Subject: I cannot believe there even needs to be a ?discussion? about ?AI? slop,

but people asked to discuss it here. So let me fullquote two
very recent posts from the Fediverse, which sum it up better
in English than I can. Even just one ought to suffice.

I've been wanting to write something lengthy about "AI" for a while, discussing the arguments for and against and arriving at a reasonable position, but I find myself completely unable to get past the most
glaring point, which is that we are facing c?a?t?a?s

?t?r?o?p?h?i?c? ?e?c?
?o?s?p?h?e?r?e?

c?o?l?l?a?p?s?e? because of capit

alist overconsumption. Our species is

literally c?o?m?m?i?t?t?i?n?

g? ?s?u?i?c?i?d?e? by c?a
?p?i?t?a?l?i?s?m? right in fro
nt of my eyes

and the architects of the climate crisis have latched onto data
centres with the energy budgets of countries running sycophantic
Markov chains as their latest engine of short term profit extraction
and people are split right down the middle either happily burning the
world down to keep continvoucly morging or whatever the newest slop
models tell them to do next, or existing in a state of angry refusal
to believe that anything is wrong as everything crumbles into ash
around them other than that some people have the wrong skin colour or
gender presentation.

*deep breath*

The few of us who don't starve to death as our food sources collapse
may eventually die of a?s?p?h?y?x?i?a

?t?i?o?n? when the wrong Amazon finishes

burning down.

How do you get past that?

How does one even hope to connect with people who prioritise
dismantling the structures of capitalism before it kills us all l?o

?w?e?r?

than asking Claude for coding tips?

I guess we know what the Great Filter is now.

???https://social.treehouse.systems/@bodil/11609369
4360361788

Emphasis not mine. But hey, a physicist says it?s not ?AI
? which is
killing the planet, it?s the german move away from nuclear power an
d
towards renewables that is [sic!].

The second one is shorter but even more of a whammy:

? https://stephvee.ca/blog/artificial%20intelligence/generativ

e-ai-is-built-on-the-exploitation-of-the-global-south/

New blog post on generative AI and the trauma endured by data labelers
in the Global South. (CW: references to sexual abuse material and
unsafe working conditions)

???https://mastodon.social/@st3phvee/11609472900749
1217

(I?ve not read Steph?s blog post yet, but others that detai
l the same
situation in the past.)

I cannot understand how some even consider use of this worth considering
and try to distract those who resist with questionable-to-outright-false argumentations about how it?s all legal (trust me, bro!) and inevit
able?
and how we as a project cannot enforce contributors to not use it. Which
in fact we can, by the way, it?s called DFSG/SC. It?s calle
d plaing nice
and acting in good faith.

? oh, ?inevitable?? I have a third post for
you:

This is something that probably sounds obvious to some folks and
unhinged to others, but I think it?s worth saying anyway: ?

?workers who

don?t adopt AI will get left behind? is right-wing propag

anda. It?s

more than just a surface-level advertising message. It makes the
unstated assumption that workers are all in a state of conflict,
racing against one another in competition for acceptable employment.
This is the literal opposite of class consciousness! All workers
should be aiming for a world where we don?t have to fight each ot

her

to earn a chance at a decent life.

???https://social.treehouse.systems/@jnkrtech/11608
3770527548518

Yet, someone with ?community team? in the signature he chos
e to use for
the eMail wrote in a direct reply to me that calling out fashtech for
what it is (and the TESCREAL ideology behind it) is ?uncalled for
? and
that ?you [I] know it?. Sorry, no. I?ve informed my
self about this, and
I have seen more results of research about this than I ever wanted and
my opinion is that this is inacceptable, in so many ways, that we must
do something against it.

And we, as project, can. Among other very large FOSS projects, both
Gentoo and NetBSD have banned slop contributions. Just like that. And
so could Debian. Easily. And I fully expect my fellow DDs to adhere to
that and not knowingly or, worse, willingly, contribute slop, to main
or contrib, *or* to project infrastructure. (We ought to also remove
rsyslog, for that matter?)


I don?t expect to reply much (if I?m even allowed after thi
s) here.
There has so much been said about this on Fedi already that I think
pointing y?all there if you weren?t will be enough (and if
you were
and still consider slop maybe-acceptable, then I ask you whether you
really have been participating). And I don?t have the energy for lo
ng
discussions in English.

bye,
//mirabilos
--
?Having a smoking section in a restaurant is like having
a peeing section in a swimming pool.?
-- Edward Burr

--- PyGate Linux v1.5.11
* Origin: Dragon's Lair, PyGate NNTP<>Fido Gate (3:633/10)

Subject: Re: I cannot believe there even needs to be a ?discussion? about ?AI? slop,

Dixi quod?

I don?t expect to reply much (if I?m even allowed after th

is) here.

? but I think I have to make one addition. I don?t normally
read the
list (too much traffic, not enough spoons), but I wanted to see on
the webinterface whether the mail made it through, and I saw the mail
from peb.

The legalities aspect (when not used as distraction or waved away) is
a bit misrepresented.

Yes, the TDM exception gives an exception to copyright for the training
of models? to analyse things, for trends and the likes. Nowhere doe
s
this allow using the models to produce output. And please, do not use
the word ?generate?, they don?t generate (generativ
e art is something
entirely different and good), they regurgitate. LLMs are a sort of
lossy compressor/decompressor, with the decompression attempting a best *average* match to continue the ?prompt? (it?s real
ly just autocomplete
with sparks).
https://explainextended.com/2023/12/31/happy-new-year-15/ demonstrated
very nicely how they actually work, using an actually obtainable model
as example.

Incidentally, this is also why their output alone is not copyrightable
as a new work: it is produced by a deterministic machine, not a human,
and therefore does not pass threshold of originality? in two differ
ent
ways, one in the legal meaning of that term, the other in the ordinary
meaning of ?originality?: there?s nothing new there
, it merely

r e g u r g i t a t e s

from its inputs. (If the companies wouldn?t filter the possible pro
mpts,
it?d be easy to extract near-complete copies of individual ?
?training
data? by the millions, as studies have shown.)

? ? ? ? ? ? ? ?
?? ? ? ? ? ? ? ?
? ? ? ? ? ? ? ?
? ? ? ? ? ? ? ?
?? ? ? ? ? ? ?

HOWEVER, this does not mean that their output is free from copyright.
Rather, due to the above-mentioned properties (machine transformation
of copyrighted works), the sum of all outputs from such a model is a
derived work from all of its inputs (and for how much this is true for
each individual combination of input and output of course depends on
the prompt, PRNG seed and output in question). This does not, of course,
give you carte blanche to just use *any* of its output? not even sm
all
ones. Citing rules do exist, after all. Especially the academics should
know some?

So.

bye,
//mirabilos
--
"Using Lynx is like wearing a really good pair of shades: cuts out
the glare and harmful UV (ultra-vanity), and you feel so-o-o COOL."
-- Henry Nelson, March 1999

--- PyGate Linux v1.5.11
* Origin: Dragon's Lair, PyGate NNTP<>Fido Gate (3:633/10)