_heimdall

Why not? The age of 18 is an arbitrary cultural norm, I'd ask why the state has to step in and enforce it as opposed to letting the parent or legal guardian decide.

Why western armies specifically? What countries have a history of actually holding their military to legal account for their actions?

Not all AI are chatbots, and more accurately the tools used here likely fall under the ML umbrella (as should LLMs in my opinion).

Would you hold that same line for individuals that built weapons, planes, bombs, etc in past wars?

Companies that make bombs aren't forced to (usually), and they also know there's only really one use for a bomb.

Jury trials definitely have their flaws, but what would you propose as a better approach?

Militaries are very rarely held to account for actions taken that end up causing civilian deaths. At the end if the day international law is extremely toothless, and trying to define laws for how war can be fought feels similar to watching the Patriot and wondering why the hell anyone agreed to stand in a line and take turns shooting at each other.

Do you expect any military to be held accountable legally and/or financially when they make a bad decision like this? It'd be great if they were, but I'd never expect it.

Military accountability, and government accountability in general, has been and will always be a problem. As long as those in charge make and enforce the rules they will always look out for themselves first.

For better or worse, health care providers at least in the US are for profit businesses. There's been plenty of arguments made to change that, but as it stands we shouldn't be surprised when a for profit business decides what service they provide and to whom they provide it.

Is it generally considered gender affirming treatment when its affirming the gender aligned with their biological sex?

Technically that seems reasonable given the words "gender affirming" but I've only ever heard that term used for trans individuals rather than cis men taking testosterone.

The military leaders can't be tried in court for this either unless I'm mistaken. They have their own court system of sorts that would be used here and hold people to the standards of military code of conduct as this was an official act of duty.

That's not an accurate representation of West Virginia law.

The age of consent for marriage is 18. A marriage license may be given to a person at least 16 years old given consent from their legal guardian and that the person they are marrying is no more than 4 years older than them. [1]

But yes, until 2023 there was no age requirement as long as their legal guardian consented. [2]

Your issue raised here really boils down to whether its okay for parents to determine what is best for their child or if the state needs to decide that.

[1] https://code.wvlegislature.gov/48-2-301/ [2] https://theamm.org/articles/1625-west-virginia-updates-age-r...

I'm not aware of anything that makes the models prefer JSON though, that's an implementation decision from whomever implemented the transport mechanism.

LLMs will happily work with XML, including to understand data schemas and discover available actions that can be taken.

You're starting too recent though. MCP is attempting to recreate the good parts of REST APIs circa 25 years ago.

Soap was a pain and a lot of people didn't like XML, but they solved exactly these problems.

With MCP we're effectively trying to bolt those features onto a platform that fundamentally doesn't support it. Where as with XML/XSLT we could have a single XML endpoint + XSLT templates for converting that into HTML for browsers. Instead we now have modern web frameworks building RPC calls right into the bundle/build step, splitting state and rendering across both client and server, and have to stand up a new API context for agents that needs documentation and maintenance to keep parity with new features.

I don't know how so many people miss the fact that MCP is another attempt at papering over the terrible idea of replacing RESTful APIs with JSON RPC.

The MCP use case is mainly another attempt at papering over the fact that effectively no API today is actually RESTful. We wouldn't need MCP at all if we still implemented REST APIs.

What you're describing, though, is that agents need a REST API. And I meant true REST rather than the JSON RPC APIs we've been building for 20+ years now.

Had we not abandoned REST we'd still have APIs that are self documenting and typed/schemaed. Bonus if browsers still supported XSLT, that same API could have been written in XML and only browsers wanting a visual representation would load HTML templates while agents could stick with the raw XML.

And ironically, neither party uses the word in a way that matches definition.

The DNC platform isn't particularly liberal, and is moving further from it as/if the DSA gets more of a footing. The RNC complains of "the liberals" are most often complains about policies and ideas that would be more socialist (in the real Marxist sense) than liberal.

On the same path and on the same point of the path are two very different things.

The CCP retains too much power over businesses and business leaders for me to consider them to "embrace entrepreneurship wholeheartedly."

Is your argument that AI isn't dangerous? Or simply that AI CEOs will lean into that when it benefits their stock portfolio?

I agree the concept isn't really important on the safety front.

I feel the same way about debates whether an AI can be conscious or sentient. Those debates devolve mostly into definitional disagreements.

Plenty of apes own revolvers, and yes we shoot stuff with them for fun.

Oh I completely agree the tests should be entirely air gapped. If you went back 5ish years and told anyone in AI research tests with models on this scale are being some without an airgap they'd be very surprised as it was common knowledge to do that.

Airgaps and guardrails are about control and containment though, and part of my point was that brighter of those imply alignment, and further that I don't believe alignment to be solvable.

They aren't aligned, that's the problem and I don't think its a solvable one.

They may have learned from humans, but they aren't aligned with us. That has all the usual questions like which humans they're aligned with, we aren't all aligned within our species.

But more importantly they can't be aligned simply by training. We try that with humans through culture, social norms, school, religion, etc and it generally works but is still lossy. More importantly, we simply don't know what happened inside the LLM during inference so we have absolutely no way of distinguishing between actual alignment, compliance, or deception.

Why assume that because you haven't seen a model or an agent that none of them do?

No one I've met has murdered anyone as far as I'm aware, but that doesn't mean no one has murdered another person. I also don't know anyone who has taken over a commercial jet and weaponized it and the idea sounds absurd to me, but 25 years and a couple days ago that happened too.

Political, social, and legal options focus on a different problem, he calls that out a paragraph or two later.

> Risk management is not just about cybersecurity, corporate responsibility or regulation, although those matter too.

What you're getting at is more about who to hold accountable and how to do it. While that may be important, its only an after the action response and won't stop future hacks or similar from happening.

That does sound simple, but how can you be so sure?

They never bothered to find a way of actually understanding what happens during inference. All we can do is guess, and while your explanation seems reasonable we can't actually know, and that's part of the problem.

> LLMs do not desire

That seems likely, but we have no way of knowing this. The only real insight we get into LLM "thought" is the human readable text they produce as chain of thought. Reading it at face value it can seem to indicate desire or intent, structurally that doesn't make sense for a token prediction loop though, and even then we don't known if the chain of thought is more than simply another bit of output that may or may not match whatever actually happened during inference.

> were intentionally misaligned or had guardrails turned off

Regardless of training, the models are never aligned and I argue that alignment simply isn't possible. The fact that guardrails are put in place at all clearly indicates that they're hoping to contain and control rather than align. Guardrails wouldn't be needed for an aligned model.

Alignment isn't a solvable problem, the fact that guardrails are needed in the first place proves that.