I’m wary of emotive terms like BS, slop, hype (not to mention woke, moral panic, alarmist). People enjoy flinging them out and instead of trying to understand what is the case they shift focus to outrage at shortcomings in opponents’ thinking.

In a coding environment (e.g. getting help from Claude inside Cursor) either the code works as expected or it doesn’t. That means the BS gets corrected by reality and Claude tries again; or we try again together. Maybe other possible forms of ‘corrective world’ vs just chat bubble.

A character in a novel can lie, obviously without sentient intention as they too are just words: what makes it lying is the interaction in context. LLM chat personas are characters whose words are generated on the fly. No reason at all they could not be configured to lie.

LLMs present an animated aggregation of human language behaviour: they certainly can lie, just as they can be polite, along with anything else that appears on the surface of human speech.

Why “say no to a minority government”?

There’s not much point in characterising that as delusion, enough just to say you think he’s wrong. Same trope as climate “alarmism” — treating other people’s errors as pathological.

If a government believes that its true enemies are internal then naturally they trade away external security in order to arm themselves against their own country.

probably = wine study, another way to induce tremors.

You can minimise anything by pointing to its constituents: but what’s the point of saying a chair is essentially an arrangement of atoms?