The core system is about 8000 lines of code. The C compiler is 1400. Its memory footprint is extremely small (1MB to have a C compiler running on bare metal in the PC architecture). From here, you have a system capable of building C programs as well as bootstrapping itself.
This is level of simplicity is staggering and empowering -- it is completely reasonable to read and understand every single line of code in Dusk OS. This would be incomprehensible in something like Debian.
I think you are missing the point. You might not have access to any modern hardware + modern, multi MLoC operating system medium/ftp server/whatever to boot in a post-apocalyptic world. This would at least provide some level of computation on simple devices.
Even if you do have access to a modern computer which is not fried, you still need a whole bunch of devices, mass storage stuff and mains electricity to be able to boot it, and then on to find and boot some software.
The OS above would run in things like graphing calculator cpus, 8-bit video games and other "almost non-computer" devices which actually have simple but useable CPUs. It would be better than going back to the 1900s.
I would frame it differently, "collapse" can mean any number of conditions that exist in the world today -- internet access that is limited, censored/liable of being shut down, or nonexistent. Same with electricity. We also rely on institutions that may or may not represent your interests to build software that you rely upon (e.g. the decline of search, the locking-down of major social media platforms, etc), and so on.
Virgil's "collapse" mindset, to me, is about building resilience and independence, and IMO Dusk OS is fantastic at this purpose.
There is a lot more software that currently exists written in C than in Dusk OS's forth dialect. The C compiler allows Dusk to run this software (once it's ported over)
One of the most interesting Forth projects to me currently is Dusk OS, a 32-bit operating system written in Forth that includes its own C compiler, with various porting efforts under way
I cannot speak to DuskOS, I don’t know anything about it.
While historically Forth has been “its own OS”, it’s a kind of crummy one. Especially for anything a modern reader might think of when they think of an OS.
I mean, for sure, what do you want from something that can run in 8K of RAM. But while it offers primitives, historically it doesn’t of any concept of things like drivers or most any other abstractions. Code is loaded from source code (which is slow). Linking is just loading blocks in the right order. Loading a new program you must first remove the current one.
Arguably not much better than CP/M, which routinely cold started the machine to exit a program, but at least it separated the BIOS from the BDOS.
I missed this one. Thanks for the link. I was just reading "Beyond the Collapse" and my Lisp and APL/J and C skills would not allow me to approach Forth's simplicity at the OS level. I'll definitely check this out by next weekend.
> their need to advertise admissions, how they talk to the media and university rating services, their relations with China, the student lawsuits they face, their need to manage relations with Oxford the political unit
This all sounds to me like DG's definition of 'bullshit'. David Graeber is an anthropologist, ie, he takes a wide view of humanity: many societies, including societies not that alien to us, manage to function without this kind of work. Why is it suddenly indispensable? It is not a business advice book about running an individual institution, it is a wider social critique about how we have set up our society and economy.
> let him spend a year managing a mid-size organization, say 60-80 employees,
Graeber is an anarchist anthropologist and would critique the idea of a top-down capitalist firm itself (ie, the deeper question: why do we insist on structuring labor this way?). He is not telling anyone the best way to run it.
I try to get to this point talking about insurance specialist working in the medical field in the US.
Those jobs are vastly more present in the US than in other places that provide adéquat level of care and also have a concept of healtcare.
Going to the ER you sign a lot of paperwork thougt your admission.
I think of it as the eastern island moais.
Something society is choosing to do, and focus on. While effort could be directed elsewhere in order to enhance our futur.
We don’t “insist” on structuring labour this way, it’s what arose as efficient from a market of voluntary private actors. Very “anarchist” in that sense. I would hope that Graeber is familiar with basic texts like Theory of the Firm.
Capitalism does not drive towards 'efficiency' in some general sense. It is full of dead weight, destruction, forced scarcity, waste, and bloat. This is not accidental, it is essential to how the system functions.
> But we're talking about New York city so you've already granted the port authority wide latitude and restricting your movements and controlling the population.
This is a bizarre argument. Cars are heavily policed: they must be registered and licensed. The state tells you where you can park it, and your ability to operate it is completely controlled by the state and can be taken away from you. In order for car infrastructure to function, there is a huge increase in police presence in people's lives to enforce traffic rules, parking, etc. None of this is true of, say, walking, biking, or taking transit, all of which are pretty unregulated, even in New York City.
The solution to a lot of problems is simply to not scale platforms past something like 1000 users. At that level, you can have a community that is guided by individual people and their relationships, rather than anonymized and centralized. And the infrastructure is much simpler to set up and maintain: we just need more tools tailored to being easy to set up and administer at this scale.
m15o has built a ton of examples along these lines. There are other communities and tools too, loosely referred to as the "small web"/"smol web"
There was an Ask HN (or similar) a while ago where someone was saying that they run a Discord server for active and committed independent game developers, and they were asking if anyone was interested in joining (I can't remember exactly). They wouldn't let just anyone in, to get in you had to schedule an interview with the moderators and show that you were a committed game developer. They were a tight community who shared experiences and helped each other. It sounds great.
The problem for me (and you) is that I'm not in it. The problem for a high school student who just wants to learn is that he's not allowed in either. The problem for most of the world is that the great insights within this community are forever hidden.
Perhaps small communities could nominate generally useful conversations and have them released as blog posts. At least then people could watch from the outside, while the community still remains small and private (not everything would be published).
The problem with small communities is that you don't necessarily have enough competent and available people to run the system reliably. Or worse, the people in charge leave, sometimes taking the servers with them, and no one else knows how to maintain it.
That's why Reddit and Discord are so successful. You can create your small community for close to zero cost and zero technical expertise, no one is going to pull the plug on the server because the one guy responsible for it didn't pay the bill, if the server in question is not an old PC in somebody's basement with no backup.
Even with good tools, building a community server still relies on an individual or a small subgroup to make an investment in money and time for the entire community, and if they stop doing so and no one wants to take over, the server is gone. On a platform like Discord, as long as there is any one person in there, it will continue running, the platform keeps the community accessible and does basic administration like spam filtering for free, even when there is no one to care, it can do that because it pays itself on user access, though premium membership and ads.
> The problem with small communities is that you don't necessarily have enough competent and available people to run the system reliably. Or worse, the people in charge leave, sometimes taking the servers with them, and no one else knows how to maintain it.
Don’t get me wrong, I acknowledge these are problems. But I have so little faith in big platforms these days that I don’t really see an alternative.
It may solve some problems, but it also cuts off the possibility of a GREAT deal of good. I get that it's hip to bash social media et al and its problems, but I believe that it's been instrumental in exposing and combatting a WIDE range of endemic problems in the world, e.g. police brutality.
"Big" spaces are important too. Maybe more important if we're trying to avert global disaster.
Sure, they vacuum things up, but people have agency. They choose to use platforms. They can choose to use different ones. I can't influence every person's choices, but I can choose how I invest my time and social energy, and present an alternative for people who are alienated with existing platforms (a very large and growing population)
This is a little atomistic view, but essentially in the correct direction. I would also add we need to nurture anti-centralization and anti-corporate sentiments that are still springing up in younger people especially, in their own way. If alternatives could bestow perceived social prestige for at least parts of the population, this would help preserve islands of free discourse for the future.
To be sure, this doesn't mean proclaiming decentralization to be cool necessarily (...fellow kids), but trying to be open-minded and friendly to the public, which I consider to be the actual better part of the ethos of the early Internet era. Being tech literate is "esoteric" by itself, and some complexities and social contracts cannot be really taken away from that, not without going back under the centralized yoke. But even moreso we should be trying to make it a little better by our attitude.
Even if a regulation of protocols for utilities will come, assuming it will be good, we need society to remain willing to preserve it.
But phpbb had its own issues. Many were run by hobbyists who eventually stopped paying the bills or updating the site, and then some bot steals the data or breaks the database.
It's like cloud-hosting email: everyone wants to not host their own email server, but then you're at the mercy of the service provider. For most people its worth it... until it isn't.
I would really like to see collaborative software development done via a usenet newsgroup. The git send-email program could be updated to send cover letters and in line patches to a newsgroup, and people could reply to those messages and review each patch.
> I'm open to giving someone free housing if they aren't on drugs
Sobriety is not a condition of housing. Many housed people drink and do drugs and don't get kicked out of their apartments for it. Why should we apply a harsher standard to our most vulnerable population?
You’re welcome to do whatever you want at your own expense.
If you can’t function to a degree where society needs to clothe you, feed you and house you; it should come with strings attached. Resources are finite and “the vulnerable population” isn’t entitled to everyone’s else labor.
My point is simply that policy should be based on what is effective and humane. Keeping people housed only based on the contingency of their sobriety is neither. Under no circumstances is it better for someone to be unhoused than housed. Housing should be a human right, full stop.
The thing is: How do we encourage or enforce good behavior? At what point do we insist that an individual try not to be a burden, and to try to be a decent participant in society? That could be as little as "don't be a public nuisance and help around the housing complex once or twice a week".
If there is no standard for behavior or "giving back" to earn one's keep, bad actors will bring everyone down.
The assumption is that drugs perpetuate the illness/uselessness of the homeless. If you have a home and can manage to afford it on your own, you get the privilege of drug consumption (within the law). If it is causing you to be unhoused or unhinged and the rest of the community is putting money into you having a place to sleep, it seems reasonable to impose some standards of behavior.
The drugs I'm talking about are illegal. Is it common to be a fentanyl addict that pays rent? Maybe but I doubt it. As a society we should do more to stop the opiod crisis. Why do we tolerate people smoking fentanyl on the bart or on our public sidewalks?
Because they're paying for it? If society is paying for your housing you need to follow societies rules. Giving addicts free housing isn't humane. It's enabling their addiction.
Ads exist because people are cheap and won't pay an equivalent amount of revenue in cash. All the complaints about ads and tracking will fall on deaf ears until that changes.
> It's looking increasingly possible that, at some point in the not-too-far future machines will be so good at creating software that humans won't be competitive in any way, and won't be in the loop at all.
This is an enormous extrapolation from what the LLMs are currently capable of. There has been enormous progress, but the horizon seems pretty clear here: these models are incapable of abstract reasoning, they are incapable of producing anything novel, and they are often confidently wrong. These problems are not incidental, they are inherent. It cannot really abstractly because its "brain" is just connections between language, which human thought is not reducible to. It can't reason produce anything really novel because it requires whatever question you ask to resemble something already in its training set in some way, and it will be confidently wrong because it doesn't understand what it is saying, it relies on trusting that the language in its training set is factual, plus manual human verification.
Given these limits, I really fail to see how this is going to replace intellectual labor in any meaningful sense.
I am extremely excited about and invested in this project, but I don't necessarily agree with the author's predictions. Rather, I see the way Virgil's design constraints have led to a very cool, novel and interesting approach to operating systems -- something so simple that you could reasonably comprehend every aspect of it starting from bare metal, yet powerful enough that you could use it to do useful computing work. I think there is value in having a system that you can comprehend fully -- something that is totally impossible with something like modern Linux. It doesn't have the same advantages as a modern operating system, but it has advantages that those systems lack: the ability to do a more independent, self-sufficient, sustainable kind of computing.
The core system is about 8000 lines of code. The C compiler is 1400. Its memory footprint is extremely small (1MB to have a C compiler running on bare metal in the PC architecture). From here, you have a system capable of building C programs as well as bootstrapping itself.
This is level of simplicity is staggering and empowering -- it is completely reasonable to read and understand every single line of code in Dusk OS. This would be incomprehensible in something like Debian.