dcm

@dcm@social.sunet.se

Assistant Professor in Philosophy of AI at Umeå University, working on and at the foundations of the sciences of mind and cognition.

Co-host of Punching Cards, the podcast about all things intelligent: https://www.spreaker.com/show/punching-cards

Older posts accessible @dim. Searchable through tootfinder.

This profile is from a federated server and may be incomplete. Browse more on the original instance.

NicoleCRust, to random
@NicoleCRust@neuromatch.social avatar

Summer reading recs?!

I’m looking to compile a big pile of IRL books for summer beach reading.

Big fan of scifi, specfic, murder mysteries. Strong female characters = bonus points. Can’t deal with darkness. Nothing high brow or tedious. Nonfiction lover but that list is already bursting at the seams.

I also love long series - 3 body problem, wheel of time, foundation … those waves all have good memories attached.

If that brings anything to mind, please send the rec!!

dcm,

@NicoleCRust The Expanse sci-fi series is very good (I'm still halfway through, though), gripping and well-written :)

dcm, to random

Very saddened to hear about Daniel Dennett's death: he was one of the philosophers who most inspired my approach to the field, and I was lucky to have met him, if only briefly, early on in my career.

https://zirk.us/@DailyNous/112298517534566334

dcm, to ai

Very interesting (and very long thread) by @pluralistic on AI, LLM booby-traps, and 'reverse-centaurism'

https://mamot.fr/@pluralistic/112196495887778626

axbom, to random
@axbom@axbom.me avatar

Looking at the definition of AI by the Organization for Economic and Co-operation Development (OECD):

"An AI system is a machine-based system that, for explicit or implicit objectives, infers, from the input it receives, how to generate outputs such as predictions, content, recommendations, or decisions that can influence physical or virtual environments. Different AI systems vary in their levels of autonomy and adaptiveness after deployment."

Based on this, if a language model has no explicit or implicit objectives it is not an AI.

dcm,

@axbom is that possible, though? Given that the training involves objectives (next-token prediction, supervised RL), LLMs always have them. The only exception would be a non-trained Transformer (so not really a language model), but as soon as it is put to some use, it would count again, if I understand this correctly.

DrYohanJohn, to random

How seriously should we take "levels" of reality? I mean the parcellation of the universe into, say, the particle level, the chemical level, the cellular level, the organism, individual level, the social/ecological level etc?

Do levels interpenetrate?

Are these metaphysical matters or just pragmatic ones?

Stimulated by the great session on mental disorders just completed, hosted by @PessoaBrain , featuring Anneli Jefferson, @awaisaftab , @NicoleCRust , Alexey Tolchinky & @eikofried

dcm,

@NicoleCRust @DrYohanJohn @PessoaBrain @awaisaftab @eikofried What do you mean by nonphysical? I would think that insomnia and fatigue are also physical phenomena, they are just multiply realisable kinds. We would lose generalisation power by focusing only on few of their realisation bases, making them useful concepts for certain causal explanations.
I'm sorry that I missed the salon, it looks like it was great!

dcm,

@NicoleCRust @DrYohanJohn @PessoaBrain @awaisaftab @eikofried right, maybe this is a terminological issue: 'nonphysical' suggests adopting some sort of dualism, which I take is not what the view involves. Studying psychological kinds does not mean rejecting the idea that some questions need to be answered by appeal to neurobiological kinds too (or exclusively).
(Strictly speaking, 'neuron', 'synapse', 'neurotransmitter' are not physical concepts either, but rather (neuro)biological ones).

dcm,

@NicoleCRust @DrYohanJohn @PessoaBrain @awaisaftab @eikofried I think that that terminology may be misleading, the distinction being best captured by talking of abstract vs concrete concepts.

'Neuron' etc are not physical concepts in the sense that they do not figure in generalisations and explanations in physics. I mentioned this because of your implied contrast between physical vs psychological concepts: many biological concepts cannot be touched either (e.g., species, function).

dcm,

@NicoleCRust @DrYohanJohn @PessoaBrain @awaisaftab @eikofried I'm a philosopher, so the philosophical weeds feel like home! :)

dcm,

@NicoleCRust @DrYohanJohn @PessoaBrain @awaisaftab @eikofried I don't think all would understand that in the way you seem to intend it. It may look like you deny the existence of insomnia and fatigue; or think that they are nonphysical 'mind' stuff; or that they are just epiphenomena.
That statement moreover presupposes lots of debatable ontological issues, and indeed I don't really agree with it (though it certainly is a legitimate philosophical position to take).

Managed without isms! ;)

dcm,

@NicoleCRust @DrYohanJohn @PessoaBrain @awaisaftab @eikofried My own preference is to talk about kinds that belong to explanations and generalisation in different sciences, which may be related in complex ways, not necessarily through neat one-to-one reductions.

ernie, to random
@ernie@writing.exchange avatar

Man, the Zoom app on Linux made me so upset this morning.

dcm,

@ernie yep, I had similar issues and started logging into X11 whenever I wanted to use Zoom. With my move to KDE, which I think defaults to X11, I stopped having to do this.

dcm, to random

@phanpy, my favourite Mastodon web client, released a great new beta function: the catch-up window.

You can choose the time window you want to catch up on, and you get a nicely organised list of posts that you can browse by author, by whether it's an original, a boost, a reply, and order by date, replies, boosts, etc. All done with a very pleasant and intuitive design.

Brilliant stuff!

dcm,

@philippsteinkrueger for multicolumn you can find an option in the menu 'Shortcuts/Columns' (bottom right of the menu). There you can choose which columns to add, which instances they should come from, etc.

dcm, to ai

Cool news: a new, publicly-funded, open-access, triple-anonymous journal in Philosophy of AI!

More info here: https://journals.ub.uni-koeln.de/index.php/poai/about

elduvelle, to random
@elduvelle@neuromatch.social avatar

Who out there has a ? I really like the idea.. any feedback on them?

From: @e_mydata
https://mastodon.social/

dcm,

@elduvelle if your interest is in a smartphone that is neither Google nor Apple, in Europe there's also Volla Phone, which are produced in Germany, and have a custom, degoogled version of Android, so most applications work, occasionally with some tweaks needed. They also have a dual-boot function for other OSs.
I've had a first-generation one for about 3 years and am happy with it!
https://volla.online/en/index.html

dcm, to philosophy

Job opening in Philosophy!

We are hiring a doctoral student in Philosophy at our (great) department here in (lovely) Umeå. (Yes, hiring is the right word, doctoral students here are employees, so maybe we should say doctoral researchers instead?)

If you or someone you know would be interested, please let them know about this!

Deadline for applications is March 10th. Information on requirements, how to apply, etc., can be found here: https://www.umu.se/en/work-with-us/open-positions/doctoral-student-in-philosophy_672424/

dcm, to ai

It's relatively commonly recognised that AI is a somewhat misleading umbrella term that covers a variety of different scientific and non-scientific projects.

In a new preprint, I articulate, defend, and illustrate a central scientific project for AI that is somewhat neglected or vaguely recognised, which I call AI-as-exploration (taking the cue from a recent paper by @olivia, @Iris et al).

https://arxiv.org/abs/2401.07964

1/n

dcm, to random

First day of a conference on and social media at HumLab in Umeå, featuring @simonlindgren and @cfiesler, among others.

Lots of great stuff, such as @cfiesler's suggestion to write short blogs about one's own papers, both for outreach and for busy colleagues.

As I don't have a blog, I'm thinking of using Mastodon for that (I couldn't find two-way integrated macroblogging for the Fediverse), linking the threads on my website.

It would be nice to hear alternative ideas, though!

dcm,

A recent initiative in a similar spirit within philosophy is the Substack newsletter New Work in Philosophy, created by Marcus Arvan: https://newworkinphilosophy.substack.com/

deshipu, to random
@deshipu@fosstodon.org avatar

deleted_by_author

    •
    dcm,

    @troos @deshipu @UlrikeHahn The issue is not about knowing names, but rather about figuring out the features that characterise a phenomenon (the specific word you choose to name the phenomenon is typically irrelevant). In this case, thinking (calling it 'Bob' would do just as well).
    If there is a phenomenon out there in the world that is thinking, we need to know what it consists in in order to understand what kinds of systems display that phenomenon: e.g., humans, octopuses, spiders, ANNs?

    dcm,

    @deshipu @troos @UlrikeHahn It's not boring at all: lots of things go on in brains and ANNs, some explanatorily relevant to certain questions and not to others. When modelling, certain things are simplified, others are left out: such choices are dependent on what phenomena the model is supposed to be modelling to start with. Without a clear picture of what a model is a model for, not much is achieved (similar issues were raised about the failures of the EU HBP).

    dcm,

    @troos @deshipu @UlrikeHahn I'm not a fan of definitions (i.e. sets of necessary and sufficient conditions): I think it's hopeless to capture the complexity of natural phenomena with them.
    But we need good characterisations of phenomena, otherwise it's unclear what it is we are investigating. They are preconditions to knowledge.
    Moreover, often lots of knowledge gathered previously informs these characterisations: they are constantly tweaked in light of empirical and theoretical considerations.

    dcm,

    @troos @deshipu @UlrikeHahn A characterisation is something that is not exceptionless like definitions are, they capture core features but can have fuzzy boundaries and admit of gray areas (definitions admit of none of that). Many biological notions are identified by characterisations, not definitions: organism, cognition, intelligence, even, under some proposals, life.
    They are more like prototypes, in the cogsci sense of the term.

    dcm,

    @troos @deshipu @UlrikeHahn We may disagree on what counts as constituting knowledge ;)
    Coming up with scientifically useful characterisations of phenomena is extremely difficult and must be sensitive to lots of theoretical and empirical considerations. It's basically to establish ways to partition the world that capture patterns that call for unified explanations, methods, that can be generalised, etc. Making progress in this, I think, is to improve our knowledge of the world.

    MolemanPeter, to random

    ... no amount of investigation of another person’ s neural processes by means of fMRI will allow us to inspect his reasoning or what he is thinking.

    Bennett, M. R.; Hacker, P. M. S.. Philosophical Foundations of Neuroscience (p. 107). Wiley. Kindle Edition.

    dcm,

    @NicoleCRust @debivort @MolemanPeter @knutson_brain @beneuroscience @kordinglab @dsmith @tdverstynen Thanks for this nice summary of your thoughts, Nicole!
    I also think that the top-down causation thing in this whole discussion is a red-herring. There's nothing mysterious about contextual influences: the activity of one brain area is modulated and differentially processed in light of the activity of other brain areas, and the latter is partly determined by the context the organism is in.

  • All
  • Subscribed
  • Moderated
  • Favorites
  • •
  • JUstTest
  • mdbf
  • ngwrru68w68
  • modclub
  • magazineikmin
  • thenastyranch
  • rosin
  • khanakhh
  • InstantRegret
  • Youngstown
  • slotface
  • Durango
  • kavyap
  • DreamBathrooms
  • megavids
  • GTA5RPClips
  • tacticalgear
  • normalnudes
  • tester
  • osvaldo12
  • everett
  • cubers
  • ethstaker
  • anitta
  • provamag3
  • Leos
  • cisconetworking
  • lostlight
  • All magazines
    •