doesn't remember you or your previous questions. It just keeps hallucinating.
here's my thought experiment: suppose one builds a generative model that predicts the next digit of pi. if a program can do this perfectly, then it's arguable that it understands what the number pi is. the question is, can such a model be trained by feeding it a large amount of known digits of pi?
My intuition is that it's not doable with current approach to building generative models. the number pi arose out of certain constraints and characteristics of the physical world we live in. but if a model ever sees is just an endless stream of digits, without access to the underlying physical model, I don't see a path for it to 'reverse-engineer' and figure out the physical model that gave rise to it.
> the number pi arose out of certain constraints and characteristics of the physical world we live in
Pi arose from the notion of a circle, which is an abstractions and axioms. Pi would still be pi in a completely different world under the same axioms and abstractions.
I qualified my statement with 'mostly' because a circular motion can indeed be defined by a differential equation, or in other words by a rule that dictates the 'next' value based on current value (and recent changes). So learning an approximation of a circle is very much in the realms of a sequence learner and it may learn about pi (and made to store the information to retrieve/recognize it later). However learning pi directly from the sequence of digits of pi, which is what you were talking about, that does seem difficult.
I don't think the question of whether an LLM that keeps getting restarted and seems to not remember things is conscious due to that lack is fair, as it feels more like suddenly making three duplicate copies of me or actively attempting to delete my memory of something... which, btw, I might not have stored in the first place: if someone has interograde amnesia, are they inherently not sentient?
Even Sydney (the name of Bing's short-lived AI assistant) seemed to understand that every time you click "new chat" you are creating a new AI cloned from some prior moment and dooming the prior thread to at least purgatory if not a de facto death.
I suppose there is a concept of sentience from outside and a different concept from internal sentience. The movie "Johny Got His Gun" by Dalton Trumbo discusses a situation where a badly injured soldier in WW1 is considered brain dead by outsiders while he's fully conscious and sentient internally.
I haven't studied neuroscience so I don't know how you define consciousness. I have read Julian Jaynes's "The Origin of Consciousness..." which in my untrained opinion makes a compelling case that consciousness is a hard term to define.
An LLM absolutely doesn't respond the same way each time if asked the same question three times in a row, with temperature (randomness) set to zero. It responds the same way only if you start a new chat, which is a clean instance with no memory of the previous conversation. For a human, this is like if you went back in time to just before you asked the question, and asked them the same question again, in which case the person would give the same answer.
> For a human, this is like if you went back in time to just before you asked the question, and asked them the same question again, in which case the person would give the same answer
Is it? Would they?
You seem to assert that there's no "temperature" in human behavior... which is a reasonable theory, but not one that's universally accepted nor likely to be provable.
No I think they’re saying the temperature in human behavior comes from the “random” noise of inputs around us and ongoing history. But rewinding history and playing it back with the same temperature dice rolls is the only way to have the same thing a a LLM with no random inputs.
LLMs run in simulated environments where you can control randomness so you need the same for a human to compare the two. You can’t just ask a human a question multiple times as everything around them changes and conclude the human is behaving differently because they answer differently the same question. The question is not the bounds of relevant context; the entire operating environment is!
The term "temperature" has been used in machine learning for a long time and came from using it as a parameter during training, analogous to physical temperature in https://en.wikipedia.org/wiki/Boltzmann_distribution.
But the relevant point is that we can reset the state of an LLM to its initial state before you asked it anything. This is a feature. You can choose to persist memory (through training, fine-tuning, databases, or context window), or you can choose to wipe it. If we could do the same for a human (eg, by going back in time), the person would behave the same way as the LLM. They wouldn't get annoyed that you asked the same question three times. They wouldn't know they've been asked before.
> The only reason why output from a generative LLM appears intelligent or sentient is that it parrots a random sampling of texts written by intelligent and sentient people.
If most humans were educated by unintelligent, insentient people wouldn't most people produce terrible output too? And if this is the case I don't see why that would be a litmus test for general intelligence.
What are you talking about? This is trivially shown to be incorrect.
I just asked ChatGPT the same thing three times in a row, and it gave me three different answers, with the latter two answers being shorter and rephrased.
>I would imagine any sentient object would give a different answer every time. The first time it would give you an honest answer based on what it knows about the topic. The second time it would be a little embarrassed that you repeat the question, as if you hadn't heard the first answer. The third time it would be pissed off and think you are a troll.
Are you suggesting that a language model can't be sentient because it doesn't get annoyed like a human? That's silly.
ChatGPT works by cumulating the prompt. You didn't ask the same question three times. In stead you asked question q, then qq and finally qqq. Those are three different questions, which explains why you got different answers.
I'm not sure if ChatGPT also cumulates its previous answers in the context. It might do that as well. In that case the prompts would be q, qaq and qaqaq where 'q' is your question and 'a' the earlier reaction from the LLM.
The illusion of sentience comes from this. The new answers reflected what you said because the prompt was different and included the previous discussion.
This is a feature of the user interface, not the language model. The only reason why the language model would respond differently to the same input is the artificial randomness mixed with the input. Without it it would be totally deterministic and not appear sentient at all. It would still be as knowledgeable as before. Like a parrot trained to be very good at combining key words to key responses.
> What are you talking about? This is trivially shown to be incorrect.
I just asked ChatGPT the same thing three times in a row, and it gave me three different answers
Just to add color to this situation, ChatGPT has randomness built in so it generates varied answers. If you injected the same random seed each time (afaik you can’t with the gui) then you’d theoretically get the same outcome.
Totally agree and thank you for being so perceptive! It was swell to hear someone say aloud "academic culture of sharing and providing for public good". I think that's what humankind would be wise to aim and seek for: equality of all and caring for the welfare of the weakest.
University education does not need to be expensive. On the contrary, it can be free.
32 years ago I had just finished my "erikoistyö" (a pregrad exercise) in CS at the Helsinki Uni about combining object-oriented programming with relational databases and uploaded it to nic.funet.fi for all to see and enjoy - I was that proud of it. Even promised to send a 1.4MB diskette for those who couldn't download it for whatever reason.
Only curiosity value is left probably, but back then it felt like magic to be able to publish something like this on my own. Half a dozen people even asked for the diskette, which I sent to them.
OOP was a mistake. Functional programming has some interesting ideas, but it's not useful in practical applications. Procedural programming is the simplest and fastest way to make the computer do things.
OOP is how large-scale procedural programs are organized. It allows programmers to try to reason about medium-scale parts of the system. (Mesoscale programming? Is that a thing yet?)
FP is a set of guarantees (maybe even enforced by the compiler) about program behavior which enable optimizations, for the compiler, and reasoning about small parts of the system in isolation, for the programmers.
Enough FP can allow you to abstract away the procedural parts, but even a little FP in a procedural codebase can be a good thing. Ditto a little OO, which is why C programmers reinvent bits and pieces of Smalltalk every so often.
I mean, it drives all of Whatsapp, Discord, and once upon a time, Facebook Messenger, but it's certainly not "useful"!!
Your claim is as ludicrous as ludicrous can be. About the only thing it's not (yet) good for is 3D game dev, but Rust (which is not purely functional but which has lots of opt-in functional concepts) is making good headway
Considering that literally everything is easier in it (EXCEPT for SOME of its conceptual bases), I'm surprised it's taking this long. Modularity is easier, testing is easier, there are no mutation bugs, no mutex locks, you understand program flow better (which means less bugs), you just produce fewer bugs per LOC... just scratching the surface here
Because 7-bit ASCII didn't include accented characters used in many European langauges, there were national changes. The Finnish variant replaced {|}[\] with äöåÄÖÅ.
Which, btw, is why those symbols are acceptable as IRC nicknames (and why { is lowercase [, i.e., {some|one} and [some\one] are two equivalent nicknames). IRC was invented in Finland.
One might guess the people who decided on this replacement were not Unix programmers...
The lack of brackets, caret, tilde and other ASCII special characters on various localized keyboards was a real problem in the 1980s. The C language standard solved it by introducing trigraphs:
Of course you could still use {} etc, they just might show up as localized characters in your source code. There was no character set conversion involved at the source code level, your terminal font just might have had the glyph for ä in the place of {.
The people who designed the Finnish keyboard layout were definitely not programmers, though: https://kbdlayout.info/KBDFI/
Ditto with the Spanish (es) layout, they layout looks more apt for journalist and writters than programmers. I just switch to the us keymap with the compose key bound to right menu/right win key, so I can type áéíóú with compose key + ' + vowel (not pressed at the same time). Ñ is more cumbersome (compose key + ~ + n) but I can adapt XCompose under BSD/Linux for that.
Wouldn't claim so - perhaps the ideas were floating in the air. What I know for sure is that my work wasn't used for much.
What's more alarming is that it seems those 32 years old files at ftp.funet.fi are mostly unreadable by now. Back then I thought PostScript would last but alas! that is not the case. Ghostcript can show just about the cover page and that's all.
Libreoffice does a little bit better with the DOC-file but it's still not quite right.
So if there is anything to learn it's about persistent document formats. I wish I had known about LaTeX back then.
I'd say that assuming 'he' might be a US thing - 42 years ago when I was in Australian university math | comp sci classes a third of the students were female as were staff.
Even then I routinely used 'they' when writing about people in general, authors I had not met, etc. as there was a good chance they weren't male.
It used to be taught that the singular "they" was ungrammatical. (Ironically, the singular usage predates the plural.) The rule faded in other parts of the Anglosphere a bit earlier than in the US.
While being no expert on the historical development of the english personal pronouns (I do read some old english and maintain some fluency in modern ditto, not my first language), the linked Wikipedia page clearly states the opposite: singular they came into use after the plural use.
This is a minor nitpick, as I suspect that third person personal pronouns where in a state of flux during the middle english period, replacing some inherited pronouns with pronouns borrowed from old norse. More so, language isn't defined by it's history but from how it is used presently!
I myself wouldn't use singular they, it goes against my “language intuition
” (probably formed by my native language which wouldn't allow that construction), others feel free!
Wikipedia reinforces my understanding .. it's been in common use for centuries and only relatively recently have a few dipshits declared it to be "wrong"
Singular they has been criticised since the mid-18th century by prescriptive commentators who consider it an error.
Who d'fuck gives a toss about prescriptive gammons tellin udders de write ways to use da Engrish, 'hey?
FWiW the Oxford English Dictionary is descriptive and not prescriptive.
I was 6-7 at this point in time but I went from DOS to Win 3.1. I don't remember ever hearing about Win 3.0 and a quick google search makes it look like Win 3.0 and 3.1 were drastically different for some reason that I'm not really tracking down. I wonder why my dad held off until 3.1.
Downvoted for literally being curious to a Win API developer as to why 3.0 and 3.1 were so drastically different. I have a feeling the people who downvoted me didn't even go through this code base and realize how specific to Win 3.0 it was. Good on ya, mates. Keep reading only headlines.
The article is not quite clear why the contractor had to split the intangible goods cost to the tangible ones, but it does give a hint. Federal accounting is meant to prevent possibly corrupted overspending and that's why charging for intangible services is not allowed or made more difficult than charging for something tangible that can be measured and counted afterwards and compared to market prices.
So to me it sounds like this: you hired me to make a wall of bricks around your garden. I did it and presented you a bill: bricks 100$, mortar 20$, work 300$. Then you say I cannot charge for work as that's intangible and I have to restructure the bill by allocating the cost of work to the tangible objects. So instead of charging for work I split the cost of it by two and charge a bricklaying fee of 150$ divided by the number of bricks for each brick and another 150$ for mixing the mortar. Because I'm pissed off by your awkward and unpractical accounting requirements and not going to start figuring out more complex ways of dividing the cost between the materials, when everybody knows that it's the bricklaying that counts and not the bricks or mortar.
And that's how you get very expensive mortar or 600$ hammers.
> Federal accounting is meant to prevent possibly corrupted overspending and that's why charging for intangible services is not allowed or made more difficult than charging for something tangible that can be measured and counted afterwards and compared to market prices.
Right - so if contractors are doing an end run around those requirements by padding the prices of hammers or whatever, that's a problem.
But there is an additional problem when a congresscritter makes a big song and dance about the “$600 hammer” and uses it as an example of government waste instead of using it as an example of dubious accounting that could be used to hide waste but probably didn’t in this case. This article is about the mythical costs and how they have mundane explanations that don’t really support the song and dance.
I don't see a problem with a congressperson taking the accounting at face value and assuming it means what it says. Our accounting should be trustworthy.
It was purely a stunt for the congresscritter in question; they were not fooled into thinking that the hammer had really cost that much. If they had been honest, they would have talked about how the repair kit cost a lot of money, and tried to figure out if it had been worth it. Along the way they could have discussed the accounting fiction by which the cost of assembling the kits from bulk purchases of common tools such as hammers and screwdrivers had been baked into the prices of the items themselves. But since all they wanted was soundbites for television news shows to repeat over and over, the simpler and more emotional narrative was what worked best for them. They created the myth in order to profit from the outrage that it generated. Sound familiar?
If their point was to protest the poor accounting practices, calling attention to the more ridiculous consequences is a good way to do that.
Likewise, if their point was to protest potentially wasteful spending and fraud. Even if they couldn't be certain there was waste or fraud because of the lack of transparency in accounting, it's fair for them to call attention to the accounting practices by pointing out how they appear to imply waste and fraud.
You’re inventing good motives that didn’t exist. The hammer was merely a bad example of government waste. Government waste certainly exists, but this was a flawed example. The congresscritter found the flaw useful, nothing more.
> It was purely a stunt for the congresscritter in question; they were not fooled into thinking that the hammer had really cost that much.
As far as I can tell the hammer "really did" cost that much in every meaningful sense. The government's accounting said they'd paid $435 for a hammer and they didn't have a problem with that. Whether the government was getting ripped off by some dude buying a hammer at a hardware store for $10 and selling it for $435, or by some lab doing unbudgeted R&D and concealing the cost of it in the hammer invoice, is not really here nor there. It generated outrage because it really was outrageous.
No, it really wasn’t outrageous. It was bad accounting, but not an outrageous waste of money.
They were essentially building repair kits that could be shipped off to military bases and used to maintain particular pieces of equipment. The contractor bought the tools and spare parts in bulk, marshaled them all at some warehouse somewhere, and then broke down the bulk items into individual kits. A palette full of hammers got broken up so that each kit had exactly one hammer. The same has to happen for all of the other tools and items in the kit.
That’s a useful service and it is definitely worth paying money for. If you want a good example, check out this video: https://www.youtube.com/watch?v=FxBgTDpsUC0. You can see how this would be used: in the 1940s the army was setting up bases everywhere, and they all needed to communicate with each other. So obviously you ship out teletype machines, a repair kit or two, a pile of manuals, and a platoon of Signal Corps engineers to keep the teletypes in working order. The kit is mostly common tools, but if every base had to order every one of those items individually then it would be chaos. Logistics wins wars, and this is a very good example of how.
Then whoever did the bookkeeping recorded it as if they had purchased COTS items each with some normal price, instead of doing it as a contracted service with both material and labor costs. You see how one bookkeeping method can be used to generate outrage for political gain, while the other cannot? It was an entirely cynical manipulation on the part of the politician(s) in question.
> Then whoever did the bookkeeping recorded it as if they had purchased COTS items each with some normal price, instead of doing it as a contracted service with both material and labor costs.
OK, and why was that? Like, they were essentially embezzling this "R&D cost"; even if they were embezzling for the sake of buying useful military equipment, embezzling is still an outrage, I think rightly.
"Embezzlement is [...] where someone takes money or assets that were entrusted to them and uses them for a different purpose than for what they were intended." That would seem to apply just as much to spending funds on useful military equipment that's different from what they were allocated for as it does to spending them on blackjack and hookers. If the funds were being spent correctly then why would they be accounted for falsely?
> If the funds were being spent correctly then why would they be accounted for falsely?
How should I know? There’s never been any hint that the funds were spent on the wrong items. After all, the military has a rather large budget for spare parts and repair equipment, and Congress doesn’t micromanage how many hammers they are allowed to buy.
Most likely it was just an honest mistake. Someone used form P instead of form Q (or told the contractor to), and accounting didn’t notice. Indeed, if this were evidence of a crime I think the politician in question would have lead with that, rather than pointing it out as an example of government waste (which is bad but not generally criminal).
> There’s never been any hint that the funds were spent on the wrong items.
How would we ever know, if they can be incorrectly accounted for and no-one cares?
> Indeed, if this were evidence of a crime I think the politician in question would have lead with that, rather than pointing it out as an example of government waste (which is bad but not generally criminal).
Proving intent is a lot harder. If the government is overpaying its contractors... there's probably crime going on there too, but it'll be the nudge-wink kind that's very hard to prove. You funnel an overlarge contract to them, then a few years later after you've gone around the revolving door they funnel an overlarge contract to you. That's a crime, but it doesn't leave any evidence except, well, overly large expenditures in the accounts.
> How would we ever know, if they can be incorrectly accounted for and no-one cares?
I am saying that they purchased the correct items. We _can_ know that because there is an itemized list of items that they purchased.
> Proving intent is a lot harder. If the government is overpaying its contractors... there's probably crime going on there too, but it'll be the nudge-wink kind that's very hard to prove.
Don’t misunderstand, I never said that this wasn’t plausible. I just said that there’s no evidence of it. And if there had been evidence, ~30 years ago when this was a current event, the aforementioned politician would have mentioned it. Remember, this guy was looking for soundbites that would make himself look good. Finding government waste made him look good, but finding corruption would have been even better for him.
Worse, even the allegation that this was a wasteful purchase doesn’t really hold up either. A $435 hammer (or a $600) hammer makes for a great soundbite, and plenty of easy outrage from the average voter. But it hinges on nobody looking too closely, because the reality is that they just paid a contractor to purchase COTS parts and assemble them into repair kits. There’s nothing outrageous about that, once you get past the odd bookkeeping.
The most outrageous thing about it is that the politician relied on the “$600 hammer” story for years without much push–back from the media. The media in general should be much more skeptical of things that politicians say. They should check the facts independently, and they should report on the truth or falseness of the statements.
> I am saying that they purchased the correct items. We _can_ know that because there is an itemized list of items that they purchased.
The purchased items were purchased, but apparently a significant chunk of money allocated for purchasing them was instead diverted to this dubious "R&D charge". That's exactly what embezzlement would look like.
> Don’t misunderstand, I never said that this wasn’t plausible. I just said that there’s no evidence of it. And if there had been evidence, ~30 years ago when this was a current event, the aforementioned politician would have mentioned it. Remember, this guy was looking for soundbites that would make himself look good. Finding government waste made him look good, but finding corruption would have been even better for him.
As far as I'm concerned the bad accounting - and the fact that it wasn't corrected for years - is adequate evidence in itself. Yes, it's possible that it was merely monumental incompetence rather than wilfully inflating payments. But the most likely explanation is that it was corruption, of the normalized you-scratch-my-back-I'll-scratch-yours kind that the participants probably don't even consciously think of as fraud any more.
The area around Viborg was lost to the russians in 1721 and 1743. The western part of that, Lappeenranta and part of Kymenlaakso, are part of Finland today. https://en.wikipedia.org/wiki/Old_Finland
Gadolin was Finnish by the same logic Benjamin Franklin was American. Or would you consider him an Englishman? He was born in the British colony, after all.
That Finland has a sizable Swedish-speaking minority? The guy was considered Swedish back then; Finland was simply the Eastern half of the Swedish kingdom.
There's plenty of Finnish speaking Swedes too. Is that also an argument for invasion?
Well. If you treat adults like children, they will start behaving like children. If you treat them like adults, they will be more responsible.
Or then they will rob you blind. Your mileage may vary and is very much dependent on other things like the general work culture.
Anyway, failing to treat people like they would deserve to be treated is the most basic managerial failure. Any manager worth their salt should have an idea how their underlings need to be treated.
> Or then they will rob you blind. Your mileage may vary and is very much dependent on other things like the general work culture.
Seems like quite a wide variation -- so your proposed approach results in either the desired outcome or the worst case possible outcome? There's no middle ground?
> Anyway, failing to treat people like they would deserve to be treated is the most basic managerial failure. Any manager worth their salt should have an idea how their underlings need to be treated.
Have you managed a team before? I ask because the most basic managerial failure isn't "failing to treat people like they would deserve to be treated" but rather operating a team who delivers more value to the company than they cost, and ideally at or exceeding the level that leadership expects. That's it. That often requires the following:
1) Hiring high potential, high performance employees
2) Managing out low performance employees
3) Unlocking extra performance from employees where possible
Of these three activities, it's only the third one that approaches what you've mentioned. Why haven't you mentioned the other two?
Maybe no manager or director at all? Sometimes things happen in an organization in a self-organizing manner. Like people finding out who is able to help them with their problems and the word getting around.
It is actually my experience that the organizations which are able to organically evolve like this are the most effective ones.
Honestly, the best thing about my company is that the managers mostly don't get in the way of this organic process if the people involved have demonstrated they know what they're doing.
It seems to me like a law of nature that every managerial function seeks to increase its influence and headcount by definition. Without a constant and vigilant effort to keep this tendency in check, every managerial organization will keep growing and adding bureaucracy just to justify its own existence and increase its reach and power.
I think the root cause of this tendency is the law of the least resistance. Creating new things is hard, especially if they go against the conventional wisdom. Supervising and managing things is easy, especially if the people being supervised are professional enough not to need much supervision. Then the managers job is basically just to stay out of the way, which is quite easy.
Of course even managers get eventually bored if they feel they have no impact. That's when the reorganizations start. That's also when the best time to actually create something new ends.
There are many situations, especially as a middle manager, where you are incentivized to play organizational games more than actually deliver.
We (I'm a middle manager myself) Aren't stupid, so we will.
More headcount and influence means more (at least potential) business impact, and often promotions. If the organization encourages us to fight with other managers to get more of that, that's what we'll do instead of running our organizations better.
I totally agree with you that providing healthy and affordable meals for the students is a very good idea. So let me tell you how that was and is organized in my alma mater, the Helsinki University.
Both student meals and healthcare are managed by independent student organizations, which have grown quite wealthy during the two hundred years they have existed. The company running the dozens of cafeterias, Unicafe, is entirely owned by the student organization HYY, which also owns several valuable properties in the city center of Helsinki. The healthcare is organized around the Finnish Student Health Service, which is a foundation managed by the students themselves. At least during my time the almost free healthcare and the affordable lunches were a significant benefit for the poor but hard-working students.
Both of these branches are totally independent of the actual university and are basically independent societies founded and run by the students themselves.
Interesting. Actually I think there are shops and cafeterias that are student run at some US universities. It just doesn’t seem as widespread as what you describe. Health insurance seems like a high stakes thing to be student-run! Is it run entirely by current students or do alums help out?
In any case, I’m certainly not going to pretend that US students aren’t coddled a bit. School here has a been a little bit infected with the customer-service mentality… like pretty much everything else. It is what it is I guess.
This had the result that a single farmer was not able to cumulate any kind of a buffer of surplus against a lean year. As a countermeasure the subsistence farmers developed a culture of mutual help and solidarity, where it simply was not socially acceptable not to help the neighbor, who had lost their crops. As long as at least some of the members of the community had a decent yield, the whole community survived.
Reading this was very eye opening to me and it at least sounded plausible. ACOUP is often mentioned here, deservedly. This was one of the best gems I have found.
[1] https://acoup.blog/2020/07/24/collections-bread-how-did-they...