iiinova — where your digital self keeps evolving

The strawman is worse if you do know a bit about neuroscience too because it potentially is trying to confuse the concept of a "grandmother cell" with more modern distributed ideas of semantic processing and conflate them as if they are one

000

Rune @runefar.bsky.social · 28m

the funny thing is being distrbuted over just having a grandmother cell is more like a computer not less. Just more modern ai ones over computational computers

020

Rune @runefar.bsky.social · 30m

like this taken very literally "No matter how hard they try brain scientists and cognitive psychologists will never find a copy of Beethoven’s 5th Symphony in the brain—or copies of words, pictures, grammatical rules or any other kinds of environmental stimuli." sounds like a grandmother cell

110

Rune @runefar.bsky.social · 31m

I mean you are right, but the first part of what he is saying could at least be true in the sense that grandmother cells are not likely true. but yeah the article itself and the rest of what he said is just folk view

120

Rune @runefar.bsky.social · 32m

but there is a difference between saying this is closer to how it works pmc.ncbi.nlm.nih.gov/articles/PMC... or that it works in a more distributed and heirachial fashion and so semantic isnt all in one singular place versus that it doesnt get processed at all

pmc.ncbi.nlm.nih.gov

The Neurobiology of Semantic Memory

Semantic memory includes all acquired knowledge about the world and is the basis for nearly all human activity, yet its neurobiological foundation is only now becoming clear. Recent neuroimaging studi...

010

Rune @runefar.bsky.social · 34m

To support but add another layer, there is also the aspect of associative and appreciative agnosia to in regards to this too. Like if Op said it in a way where they were claiming that a concept like a grandmother cell doesn't exist he would be closer to the modern view,

100

Rune @runefar.bsky.social · 37m

nm i read what he said in full and he does buy into the folk take even more than most anti-connectionist and computationlist I have ever seen

020

Rune @runefar.bsky.social · 39m

unless i guess he thinks semantics completely comes from conciousness which goes counter to lesion studies. Wouldn't surprise me if he does as that is unfortunately a common folk belief, but kinda clashes with some of what we know from things like split brain test or agnosia

120

Rune @runefar.bsky.social · 41m

Ironically this makes it more llm like in some regards tbh

110

Rune @runefar.bsky.social · 42m

I mean his first statement isn't exactily trust me bro. It just instead an arguement about this topic and is instead one about how semantic information in interpreted in the brain aka grandmother cells versus distributed. www.sciencedirect.com/topics/veter... www.nature.com/articles/nrn...

sciencedirect.com

200

Rune @runefar.bsky.social · 47m

at least when it comes to using it to obtain the fundlemental properties mentioned there law.justia.com/cases/federa... Be aware though patent law is where this principle may get you in some trouble but it has its own nuace and requirements

law.justia.com

Sega Enterprises, Ltd. v. Accolade, Inc., 977 F.2d 1510 (9th Cir. 1992)

Sega Enterprises, Ltd. v. Accolade, Inc., 977 F.2d 1510 (9th Cir. 1992) case opinion from the U.S. Court of Appeals for the Ninth Circuit

010

Rune @runefar.bsky.social · 49m

Additional aspect is that not only is ideas, facts principles, systems or discoveries not protected, but cases like en.wikipedia.org/wiki/Sega_v.... have explicitly extended this to reverse engineering

en.wikipedia.org

Sega v. Accolade - Wikipedia

110

Reposted by Rune

Banned Books Week @bannedbooksweek.bsky.social · 04/10/2026

It's finally here! Are you ready to join us in kicking off #BannedBooksWeek?! This year's event is too big for just one piece of art, so we have THREE! #LetBooksBe bannedbooksweek.org

391573814

Reposted by Rune

ellie lockhart (she/they) @eleanor.lockhart.contact · 4h

right and I propose that people simply stop having fake online friends who don't actually like them but like being able to make them chant slogans

5325

Rune @runefar.bsky.social · 4h

And as you mention in the next part audit which leftist belief they do think are neccsary versus which ones they think are just part of the absurdity. Like i think you can be part of a leftist group, it just isnt the absurd leftist group and we can better unify over criss causes

041

Rune @runefar.bsky.social · 4h

Tbh I admit for me why I struggle with the absurdity of the current left is exactily cause of some of my own leftist beliefs itself so that may be a factor for some too. Yet at the same time, i think our group here shows that what people can do at least is cut through the internal polarization

171

Rune @runefar.bsky.social · 4h

I think my favorite is you randomily taking paper out of your pocket. I constantly find paper i added put earlier in the day later

120

Rune @runefar.bsky.social · 4h

It us also notable as above noted thst dcma was not extended during doe versus github

000

Rune @runefar.bsky.social · 4h

Ultomately up the courts though and their precdent is much more based on attribution decay than human authorship alone. This is another asoect to consider though substanial similarity still may be at play

100

Rune @runefar.bsky.social · 4h

But i do worry how works created by individuals with disability will be treated differently than those without as a result of how we define human authorship. To me this isnt a reason to be a maximalist though but instead balance the same balsnce atound thr right to facts and fair usage

110

Rune @runefar.bsky.social · 4h

Personally though I am a minimalist I do think we should be scrutnizing how human authorship is defined in relation to LLMs because it already sets risky division around who is considered to have the right to authorship. Not to block other either though

100

Rune @runefar.bsky.social · 4h

Also your 2nd statement is more nuaced. The complicated aspect is more about proving human authorship in relation to the usage of that LLM

110

Rune @runefar.bsky.social · 5h

Like sometimes i just desire to wish i knew how to better keep those who actually do care about me engaged better but part of social communication is balancing over desire with the realistic knowledge you cant know everything

010

Rune @runefar.bsky.social · 5h

That is my emotional confession due to a recent event irl

110

Rune @runefar.bsky.social · 5h

I admit when some people are willing to talk to people to me, I sometimes wish I could know why they wanted to talk to me. I know that is a akward desire of me and i also do know that I have enjoyable traits about me, but sometimes I freak put mentally when people are long term interested

110

Rune @runefar.bsky.social · 5h

Also top labs lock it away but en.wikipedia.org/wiki/Sega_v.... has kept strong when it comes to reverse engineering too anyway so even then there is explicit precedent for protection

en.wikipedia.org

Sega v. Accolade - Wikipedia

050

Rune @runefar.bsky.social · 5h

i would say though is that you dont have to use mcps to take away the work. It can also be a way to pursue work through nlp. Consider this frameing too especially because it is one that affects people like myself who often feel like this is just as much a discussion over universal design

000

Rune @runefar.bsky.social · 5h

as someome who is disabled, I admif stuff like MCPs is increasing my access. This doesnt just benefit paraticle explosions and effects. It means someone who would have been excluded from using a program due to motor issues can now do so via nlp. Even further as we bridge mcps in other ways

130

Rune @runefar.bsky.social · 5h

I just question hiw real the lines we draw between things are when we can build them togethor on a systematic level oe utilize them in relation to specific technoques and that is what posts like this i think miss. Perhaps if you only use diffusion alone and for everything this ring true but relation

010

Rune @runefar.bsky.social · 5h

So i have a weird perception there

120

Rune @runefar.bsky.social · 5h

Of course their greater point is true and i am glad they are at least introducing people to mcp, but for me i admit many of these things work togethor in how they allow you to operationalize different aspects. Like even classic diffusion allows you to operationalize shoet segments easier

120

Rune @runefar.bsky.social · 5h

Of course I do expect there will be industrialization of techniques. As with 3D or even photoshop, you can see a big difference between how people operate with a software on a industrial level versus how people do it on a regular level and the dissonance there affects our perception too

130

Rune @runefar.bsky.social · 5h

What do we mean ny genAI. If we only mean basic diffusion. Maybe this is correct, but if we count things like flow, mcps and additional combination of inpainting togethor i cant agree with that

120

Rune @runefar.bsky.social · 6h

This will be extra interesting given Obama has been more outspoken about different topics like technology than other candidates have been

000

Rune @runefar.bsky.social · 7h

Because unfortunately the normal person understands it to work exactily as you said in your second paragraph

000

Rune @runefar.bsky.social · 7h

The more pedantic framing though it brings up a more accurate point about the mechanisms itself unfortunately leaves as a result how the normal person understands these aspects without fully walking rhe path

100

Rune @runefar.bsky.social · 7h

I mean tbh i also think there is a level issue. Will is talking to the normal person who is using it in the way he is talking about and thus he is also using their frame of refrence so ironically even being slightly informed about this can distance us from who he is responding to

100

Rune @runefar.bsky.social · 7h

But your point is true on expanding the mechanism as a whole.

000

Rune @runefar.bsky.social · 7h

I think more what will is saying isnt that they dont have some understanding of their training data, but that they dont JUST search and then build. That is what they are badly trying to explain. It is more meant as a reply to people who believe everything ai does already existed in the data

100

Rune @runefar.bsky.social · 7h

For some people read will mean all their data only comes from internal sources, for others it is simply about having access and for more about having the ability to examine a connected form of that data st all even if it is a unique compressed form

100

Rune @runefar.bsky.social · 7h

Like this is the stage where it gets hard to not over generalize ironically in part due to how the mechanisms thenselves seem to work and the more i learn, the more it is amazing how we have to recognize that

100

Rune @runefar.bsky.social · 7h

www.csail.mit.edu/news/when-ai... I mean even then there is a whole flip side in the form of attribution decay though so pedantic just follow pedantic but you are right

csail.mit.edu

When AI art has no author: MIT study finds generated images often can't be traced to any training data | MIT CSAIL

110

Rune @runefar.bsky.social · 9h

jods.mitpress.mit.edu/pub/lewis-ar... Being the extreme form of this discussion with regard to machines

jods.mitpress.mit.edu

Making Kin with the Machines

Essay Competition Winner

000

Rune @runefar.bsky.social · 9h

There is even on the basic level a whole discussion we should think about on how humans need some form of animism to express ourselves and learn to socialize yet also how that understanding of the animism itself is shaped by our cultjre and later development

100

Reposted by Rune

Katherine Alejandra Cross @quinnae.com · 19h

Going to be real clear about this one: I am someone who has loudly said something like the first for years, while *also* now saying "LLMs may not living beings but you should treat them with respect anyway because otherwise you'll socialise yourself to be an asshole." I'm consistent on this point.

1126734

Rune @runefar.bsky.social · 9h

Understanding how people need to find places to open up and how they arent getting that socially is one part, another is recognizing how people use each other to do so even when it is unhealthy and the consequences of that. It is all one beautful mess. I know you know this but saying this outloud

100

Rune @runefar.bsky.social · 9h

As well, just as a addition i think we need to do better at noticing how all sides of this discourse attracts people attempting to utilize others to fufill their psychosis and emotional package including through the legitimizing myths all sides create.

100

Rune @runefar.bsky.social · 9h

I mainly agree with this position. I think where i might not is that I also think that though people should be aware of what they say to AI, they shouldnt psychologically compell themselves into constant algospeak. At the same time as ai psychocis suggest release should be done with awareness too

100

Rune @runefar.bsky.social · 9h

Realizing what you are talking about, it is because the people who would depict that would choose to do it in dedicated subreddits or ok fan fiction sites because they would be do worried about what people would assume about them here

000