advael

This might be my new favorite hn comment. At the very least it captures the zeitgeist perfectly

Those are the same thing. The whole point of saying "too big to fail" is to evoke the moment in the housing crash where governments largely threw most of their citizens under the bus by bailing out banks rather than homeowners for the banks' wildly irresponsible decisions. "Too big to fail" means the government steps in and bails you out, and that phrase became popular because for many it was the final nail in the coffin for their trust in government

I can use ssh as the transport and authentication layer over any internet connection through a fairly easily learnable set of ssh flags that can be further simplified through aliasing. As a bonus it's e2ee. The overhead from that affects latency but not bandwidth. Let's set aside ssh for a second. Streaming voice over the internet is a long-solved problem, I've been able to host a mumble server on a toaster since forever ago. So if the local machine can recognize voice commands, talking through a phone shouldn't make a difference. Like, if this works locally, and it works on the phone or whatever, why is having it talk over the internet the hard part? Whatever you think of the application itself, this is a weird failure mode

The LLM being able to generate random games is kinda expected behavior. It's trained on sequence probability distributions that include the code for a great many games and nudged toward doing so by the human user. I'm disappointed that the dog is basically used as a noise generator here. A process driven by the desires of the dog in a meaningful way would be more interesting and at least seems somewhat plausibly technologically feasible, and the article's title kind of implies it. I am especially disappointed because what most excites me about new technologies is applications that were not possible before its invention, which this seemed like it could be an example of

Dam, I really thought this would be much more interesting than it is

People have been doing some cool stuff for like a decade with giving dogs buttons to use human language, something they can seemingly get decent at communicating effectively with if they can get around the pesky issue of not having the sophisticated vocal machinery needed to produce recognizable phonemes, through the power of a good interface for them, even if the output is discretized to the level of words

I thought maybe this would be about creating a way for a dog to create stuff said dog might actually want or enjoy via the more powerful lever of effective long-context natural language processing that came of a similar tokenization approach - which can even sometimes churn out working code - that we have now

Instead it seems to be an exploration of how the capabilities you can produce from essentially random noise from this technology is less distinguishable from the result of thoughtful input than I might have hoped. Still interesting, but way less so

Yea I've noticed that most things made by a company are worse and less reliable than things that are free and maintained by volunteers

Weird all these companies struggle so much to support remote services, ssh has been working for me pretty seamlessly for like the 20 years I've been using it and has allowed me to remote-control any computer I own with relatively reliable authentication (with some hiccups that tend to be patched pretty rapidly when found) throughout that entire period. I hear tell it worked even before I was using computers professionally, too

Keypairs are fairly easy to use if you're on a reasonable unix-like OS and if you're not then frankly nothing is easy to use. Unfortunately this does mean that your statement is true for the majority of devices people use to access social media

Yea, the pervasiveness of this analogy is annoying because it's wrong (because a compiler is deterministic and tends to be a single point of trust, rather than trusting a crowdsourced package manager or a fuzzy machine learning model trained on a dubiously-curated sampling of what is often the entire internet), but it's hilarious because it's a bunch of programmers telling on themselves. You can know, at least at a high level of abstraction, what a compiler is doing with some basic googling, and a deeper understanding is a fairly common requirement in computer science education at the undergrad level

Don't get me wrong, I don't think you need or should need a degree to program, but if your standard of what abstractions you should trust is "all of them, it's perfectly fine to use a bunch of random stuff from anywhere that you haven't the first clue how it works or who made it" then I don't trust you to build stuff for me

I think robust crowdsourcing is probably the biggest capital-A Advancement in humanity's capabilities that came out of the internet, and there's a huge disparity in results that comes from how that capability is structured and used. Wikipedia designed protocols, laws, and institutions that leverage crowdsourcing to be the most reliable de facto aggregator of human knowledge. Social media designed protocols, laws, and institutions to rot people's brains, surveil their every move, and enable mass-disinformation to take over the public imagination on a regular basis.

I think LLMs as a technology are pretty cool, much like crowdsourcing is. We finally have pretty good automatic natural language processing that scales to large corpora. That's big. Also, I think the state of the software industry that is mostly driving the development, deployment, and ownership of this technology is mostly doing uninspired and shitty things with it. I have some hope that better orgs and distributed communities will accomplish some cool and maybe even monumental things with them over time, but right now the field is bleak, not because the technology isn't impressive (although somehow despite how impressive it is it's still being oversold) but because silicon valley is full of rotten institutions with broken incentives, the same ones that brought us social media and subscriptions to software. My hope for the new world a technology will bring about will never rest with corporate aristocracy, but with the more thoughtful institutions and the distributed open source communities that actually build good shit for humanity, time and time again

I can't parse what distinction you're trying to draw here. When I say "collective action" I mean exactly things like exerting political pressure toward objectives like expanding rather than reducing healthcare coverage provided by governments, such as medicaid. The notion that solutions that involve changing the policy of governments "require[s] Washington's permission" seems to reject the notion that we exert power collectively via democracy, but your proposed example of what someone shouldn't accept suggests that this is an issue of how people fail to see the value in exerting said power, for which the prescription would presumably be doing so? I don't understand what you're driving at or even why you think we're in disagreement

To "decenter [one's] self worth from [one's] job" would presumably require the fact that one's "access to basic rights is intermediated by a corporate job", no? This is a policy problem that needs to be solved by collective action, not a mindset problem that can be solved by personal growth

I think system package managers do just fine at wrangling static library dependencies for compiled languages, and if you're building something that somehow falls through the cracks of them then I think you should probably just be using git or some kinda vcs for whatever you're doing, not a package manager

But on the other hand, I am used to arch, which both does package-management ala carte as a rolling release distro and has a pretty extensively-used secondary open community ecosystem for non-distro-maintained packages, so maybe this isn't as true in the "stop the world" model the author talks about

One is an operating system maintained and run by a diffuse community of people, (albeit a flavor of linux with the explicit backing of at least one large company), whose primary goal is to create and maintain a functional operating system that can be used by many people for many different purposes. The other is a product whose primary goal is to convince investors that the company is on a growth trajectory that will continue into the next quarter by extracting value from that product's customers. We've now seen decades of data that suggest disparate results stemming from these priorities in a lot of contexts. I think viewed through that lens, the reason is obvious and was inevitable over time no matter what your threshold of "sucks" is, so long as it has something to do with the thing's function as an operating system

It's probably not worth it to say more than that my experience simply differs from yours. I've found it incredibly unproductive to quibble with people who have jumped to the conclusion that some difference of opinion must stem from some kind of identity-justification/confirmation bias delusion out the gate. This seems to be the most common mindless kneejerk criticism people jump to these days when they're engaging not with the person they're talking to, but a strawman or stereotype they believe that they conceptually represent, which in turn seems to be the most common failure mode of internet argumentation in general. It's interesting to see how the real phenomenon of confirmation bias and some relatively well-respected theories about opinion-as-identity have jumped from psychological literature to being basically pervasive thought-terminating cliches. But I like writing out my thoughts so... against my better judgment I'll write 'em out here. As a treat

My experience with the linux ecosystem overall, which seems consistent with that of the person you're responding to from what little information that post gives, has been of consistent improvement over a long timescale with an increasingly capable stack of open-source software whose exact pieces have shifted with various community and maintainer dramas and the natural process of the birth of new projects and death of old ones over time. I've found that I have my preferences within that ecosystem, like I settled on archlinux as a distro about ten years ago and haven't really seen a strong reason to switch, despite periodically working with other popular ones in the course of a career as a software engineer and researcher. I have strong reasons to prefer a modular, composable operating system that I control, so I wouldn't consider using proprietary software if there's a working FOSS alternative. This is a bias for sure! But I find my frustration with these things has decreased in aggregate over time, even as I've changed tools and suffered switching costs for it numerous times, and dealt with the general hostility with which a lot of manufacturers seem to view open-source software running on their hardware, and their attempts to make this more difficult. However, the aggregate experience of proprietary software users seems to have significantly degraded over the same period. They generally insist that this is still worth it to them over doing what I do, and again I've been in enough dumb internet arguments to know that it's not worthwhile to do more than gently suggest that alternatives exist and may be worth trying unless I know them personally.

I do get a window into proprietary ecosystems nonetheless, because I still don't feel I can replace the use cases required of me on mobile phones with an open-source alternative yet, and have seen my frustrations steadily increase over time with both these and SaaS products that I've been required to use for work. I also got frustrated enough with game consoles that I've entirely switched over to using PCs, running linux, for any games I want to play. At every turn, I have found that while computers are always error-prone in some way or another, and using them extensively will result in some frustration, this is significantly less when I have more control over the computer, and has become less rather than more frequent as open-source projects mature. I can not only observe that my own experience with proprietary products has followed the opposite pattern, but that more and more people talking about tech companies with scorn rather than effusive praise, yelling at their phones, and the public discourse adopting terms like "platform decay", "enshittification", "tech rot", etc all suggest that this is a general trend rather than my biases

Again, your mileage may vary, but I do find it odd that you are so immediately dismissive of this perspective, accusing a pretty innocuous comment about it of reactionary identity-defense basically immediately without engaging at all. If you're inclined to listen to a zealot like me at all, I would only urge you to consider why you have assumed this so quickly, why you are so adamant that this is the only sort of person who could form such an opinion

I feel vindicated by this article, but I shouldn't. I have to admit that I never developed the optimism to do this for two years, but have increasingly been trying to view this as a personal failing of closed-mindedness, brought on by an increasing number of commentators and colleagues coming around to "vibe-coding" as each "next big thing" in it dropped.

I think the most I can say I've dove in was in the last week. I wrangled some resources to build myself a setup with a completely self-hosted and agentic workflow and used several open-weight models that people around me had specifically recommended, and I had a work project that was self-contained and small enough to work from scratch. There were a few moving pieces but the models gave me what looked like a working solution within a few iterations, and I was duly impressed until I realized that it wasn't quite working as expected.

As I reviewed and iterated on it more with the agents, eventually this rube-goldberg machine started filling in gaps with print statements designed to trick me and sneaky block comments that mentioned that it was placeholder code not meant for production in oblique terms three lines into a boring description of what the output was supposed to be. This should have been obvious, but even at this point four days in I was finding myself missing more things, not understanding the code because I wasn't writing it. This is basically the automation blindness I feared from proprietary workflows that could be changed or taken away at any time, but much faster than I had assumed, and the promise of being able to work through it at this higher level, this new way of working, seemed less and less plausible the more I iterated, even starting over with chunks of the problem in new contexts as many suggest didn't really help.

I had deadlines, so I gave up and spent about half of my weekend fixing this by hand, and found it incredibly satisfying when it worked, but all-in this took more time and effort and perhaps more importantly caused more stress than just writing it in the first place probably would have

My background is in ML research, and this makes it perhaps easier to predict the failure modes of these things (though surprisingly many don't seem to), but also makes me want to be optimistic, to believe this can work, but I also have done a lot of work as a software engineer and I think my intuition remains that doing precision knowledge work of any kind at scale with a generative model remains A Very Suspect Idea that comes more from the dreams of the wealthy executive class than a real grounding in what generative models are capable of and how they're best employed.

I do remain optimistic that LLMs will continue to find use cases that better fit a niche of state-of-the-art natural language processing that is nonetheless probabilistic in nature. Many such use cases exist. Taking human job descriptions and trying to pretend they can do them entirely seems like a poorly-thought-out one, and we've to my mind poured enough money and effort into it that I think we can say it at the very least needs radically new breakthroughs to stand a chance of working as (optimistically) advertised

My days of not believing people's gushing praise about "just works" about any proprietary technology are certainly coming to a middle

That seems a hell of a lot better than repealing section 230 altogether. I also agree with the rest of this argument. Either the editorial choices made by an algorithm are a neutral platform or they're protected speech. They certainly aren't just whatever's convenient for a tech company in any given moment

I think by now roughly half of us grew up in a world where global reach has been simply taken for granted. I don't think it's particularly onerous to say that there should be some oversight on what a business can and can't do in the context where that business is relying on public infrastructure and can affect the whole-ass world, personally

Yea I think my own preference for self-hosting boils down to a distrust of a continuous dependency on a service in control of a company and a desire to minimize such dependencies. While there are FOSS and self-hostable alternatives to tailscale or indeed claude code, using those services themselves simply replaces old dependencies on externally-controlled cloud-based services on new ones

I dunno, I spend less time fighting with any of my several linux systems than the macbook I'm required to use for work, even without trying to do anything new with it. I choose to view this charitably and assume most of the time investment people perceive when switching operating systems is familiarity penalties, essentially a switching cost. The longer this remains the case, the less charitably I'm willing to view this.

We live in a world with the internet and distributed version control, so essentially every piece of software in the world has a tradeoff where the people maintaining it might push an update that breaks something at any time, but also those updates often do good things too, like add functionality, make stuff more efficient, fix bugs, or probably most crucially, patch out security vulnerabilities.

My experience with FOSS has mostly been that mature projects with any reasonable-sized userbase tend to more reliably not break things in updates than is the case for proprietary software, whether it's an OS or just some SaaS product. YMMV. However, I think probably the most potent way to avoid problems like this actually ever mattering is a combination of doing my updates manually (or at least on an opt-in basis) and being willing to go back a version if something breaks. Usually this isn't necessary for more than a week or so for well-maintained software even in the worst case. I use arch with downgrade (Which lets you go back and choose an old version of any given package) and need to actually use downgrade maybe once a year on average, less in the last 5

I like it, though I do notice that like most LLMs it seems to take criticism of LLM rollout kinda personally (actually circa 2023 this was considerably less true across every popular model)

I also think the predictions section seems kinda generic where the roast section felt better-personalized, which seems like a prompting issue

That sounds about right to be honest

In many moral frameworks, inconsistency isn't the only wrong someone can commit. The argument constructed in this article is essentially utilitarian, making the claim that the mechanisms of surveillance and privacy make this behavior harmful to others, regardless of their intentions or internal sense of morality. In fact, the author doesn't mention hating these people at all, although I suppose that's not a completely unreasonable thing to infer. From the perspective of this argument, this only lacks the harm the "deviancy signal" would itself do to the individual, though in the oppressive regime proposed they would perhaps take greater risk by openly deviating

This quote gives me such pause. I came back to read it again several times today. Conversations about echo chambers and filter bubbles are everywhere, and it's hard to sort into real data-driven arguments that there is an upward trend in the tendency to consume information that reaffirms our beliefs, but it does seem like our mechanisms for doing this have gotten a lot better, and that one could stay in a bubble indefinitely and never run out of content. I wonder if Murch is even still right to assume that we search for balance and harmony with an outside world we more often interact with through our abstractions, many chosen by us, most at least chosen by something. I wonder how many glaciers to read of restraint, how many volcanoes read of passionate abandon today, whether the feedback loop of escalation of flattery drives people to disappear into cages of their own making, or to burn themselves out, to use only this one dichotomy. I wonder how many of these feedback loops anyone is in about how many things. I wonder if I can even know which ones I'm in. Even if anyone succeeds at questioning everything they believe all the time, are they actually better off being a leaf on the wind, unable to form opinions?

I guess in short, this quote brings into sharp focus how brainrotting all this information and curation is, automatic and pervasive as it's become

I think a major problem with advice for a general audience is that different people need different advice. I agree with you that a path to mastery usually involves putting in a lot of effortful practice and then learning to operate without conscious effort, to let muscle memory and such take over. I think people fail at this in different ways, however. I'm sure a lot of people fall off of mastery because they mistake the feeling of effort for lack of an innate talent or the endeavor being futile, and a lot of people fail to achieve fluency because they're unable to let go of the effortful, conscious mode of thinking. Advice for either of those groups is probably going to be counterproductive for the other

That said, I do think this article frames its advice in a clickbaity way by handwaving cumulative effort while talking about instantaneous effort

From my experience riding in them and news reports I've read, any tesla fits the bill

Sadly, the most reliable signal american tech companies send is that they are primarily concerned with building a surveillance state. Whether this is for the US government or just their own fiefdoms (franchulates?) seems to vary a lot both within and between them, but neither prospect is particularly appealing to me as a prospective customer and/or target

This statement is, while true, quite vacuous. What's your point?

This is a phenomenon that, like many social phenomena, seems to scale superlinearly, and you've described a gradient of economic advantage along which this will tend to accelerate. These properties suggest to me that an attitude like "it's fine as long as it's not taken too far" is at best pretty naive