And what is using Confluence in the first place? Your MacBook Pro is faster than a supercomputer from 20 years ago. As we make compute cheaper, we find ways to use it that are less efficient in an absolute sense but more efficient for the end user. A graphical docs portal like Confluence is a hell of a lot easier to use than EMacs and SSH to edit plain text files on an 80 character terminal. But it uses thousands of times more compute.
It seems ridiculous right now because we don’t have hardware to accelerate the LLMs, but in 5 years this will be trivial to run.
I'm confused by your analogy. A wiki run server is extremely efficient to run, and can be hosted from a tiny little raspberry pie. A search engine can be optimized to provide results near O(1). You can even pull up and read results on a very old computer. All of the concerns around cost and resource efficiency can be addressed as all of this is a solved problem.
Even with an LLM agent getting cheaper to run in the future, it's still fundamentally non-deterministic so the ongoing cost for a single exploration query run can never get anywhere near as cheap as running a wiki with a proper search engine.
This thread is about how a static text article loaded 500 megs in the background. How would someone prepare for that exactly? This is effectively malware as far as your bandwidth is concerned.
I’ve used travel SIMs that only give you about 5GB. You avoid using the web at all, unless you are on WiFi. You can use maps, train and bus apps, banking apps, messaging, AirBnB etc, but not the web. If you go to some place and they want to use a QR code to buy a ticket or use a menu you may as well forget about it.
People on government assistance are just casually going to Starbucks for free wifi? They probably don’t even have a reliable way to get around. Let them eat cake?
There aren't even any Starbucks in these areas :( Best you can do is a McDonald's or Burger King and they hate opening the doors most of the time as they tend to prefer only allowing drive-thru orders.
Thank you for posting this. It’s totally understandable and believable that you simultaneously love them and regret some things about it. There’s this insane pressure in our society to never acknowledge the toll that kids have and to never speak out about this. I remember when this article was first posted and how I received it, like I was wrong for not being sure about kids, and that some change would come over me when I had them. Truth is, that doesn’t happen with everyone. Then the world tries to gaslight those people who don’t feel that way into feeling like they’re broken somehow.
I’m sure you love your kids and take great care of them, and it’s not your fault that you feel this way.
It would benefit all of us if this taboo was lifted, so that we could speak truthfully about the impact of kids on families, and maybe then we’d have to provide more support and encouragement to convince people to have them. Not everyone has free daycare from their grandparents or a large social network to babysit or the finances that make having a child less of a burden.
Saying that the paradigms of Python and Haskell are the same makes it sound like you don’t know either or both of those languages. They are not just syntactically different. The paradigms literally are different. Python is a high level duck typed oo scripting language and Haskell is a non-oo strongly typed functional programming language. They’re extremely far apart.
They are different, but on some fundamental level when you're writing code you're expressing an idea and it is still the same. The same way photograph and drawing of a cat are obviously different and they're made in vastly different ways, but still they're representations of a cat. It's all lambda calculus, turing machines, ram machines, combinator logic, posts' dominoes etc., etc. in the end.
Other people have replied but to clarify my point, while two languages may operate with a focus on two seperate paradigms, the actual paradigms do not vary. OOP is still OOP whatever language you use, same for functional et al. Sure some languages are geared towards being used in certain ways, some very much so, but if you know the actual ways the language is largely irrelevant.
They’re both turing complete, which make them equivalent.
Code is a description of a solution, which can be executed by a computer. You have the inputs and the outputs (we usually split the former into arguments amd environment, and we split the latter into side effects and return values). Python and Haskell are just different towers of abstraction built on the same computation land. The code may not be the same, but the relation between inputs and outputs does not change if they solve the same problem.
> The paradigms literally are different. […] They’re extremely far apart.
And yet, you can write pure-functional thunked streams in Python (and have the static type-checker enforce strong type checking), and high-level duck-typed OO with runtime polymorphism in Haskell.
The hardest part is getting a proper sum type going in Python, but ducktyping comes to the rescue. You can write `MyType = ConstructA | ConstructB | ConstructC` where each ConstructX type has a field like `discriminant: Literal[MyTypeDiscrim.A]`, but that's messy. (Technically, you can use the type itself as a discriminant, but that means you have to worry about subclasses; you can fix that by introducing an invariance constraint, or by forbidding subclasses….) It shouldn't be too hard to write a library to deal with this nicely, but I haven't found one. (https://pypi.org/project/pydantic-discriminated/ isn't quite the same thing, and its comparison table claims that everything else is worse.)
You can also use cloudflare to create a dns record for each local service (pointed to the local IP) and just mark it as not proxied, then use Wireguard or Tailscale on your router to get VPN access to your whole network. If you set up a reverse proxy like nginx proxy manager, you can easily issue a wildcard cert using DNS validation from your NAS using ACME (LetsEncrypt). This is what I do, and I set my phone to use Wireguard with automatic VPN activation when off my home WiFi network. Then you’re not limited by CF Tunnel’s rules like the upload limits or not being able to use Plex.
This is exactly what I do. I have a few operators set up in k8s that handle all of this with just a couple of annotations on the Ingress resource (yeah, I know I need to migrate to Gateway). For services I want to be publicly-facing, I can set up a Cloudflare tunnel using cloudflare-operator.
It seems ridiculous right now because we don’t have hardware to accelerate the LLMs, but in 5 years this will be trivial to run.
reply