Mandelson was given secret spreadsheet to vet left-wing MPs in Labour power struggle Whistleblower's revelation shows extent of his direct involvement in bitter battle to gain control of the party ahead of 2024 election

Your regular reminder Peter Mandelson handpicked every new Labour MP in 2024.

inews.co.uk/news/politic...

If you can substitute "hungry ghost trapped in a jar" for "AI" in a sentence it's probably a valid use case for LLMs. Take "I have a bunch of hungry ghosts in jars, they mainly write SQL queries for me". Sure. Reasonable use case.

"My girlfriend is a hungry ghost I trapped in a jar"? No. Deranged.

choose your faction:
- I made a vape that's also a hammer
- Hammers are going to become god & change everything
- I tried holding the hammer by the head and it didn't work, this is bullshit
- Hammers are useful tools sometimes
- Hammers? You mean murder-tools, as seen in the documentary The Raid 2?

Prosentences or bust!

remember when

Fascinating paragraph from Scott Alexander, where he admits he’s the dog that caught the car

As the crowdfunder for the first radical bookshop icon Newcastle for 40 years gets closer to the end I’m making the bold move of calling on all the academics I sort of vaguely know and also those I don’t to share the link for our crowdfunder: www.crowdfunder.co.uk/p/booksfromb...

This is the most UK coded experience it is possible to have.

The summoning table from Lamentations of the Flame Princess always sticks with me. A delightful injection of absurdity in an otherwise grimdark OSR D&D clone.

I can send you an essay that lays out the basic picture of you like?

I’m not that worried about this, as my position is much weaker than Kant’s in several ways (and closer to Heidegger’s in some). I think we can live up to the ideal quite poorly and yet still count as autonomous in virtue of an in principle capacity for (rational) change in motivation.

A pith off! 😆

And well, it's those formal features that I'm interested in. That which enables us to want to create sustainable small scale nuclear fusion despite never having observed it, because we have developed the conceptual repetoire to theoretically represent it.

That's describing something at the level of the 'program' itself for want of a better word, than it is sideloading a new program entirely. Closer to adding a plugin, but then, the plugin framework would have to satisfy certain formal features in order to be sufficiently extensible.

It's always tricky to precisely pin down modal language in giving definitions like this, and I doubt I will satisfy your Cynicism here, but I'm trying to describe a computational process that can (in principle) endogeneously revise and extend the types of information it can process.

So, I suppose you could argue that anything capable of general purpose computing can feasibly emulate anything else if you can load a program onto it and expand its memory/storage on the fly, but this is a little like saying a cat can become a dog if we can rearrange its atoms in the right way.

I could push back on the point about language leading to metaphysics with Sellars' argument against the myth of the categorical given, but I should probably stop somewhere!

The differences might come out in the wash, as my view of the entry-exit component is more or less psychofunctional (capacities to attend to and modulate internal states), but these need to be linked to inference to be a concept proper. This is basically Kant's difference between schema/concept.

Where's it's ability to extend what's it's capable of representing (and thereby aiming for)?

To quote my pithy definition: “a person is a dynamic system the behaviour of which is driven by the divergence between two consistent, extensible, and revisable representations of the world as a whole: one of the way it is (real), and one of the way it should be (ideal).”

What I take to be my main original contribution is the idea that general intelligence and autonomous agency are in some sense dual: the former requires in principle revisability of any of one's theoretical commitments, and the latter requires in the same for practical commitments.

Just to be clear, I define the capacity for reflection in terms of holistic revision, rather than the other way around, so I'm not sure Kornblith's worries will apply. There's an extremely condensed account in the 'Geist in the Machine' essay I linked, but there's a more elaborate one in my book.

I'll take a look too, but I've yet to find a framing of the problem that suggests there's more going on than the infelicitousness of Nagellian language.

I'll take a look. I've generally avoided PCS stuff just because I'm a Sellarsian/Brandomian inferentialist about concepts, and so think that whatever representational differences are going on here have to be explained in terms of inferential role (including language-entry/exit) anyway.

As for autonomy, I've got a somewhat nuanced account of it, and I've already made this reply thread too long. But, roughly, I see it as the capacity for open-ended reflection on one's priorities, which essentially means ongoing holistic revision. It's not about where you start, but how you evolve.

Knowing 'what it's like' is then a form of practical know-how rather than a special sort of knowing-that (one of the few things I agree with David Lewis about). We can happily talk about 'contents' of consciousness, even structural patterns thereof, without rendering them metaphysically mysterious.

My rough proposal is that we have second-order reliable differential responsive dispositions that let us report on the informatic content of the mechanisms underpinning the first-order ones, but the resulting reports are used in actively calibrating our perceptual capacities.

...saying we're reporting a special sort of fact/property ('what it's like'), and postulating a special sort of knowledge by acquaintance, is to bake an introspective framing into the explanatory problem that makes it irresolvable from the outset, IMHO. We're better off extending Sellars' strategy.

In conversations like:

'What's it like to have astigmatism?',
'Lights give off these alternating, angular streaks at night'

or

'What's it like to read Ligotti?',
'Oh, a bit like reading Lovecraft, but more depressing, in a Kafkaesque sort of way...',

We are reporting something, but...

But we obviously do a lot more than this. This is also where late Wittgenstein falls down, as he is surprisingly incurious about the sheer richness of language we use to successfully communicate features of our internal lives. However, the Nagellian 'what it's like to...' only explains by reifying.