Vincent Carchidi
@vcarchidi.bsky.social
Defense analyst. Tech policy. Have a double life in CogSci/Philosophy of Mind. (It's confusing. Just go with it.)
https://philpeople.org/profiles/vincent-carchidi
All opinions entirely my own.
https://philpeople.org/profiles/vincent-carchidi
All opinions entirely my own.
Alright now we're getting somewhere
November 18, 2025 at 10:56 PM
Alright now we're getting somewhere
I was honestly considering putting a line where I basically say "we shouldn't expect this to happen" but then why the hell am I writing it
November 18, 2025 at 10:53 PM
I was honestly considering putting a line where I basically say "we shouldn't expect this to happen" but then why the hell am I writing it
a mental object ie mobject
November 18, 2025 at 10:16 PM
a mental object ie mobject
I hear both points. On the second tho, if Gemini's answers *were* wrong, and Dave was not an expert, how would he know?
November 18, 2025 at 9:49 PM
I hear both points. On the second tho, if Gemini's answers *were* wrong, and Dave was not an expert, how would he know?
My concern has always been that they are only (let's just say) expert-level if the person using them is expert-level, if that makes sense.
November 18, 2025 at 9:38 PM
My concern has always been that they are only (let's just say) expert-level if the person using them is expert-level, if that makes sense.
Also can be a good one. Asking for critiques to see if I've forgotten something is another.
November 18, 2025 at 9:37 PM
Also can be a good one. Asking for critiques to see if I've forgotten something is another.
This is fair, but I gotta be honest, the company I work for has the enterprise versions of ChatGPT and Gemini, and the most I can really squeeze out of them (for work only here) are better web search.
Which *is* useful but I wouldn't call it life changing.
Which *is* useful but I wouldn't call it life changing.
November 18, 2025 at 9:31 PM
This is fair, but I gotta be honest, the company I work for has the enterprise versions of ChatGPT and Gemini, and the most I can really squeeze out of them (for work only here) are better web search.
Which *is* useful but I wouldn't call it life changing.
Which *is* useful but I wouldn't call it life changing.
Agree about bluesky, but do you think that's true for Twitter/X? My sense is that Twitter was particularly influential for SV, both within AI firms and the VC types.
I'd have to dig it up, but I recall reading from a former OpenAI employee that Twitter vibes were a big factor for the company.
I'd have to dig it up, but I recall reading from a former OpenAI employee that Twitter vibes were a big factor for the company.
November 18, 2025 at 7:28 PM
Agree about bluesky, but do you think that's true for Twitter/X? My sense is that Twitter was particularly influential for SV, both within AI firms and the VC types.
I'd have to dig it up, but I recall reading from a former OpenAI employee that Twitter vibes were a big factor for the company.
I'd have to dig it up, but I recall reading from a former OpenAI employee that Twitter vibes were a big factor for the company.
You have to study it, ascertain whether the data you are collecting from it is best explained by appealing to something humans might already do. That brings you to areas of cognitive science, including (but not only) linguistics.
It's a heavy lift and most people stop at the initial observations.
It's a heavy lift and most people stop at the initial observations.
November 18, 2025 at 5:34 PM
You have to study it, ascertain whether the data you are collecting from it is best explained by appealing to something humans might already do. That brings you to areas of cognitive science, including (but not only) linguistics.
It's a heavy lift and most people stop at the initial observations.
It's a heavy lift and most people stop at the initial observations.
this, and the pessimists see this. We all agree that it looks like this.
But the results of an experiment have to be interpreted: what is the underlying process by which those sentences are constructed? Is it human-like?
Maybe! But you *cannot know that from merely personal interactions with it.*
But the results of an experiment have to be interpreted: what is the underlying process by which those sentences are constructed? Is it human-like?
Maybe! But you *cannot know that from merely personal interactions with it.*
November 18, 2025 at 5:34 PM
this, and the pessimists see this. We all agree that it looks like this.
But the results of an experiment have to be interpreted: what is the underlying process by which those sentences are constructed? Is it human-like?
Maybe! But you *cannot know that from merely personal interactions with it.*
But the results of an experiment have to be interpreted: what is the underlying process by which those sentences are constructed? Is it human-like?
Maybe! But you *cannot know that from merely personal interactions with it.*
One thing, related to Jefferson's remark, that I think has gotten lost is this: an AI model is effectively an experiment. The results of that experiment are whatever the model outputs triggered by an input.
GPTs of a certain size output things that look like sentences. I see this, the optimists see
GPTs of a certain size output things that look like sentences. I see this, the optimists see
November 18, 2025 at 5:34 PM
One thing, related to Jefferson's remark, that I think has gotten lost is this: an AI model is effectively an experiment. The results of that experiment are whatever the model outputs triggered by an input.
GPTs of a certain size output things that look like sentences. I see this, the optimists see
GPTs of a certain size output things that look like sentences. I see this, the optimists see
It's often said that Turing anticipated many of the objections to AI that we debate today. Often true. So did critics before the rise of digital computers, from the 20th century and further back.
November 18, 2025 at 5:34 PM
It's often said that Turing anticipated many of the objections to AI that we debate today. Often true. So did critics before the rise of digital computers, from the 20th century and further back.
Nor do I blame him - I've come around to the idea that the Imitation Game was a thought experiment for Turing, and the 1950 paper was likely tool of propaganda, which has the virtue of explaining why the paper is structured so strangely.
www.tandfonline.com/doi/epdf/10....
www.tandfonline.com/doi/epdf/10....
Galilean resonances: the role of experiment in Turing’s construction of machine intelligence
You have to enable JavaScript in your browser's settings in order to use the eReader.
www.tandfonline.com
November 18, 2025 at 5:34 PM
Nor do I blame him - I've come around to the idea that the Imitation Game was a thought experiment for Turing, and the 1950 paper was likely tool of propaganda, which has the virtue of explaining why the paper is structured so strangely.
www.tandfonline.com/doi/epdf/10....
www.tandfonline.com/doi/epdf/10....
it would be doing no more than telegraphic systems do already."
- Geoffrey Jefferson, 1949, 'Mind of Mechanical Mind,' one of the papers that Turing (1950) took pains to respond to.
It's not clear to me that Turing actually responded to this argument, accusing Jefferson of risking solipsism.
- Geoffrey Jefferson, 1949, 'Mind of Mechanical Mind,' one of the papers that Turing (1950) took pains to respond to.
It's not clear to me that Turing actually responded to this argument, accusing Jefferson of risking solipsism.
November 18, 2025 at 5:34 PM
it would be doing no more than telegraphic systems do already."
- Geoffrey Jefferson, 1949, 'Mind of Mechanical Mind,' one of the papers that Turing (1950) took pains to respond to.
It's not clear to me that Turing actually responded to this argument, accusing Jefferson of risking solipsism.
- Geoffrey Jefferson, 1949, 'Mind of Mechanical Mind,' one of the papers that Turing (1950) took pains to respond to.
It's not clear to me that Turing actually responded to this argument, accusing Jefferson of risking solipsism.
an improvement on the typewriting monkeys which would accidentally in the course of centuries write Hamlet. A machine might solve problems in logic, since logic and mathematics are much the same thing. If the machine typewrites its answers, the cry may rise that it has learned to write, when in fact
November 18, 2025 at 5:34 PM
an improvement on the typewriting monkeys which would accidentally in the course of centuries write Hamlet. A machine might solve problems in logic, since logic and mathematics are much the same thing. If the machine typewrites its answers, the cry may rise that it has learned to write, when in fact
Seeing enough of these things (collaborations, strategic partnerships, MoUs, etc.), you start to appreciate when companies actually do something *worth* publicizing.
November 18, 2025 at 2:26 PM
Seeing enough of these things (collaborations, strategic partnerships, MoUs, etc.), you start to appreciate when companies actually do something *worth* publicizing.
But tbh, I don't know how much actual power the nonprofit has over OpenAI the for-profit company. And I don't know if they know either.
November 18, 2025 at 2:05 PM
But tbh, I don't know how much actual power the nonprofit has over OpenAI the for-profit company. And I don't know if they know either.
One of the reasons why debates about phenomena like consciousness seem unending is because the only way to observe it is to be a subject with consciousness. So far as we know, only humans have consciousness. We can only be sure of our own. We assume others are because we're made similarly. 🧵
i'm made of magic meat i'm made of magic meat i'm made of magic meat i'm made of magic meat i'm made of magic meat i'm made of magic meat i'm made of magic meat i'm made of magic meat i'm made of magic meat i'm made of magic meat i'm made of magic meat i'm made of magic meat i'm made of magic meat i
November 18, 2025 at 3:11 AM
Could a non-human being replicate it? Sure! Could a computational system - whose operations seem to fall within the range of determinacy and randomness - replicate it? I have my doubts.
At the least, I think it sheds like on something we just don't understand about humans
bsky.app/profile/vcar...
At the least, I think it sheds like on something we just don't understand about humans
bsky.app/profile/vcar...
Sharing a new preprint on AI and philosophy of mind/cogsci.
I make the case that human beings exhibit a species-specific form of intellectual freedom, expressible through natural language, and this is likely an unreachable threshold for computational systems.
philpapers.org/rec/CARCBC
I make the case that human beings exhibit a species-specific form of intellectual freedom, expressible through natural language, and this is likely an unreachable threshold for computational systems.
philpapers.org/rec/CARCBC
Vincent Carchidi, Computational Brain, Creative Mind: Intellectual Freedom as the Upper Limit on Artificial Intelligence - PhilPapers
Some generative linguists have long maintained that human beings exhibit a species-specific form of intellectual freedom expressible through natural language. With roots in Descartes’ effort to distin...
philpapers.org
November 18, 2025 at 3:11 AM
Could a non-human being replicate it? Sure! Could a computational system - whose operations seem to fall within the range of determinacy and randomness - replicate it? I have my doubts.
At the least, I think it sheds like on something we just don't understand about humans
bsky.app/profile/vcar...
At the least, I think it sheds like on something we just don't understand about humans
bsky.app/profile/vcar...
Put another way: even if we look internally, and (rightly) assume there is a physical basis for every word a person speaks, it is not clear how individuals *use* language in a manner that coheres with the thoughts of others.
Language I believe is computable. Language use may not be.
Language I believe is computable. Language use may not be.
November 18, 2025 at 3:11 AM
Put another way: even if we look internally, and (rightly) assume there is a physical basis for every word a person speaks, it is not clear how individuals *use* language in a manner that coheres with the thoughts of others.
Language I believe is computable. Language use may not be.
Language I believe is computable. Language use may not be.
Human language use, in other words, appears free of any *fixed and identifiable* set of stimuli, frequently novel in what it expresses, yet appropriate to the mental states of others who hear or otherwise interpret the remarks. It appears to be neither deterministic nor random, yet appropriate.
November 18, 2025 at 3:11 AM
Human language use, in other words, appears free of any *fixed and identifiable* set of stimuli, frequently novel in what it expresses, yet appropriate to the mental states of others who hear or otherwise interpret the remarks. It appears to be neither deterministic nor random, yet appropriate.
I think there is. Not doing details here, but it is not clear to me why human actions - primarily expressible through natural language - "fit" the situations in which they occur while not evidently being "caused" by the situations in which they occur.
November 18, 2025 at 3:11 AM
I think there is. Not doing details here, but it is not clear to me why human actions - primarily expressible through natural language - "fit" the situations in which they occur while not evidently being "caused" by the situations in which they occur.