Conversation

Notices

1. Embed this notice
   Paul Cantrell (inthehands@hachyderm.io)'s status on Wednesday, 01-Jul-2026 12:35:36 JST Paul Cantrell

   RE: https://mastodonapp.uk/@JohnSullivan/116842603611650876

   “Some of the company’s most experienced personnel left before all of their accumulated knowledge could be fully transferred into Ford’s automated systems…”

   This is apparently the new line for blaming companies for AI failures: you didn’t fully upload the brains of your senior people before firing them! You have to complete the extraction process first! Suck them dry, •then• discard them!

   Machine learning does not actually work that way, but…in this FOMO-driven industry, anything goes to blame the customer.

   • Minoru Saba repeated this.
   • Embed this notice
     Paul Cantrell (inthehands@hachyderm.io)'s status on Wednesday, 01-Jul-2026 12:51:29 JST Paul Cantrell

     in reply to

     There’s a common misconception about how learning works, about what knowledge •is•. In this misconceived metaphor:

     - the mind is a container,
     - knowledge consists of items (or maybe a substance) in this container, and
     - learning consists of transferring those items from the mind-container of the teacher into the mind-container of the student.

     This is wrong — knowledge doesn’t work this way, learning doesn’t work this way, brains don’t work this way — but we constantly fall into the trap of imagining that this is reality.

     2/

   • Embed this notice
     Paul Cantrell (inthehands@hachyderm.io)'s status on Wednesday, 01-Jul-2026 13:17:09 JST Paul Cantrell

     in reply to

     • V

     @miss_rodent

     Right. Because knowledge simply does not work this way. Doesn’t. It’s not a pile of facts.

   • Embed this notice
     V (miss_rodent@girlcock.club)'s status on Wednesday, 01-Jul-2026 13:17:11 JST V

     in reply to

     @inthehands Yeah... "all of their accumulated knowledge" can't be transferred in the first place - they almost certainly can't even articulate large amounts of it, even while they use it...
     Even if you extracted all the *facts*, you'd still have a machine that's at best ... kinda okay at trivia, and still bad at actually doing things.

   • Embed this notice
     Iris Young (he/they/she) (PhD) (iris@neuromatch.social)'s status on Wednesday, 01-Jul-2026 13:32:21 JST Iris Young (he/they/she) (PhD)

     in reply to

     @inthehands also notice the verb choices: the people left! before the work was done! I don't think many left by choice, even where it looks like it on paper.

   • Embed this notice
     Paul Cantrell (inthehands@hachyderm.io)'s status on Wednesday, 01-Jul-2026 13:50:19 JST Paul Cantrell

     in reply to

     This is closely related to / partly overlapping with what linguists call the “conduit metaphor,” in which language is a container into which the speaker/writer places meaning and from which the listener/reader extracts meaning:

     https://en.wikipedia.org/wiki/Conduit_metaphor

     Study the example sentences in that article. Notice how they are •metaphorical•, but also notice how normal, how •literal• they sound to our ears.

     3/

   • Embed this notice
     Glyph (glyph@mastodon.social)'s status on Wednesday, 01-Jul-2026 14:26:53 JST Glyph

     in reply to

     @inthehands my go-to for explaining why the conduit metaphor is wrong is wittgenstein’s sprachspiel, but I don’t know a more empirically grounded source

   • Embed this notice
     Paul Cantrell (inthehands@hachyderm.io)'s status on Wednesday, 01-Jul-2026 14:34:20 JST Paul Cantrell

     in reply to

     • Glyph

     @glyph

     I love his discussion of that idea, but it’s hard to summarize convince to those who haven’t worked through the details! It’s a thing I bring up somewhat clumsily when talking about the difference between programming languages and natural languages.

   • Embed this notice
     Paul Cantrell (inthehands@hachyderm.io)'s status on Wednesday, 01-Jul-2026 14:48:03 JST Paul Cantrell

     in reply to

     What I’m talking about is not exactly the Conduit metaphor in which •language• is the idea-container during idea transit, but rather a related metaphor in which the mind is the idea-container for long-term storage (e.g. “Get it into your head,” “Tried to remember but I’d lost it,” “My brain is full,” etc etc).

     It’s a handy way to talk, but it does mislead us.

     (I’m sure somebody’s worked out the linguistics side of this, not to mention the psych + neuroscience sides, but that’s well beyond my knowledge.)

     4/

   • Embed this notice
     Real Programmer (sleepyfox@hachyderm.io)'s status on Wednesday, 01-Jul-2026 19:17:34 JST Real Programmer

     in reply to

     @inthehands Every metaphor is a double-edged sword.

   • Embed this notice
     Real Programmer (sleepyfox@hachyderm.io)'s status on Wednesday, 01-Jul-2026 20:32:47 JST Real Programmer

     in reply to

     @inthehands Metaphors we live by.

   • Embed this notice
     George Dinwiddie (gdinwiddie@mastodon.social)'s status on Thursday, 02-Jul-2026 01:42:12 JST George Dinwiddie

     in reply to

     @inthehands
     Learning often comes from “prediction error, a gap between what you had expected and what your senses reported was on offer at that moment”
     See the book, Uncertain by Maggie Jackson.

   • Embed this notice
     Paul Cantrell (inthehands@hachyderm.io)'s status on Thursday, 02-Jul-2026 01:43:44 JST Paul Cantrell

     in reply to

     • V

     @miss_rodent

     This is especially true about how people in business •understand• what LLMs are and what they're good for. The technology in the abstract doesn't necessarily rest on those assumptions — detecting and reproducing patterns in language is not something we necessarily have to understand as “learning” or “thinking” at all — but the technology as sold certainly does.

   • Embed this notice
     V (miss_rodent@girlcock.club)'s status on Thursday, 02-Jul-2026 01:43:45 JST V

     in reply to

     @inthehands Yeah, LLMs are built on some... Really shaky ground in terms of the underlying philosophy-of-mind their claims depend on.
     Even ignoring the psychology side of it - they are depending on a computational theory of mind, and taking the 'brain as computer' metaphor entirely literally.

   • Embed this notice
     Paul Cantrell (inthehands@hachyderm.io)'s status on Thursday, 02-Jul-2026 01:57:54 JST Paul Cantrell

     in reply to

     • Aaron

     @hosford42

     This understanding goes a long way and is useful, but unravels under enough examination. It’s worth reading Wittgenstein’s critique of this idea in _Philosophical Investigations_, where he talks about the “language game.”

   • Embed this notice
     Aaron (hosford42@techhub.social)'s status on Thursday, 02-Jul-2026 01:57:55 JST Aaron

     in reply to

     @inthehands Craft your words carefully, to map out your ideas and thereby shape the thoughts of another and synchronize your perspectives.

     Just playing with alternative metaphors. Really enjoyed the article you linked to!

     I actually think of language as a serialization protocol. You take a complex structure and encode it using grammatical rules as a sequence. The listener then decodes (what we hope is) a similar structure using supposedly the same rules. The obvious source of communication failure is when the structural encoding rules, learned entirely by example, differ between two people. The less obvious one is when it come times to integrate that structure into the larger semantic model: The structure might integrate at the wrong place. And the most frustrating one is when the listener's mental model doesn't even have a similar enough shape for the new structure to be incorporated. This is when you find yourself staring at the person, trying to understand how the words they just uttered can possibly make sense (metaphor!) to them. Often in the last case, you don't even know how to ask clarifying questions (another metaphor!) to make any progress.

     I built an NLU engine that worked on these principles, back before LLMs existed, and it could actually connect the dots (metaphor!) between things you said to it, integrating the information so that it could then answer questions that bridged (metaphor!) across them. If only I had had billions of dollars in investment, instead of being some autistic guy that nobody took seriously, we might have had language models today that didn't hallucinate.

   • Embed this notice
     Charl van der Walt 🌻🇵🇸 (charlvdwalt@infosec.exchange)'s status on Thursday, 02-Jul-2026 02:08:03 JST Charl van der Walt 🌻🇵🇸

     in reply to

     @inthehands this exact idea is discussed as an ideological objective in the book Muskism. Worth a read. https://a.co/d/0dSua7dM

   • Embed this notice
     Paul Cantrell (inthehands@hachyderm.io)'s status on Thursday, 02-Jul-2026 02:08:03 JST Paul Cantrell

     in reply to

     • Charl van der Walt 🌻🇵🇸

     @charlvdwalt

     Noted.

   • Embed this notice
     Paul Cantrell (inthehands@hachyderm.io)'s status on Thursday, 02-Jul-2026 02:32:37 JST Paul Cantrell

     in reply to

     • Andy Wootton

     @woo

     “Disciplinary boundaries are lines we draw in the ocean” is my Paulism for this.

   • Embed this notice
     Andy Wootton (woo@fosstodon.org)'s status on Thursday, 02-Jul-2026 02:32:38 JST Andy Wootton

     in reply to

     @inthehands I believe I have ADHD so, more than most people, I depend on the connections between items, even when I can't remember what the items are called. I think and make notes in graphs of connected ideas but I've noticed that others think, or perhaps communicate, in trees or in linear narratives about any 'subject'. I don't accept that there are separate subjects. "Everything is deeply intertwingled", as Ted Nelson wrote in 1974.

   • Embed this notice
     Marsh Gardiner 💡🐝🔧 (earth2marsh@hachyderm.io)'s status on Thursday, 02-Jul-2026 04:07:11 JST Marsh Gardiner 💡🐝🔧

     in reply to

     @inthehands This whole thread is 🔥. One thing it makes me wonder about is how much of this stems from the challenge of attempting to scale education. Learning is intensely personal--I'm thinking about zen koans and of "… the student was enlightened."