3cats-in-a-coat

Ancient world? The Roman Empire fell apart. You know that, right? So, turns out their worries were warranted. A culture, a civilization can go through a period of stagnation and decay, and it eventually dies. Then there's a lengthy period of chaos and eventually something else may arise. But we're talking countless generations lost in that noise.

Currently, our civilization is more united than ever. Monoculture. Not only the Western World, but The World. We're united through communication, technology, everything. And when we all start declining at the same time, through many objective metrics, mind you, including the decline of democracy worldwide, the terrifyingly low level of public discourse, reduction and profanation of our vocabulary, inequality, collapsing demographics, climate, wars etc. it's grim. It's grim, because there is no clear alternative arising... except AI.

And I don't think we want AI to be an "alternative" to human civilization, that wasn't the plan, was it?

See, that's the thing, we look around, we see the technology we live with, and we feel very personal about technological progress, as if you and I personally invented everything from the wheel, through electricity, to computers, networks, and AI. We feel in control, we feel smart, we feel so Personally Intelligent (tm) for having technology. We feel "wow, clearly if we have these sci-fi things, we're on the up and up". But technology is not you. It's not me. It has taken on a life of its own. It serves capital, not the people. As technology gets betters, humans get worse. So tech progress is not the story you think it is. And it doesn't end how you think it does.

Things are VERY fragile right now.

The way language is eroding is very indicative of our overall social and cultural decay.

AI currently is a broken, fragmented replica of a human, but any discussion about what is "reserved" to whom and "how AI works" is only you trying to protect your self-worth and the worth of your species by drawing arbitrary linguistic lines and coming up with two sets of words to describe the same phenomena, like "it's not thinking, it's computing". It doesn't matter what you call it.

I think AI is gonna be 99% bad news for humanity, but don't blame AI for it. We lost the right to be "insulted" by AI acting like a human when we TRAINED IT ON LITERALLY ALL OUR CONTENT. It was grown FROM NOTHING to act as a human, so WTF do you expect it to do?

Reproduction in time and reproduction in space are connected. If atoms couldn't reproduce (I can sense most readers knee-jerking reading this), they'd be all unique, wouldn't they. And yet they aren't.

You could say "they have no heritability", and not the way you expect, I guess, but they all inherit the same local laws of physics, and they may even impact those laws, thus forming a feedback loop, and clearly there are googols of them in clusters, same weight, same energy, same polarity, same properties, same states, much like you see with any other species in nature, in fact in far lesser numbers.

If robots are made in a factory, does this count as reproduction? If not, why not. Does a mother's womb not resemble a "baby factory". A baby does not create itself. Always something else creates you.

We have clusters of "common sense" about these things, and most of what I said immediately sounds stupid to "common sense". Yet common sense falls apart if you start thinking about it. But Internet is not EXACTLY conductive to "thinking about it". It's all about the hot takes and the current consensus. Then time passes, and that consensus seems truly unenlightened.

Atoms and sub-atomic particles fit this definition.

Machines fit this definition.

Fire fits this definition.

Truth is "life" is not a distinct category. We just think of life as complex life. A complex system that mines energy gradients to preserve and replicate its forms.

But there's no hard boundary. It's just in our head.

No one forces us to use objects in JSON with repeated keys you know.

I'll say the obvious. A lot of this you can just do in JSON.

Let's take the example:

    {
      "users": [
        { "id": 1, "name": "Alice", "role": "admin" },
        { "id": 2, "name": "Bob", "role": "user" }
      ]
    }

    users[2]{id,name,role}:
      1,Alice,admin
      2,Bob,user

We can keep it JSON, but use more compact list expressions, as tuples when pragmatic:

    ["users",
       [1, "Alice", "admin"],
       [2, "Bob", "user"]
    ]

The thing is the game with LLMs is not what's shortest, but what's:

1. Mainstream, so they understand it.

2. What they're tuned for, and their tuned for what's mainstream (JSON).

If you want to go extreme compression you can shove it all in JSON strings too and keep the larger structure JSON:

    ["users",
       "1:admin:Alice",
       "2:user:Bob",
    ]

You may say "how is this better". Well it's better because it's still JSON, there's less to explain to the LLM, and to your other devs. Even if we use a weird compact format like "id:role:name" this is still shorter to explain than a completely different syntax with its whole world of rules.

On the path they've started with pipelining calls parametric on previous calls, they'll quickly realize that eventually they'll need basic transforms on the output before they pass it as input i.e. necessitating support for a standard runtime with a standard library of features.

I.e. imagine you need this:

    promise1 = foo.call1();
    promise2 = foo.call2();
    promise3 = foo.call2(promise1 + promise2);

Can't implement that "+" there unless...

    promise1 = foo.call1();
    promise2 = foo.call2();
    promise3 = foo.add(promise1, promise2)
    promise4 = foo.call2(promise3);

You can also make some kind of a RpcNumber object so you can use their Proxy function to do promise1.add(promise2) but ultimately you don't want to write such classes on the spot every time. Or functions on the server for it.

The problem is even that won't give you conditions (loops, branches) that run on the server, the server execution is blocked by the client.

Once you realize THAT, you realize it's most optimal if both sides exchange command buffers in general, including batch instructions to remote and local calls and standardized expression syntax and library.

What they did with array.map() is cute but it's not obvious what you can and what you can't do with this, and most developers will end up tripping up every time they use it, both trying to overuse this feature and underusing it, unaware of what it maps, how, when and where.

For example this record replay can't do any (again...) arithmetic, logic, branching and so on. It can record calling method on the Proxy and replaying this on the other side, in simple containers, like an object literal.

This is where GraphQL is better because it's an explicit buffer send and an explicit buffer return. The number of roundtrips and what maps how is not hidden.

GraphQL has its own mess of poorly considered features, but I don't think Cap'n Web survives prolonged contact with reality because of how implicit and magical everything is.

When you make an abstraction like this, it needs to work ALL THE TIME, so you don't have to think about it. If it only works in demo examples written by developers who know exactly when the abstraction breaks, real devs won't touch it.

TLDR: It looks like an operating system so it can get viral, like get an article on HN and so on.

You need some sort of critical majority that cares about deeper aspects of whatever (over superfice) or else everyone focuses on superfice, as that's what all the public debates are about.

But lacking that, lisp could improve its syntax (did that, BTW not hard, just added some containers and a few standard infix operators, which all compile to lists).

Microsoft just took what they had and directly translated it to XML. It's not intentionally messy, it's just a big corporation with old product acting like it.

If you wanted a more equitable experience for all you wouldn't just limit the high-end users, but return the money to low-end users.

Charging a low flat fee per use and still warning when certain limits hit is possible. But it's market segmentation not to do it. Just charge a flat fee, then lop off the high-end, and you maximize profit.

We're zooming into their miiiiiind!

That's just how the Babel Fish works. It translates perfectly, but it tacks on a random accent.

Every time someone mentioned "days" or "months" or "years" in Andor I had to mentally zap my brain not to think about how it doesn't make a sense across a galaxy.