Maybe it’s parasocial (derogatory) of me to imagine that anyone thinks we’re a community in the first place. Anyway. It’s been nagging at me for a while.

What *are* we doing here?

I want to reflect on that. What are we to one another? What obligations do we have to one another? To hear one another? Is communication and contact on here multi-directional? Are we both seeing and being seen?

The educator training I have had has been SO BAD that I honestly 100% believe this poster. Teacher training reform in general needs to have happened yesterday. (And yes, I know there are amazing schools of education out there. However. In my experience, the truly terrible ones are far too common)

I’m captivated by the fluid dynamics turbulence metaphor, but I don’t yet see how to apply it to her own examples. Wouldn’t she try to establish a kind of turbulence—difficulty that could account for them?

…buckets would hallucinating a Venmo account fall into? Or just failing to check prices on the cubes? Or the way the multimodal LLMs could not tell how many times the lines crossed? I’m wondering whether I’m missing something.

I loved reading this, and I have so many questions. Not sure you can answer this, but I’m interested in the disconnect between the types of mistakes the Claude drinks vendor was making, and her list of sources of difficulty or jaggedness (reward signals, context windows, etc). Which of those…

It's only a matter of time before humanities departments will be forced to accept AI-authored assignments, as part of revised university policy to cooperate with these billionaires. It's already happening, and our response needs to be decisive. Because our students' ability to *think* is at stake.

I mean—maybe! Certainly seems to be. But I still think it’s worth saying. I currently don’t use LLMs at all, and so clearly, at this moment in history, it’s not a foregone conclusion. Plus, if people reflect on what could could be at stake cognitively, maybe they’ll approach the tech with caution

Yeah, I’m definitely assuming that, humans being humans and competitive pressure being competitive pressure, there is very little halfway with LLMs. If people use them, I’m imagining they’ll be used way too much, and it would be better socially if they didn’t exist at all.

So decoupling basic acts of communication from the human thought process is counterproductive. What you need is the thought, and to get the thought, you have to write. Which is why I think framing written words as merely tools is limiting and missing something crucial about the problem with LLMs.

Again, in the short term, maybe the effects of cognitive offloading aren’t obvious, but I think the act of thinking matters, and you don’t think the same way when you’re not composing. You often don’t know what you think until you’ve tried writing it (I would argue).

But what’s in the thing presented matters, right? On some level? It’s like ground zero of what’s in our brains and what matters to us. So how can you evaluate the presentation and have a vote if you haven’t done the thinking?

Okay—well, I would say that the PowerPoint is pretty much meaningless if a human didn’t make it. Maybe not at first, but over time. Obviously written documents have uses and meaning, but I’m saying the most important way they do is as a trace of thought, and as an act of communication.

Not a latrine. It’s more: claims about latrines, PowerPoints about latrines, descriptions of latrines—all of which have been radically decoupled from actual human experience.

Your thread assumes an LLM is a tool, but I’m not convinced it is. What is the purpose of this tool? What does it create? And does that product, that output, have any meaning beyond the human-to-human thought process it instantiates? Take the human out, and what do you have?