@tante@tldr.nettime.org avatar

tante

@tante@tldr.nettime.org

Sociotechnologist, writer and speaker working on tech and its social impact. Communist. Feminist. Antifascist. Luddite. Email: tante@tante.cc | License CC BY-SA-4.0 tfr

"Ein-Mann-Gegenkultur" (SPIEGEL)

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

tante, to random German
@tante@tldr.nettime.org avatar

Also für jede Organisation, die geballtes Wissen über Digitales und Politik braucht, ist das hier die Chance. Sprecht Anne jetzt an, bevor es jemand anderes tut.

https://systemli.social/@anneroth/111370863409195004

tante, to random
@tante@tldr.nettime.org avatar

Really disappointing to see CreativeCommons argue that "generative AI" on the basis of the appropriation of cultural works often against their creators' wishes is somehow like remixing and in the "public interest".
https://creativecommons.org/2023/02/06/better-sharing-for-generative-ai/

tante, to random
@tante@tldr.nettime.org avatar

Everyone that Microsoft has fired in the last few rounds to be more lean must feel super good about them basically hiring everyone who ever stumbled into an OpenAI job, a massively overstaffed company that has never made any money and no real medium-term pathway towards profitability.

tante, to random
@tante@tldr.nettime.org avatar

Of course OpenAI has to stop doing business in the EU if they can't comply with the laws here. Why is that a scandal or debate?

tante, to random
@tante@tldr.nettime.org avatar

The fact how quickly OpenAI crumbles when the hype man gets fired shows just how little material innovation and substance there is in the field.

https://fosstodon.org/@chrisg/111436543757025986

tante, to random
@tante@tldr.nettime.org avatar

I can't believe that anyone would run an event accepting Jake Appelbaum in 2023.

I can believe that the CCC does.

https://tldr.nettime.org/@eaon/111658045309830027

tante, (edited ) to random German
@tante@tldr.nettime.org avatar

Die Behauptung, private Unternehmen würden (gerade in Strukturen, wo man die öffentliche Hand als Kunden hat, aber auch generell) so viel effizienter wirtschaften als Behörden oder sowas ist ja auch nicht tot zu bekommen.

Aber selbst wenn das stimmte (und das tut es nicht wie jeder, der mal mit einem beliebigen DAX Konzern zu tun hatte weiß): Es geht bei sowas wie Daseinsvorsorge nicht um maximale Effizienz, es geht um Ergebnis. Wenn ich super effizient bin, aber die Leute keine gute Gesundheitsversorgung haben, dann habe ich meine Aufgabe nicht begriffen. Natürlich sollte man Ressourcen zielgerichtet (schon wieder Ergebnis!) einsetzen, aber mit Erbsenzählerlogik kommt am Ende nur wieder Austerität raus. Unter der wir alle leiden.

tante, to random German
@tante@tldr.nettime.org avatar

Ich finde "Wir brauchen Zuwanderung, damit jemand die Jobs, die wegen mieser Arbeitsbedingungen und katastrophaler Bezahlung niemand machen will, macht" jetzt auch nicht unbedingt eine progressive Position.

Wir in Deutschland haben, insbesondere auch aufgrund des Beitrags, den wir an der Klimakatastrophe haben, eine Pflicht, Menschen Raum und eine Leben zu bieten. Egal ob sie nützliche Arbeit (welcher Art auch immer) machen wollen oder nicht.

tante, (edited ) to random
@tante@tldr.nettime.org avatar

I really like @Mer__edith's point here, very well put.

But I think there's something to add (that Meredith added in another segment as well): This true analysis comes from an output perspective. "Generative AI" produces "plausible" text and "passable though somewhat off" imagery so these systems are not that useful in serious contexts. 100%

But these systems are massively useful to undermine labor. To burden a decreasing amount of employees with the work of fixing the output of these systems (probably for bad pay because "you are just fixing it a bit"). And to repeat the last automation cycles of producing a lot more stuff for a lot less cost per unit which only works by reducing quality.

You know how bad fast fashion is quality-wise? That is what "generative AI" does to text, image, media. But that fact that fast fashion is bad doesn't stop it because the cost argument wins. And that is what "generative AI" is useful for.

"Useful" is a word that only makes sense when adding for whom and for what purposes and goals.

https://mastodon.social/@hagen/111328641537069348

tante, to random
@tante@tldr.nettime.org avatar

The biggest image data set used in "AI" contains examples of child sexual abuse material
(Title: Largest Dataset Powering AI Images Removed After Discovery of Child Sexual Abuse Material) https://www.404media.co/laion-datasets-removed-stanford-csam-child-abuse/

tante, to random
@tante@tldr.nettime.org avatar

What if the answer to "SEO and SPAM have poisoned the well for traditional search engines and their rankings" is not "so we'll insert a wasteful lying machine in there that makes up shit"?

tante, to random
@tante@tldr.nettime.org avatar

I have been talking about this for months. OpenAI fights regulation where it can and every photo of a high ranking official with Sam Altman should be seen as a threat.

https://time.com/6288245/openai-eu-lobbying-ai-act/

tante, to random
@tante@tldr.nettime.org avatar

"Do not think that thoughtful design is just for the elderly, or the sick, or the disabled. In the field of design, this is called “inclusive design” for a reason: It helps everyone."

https://www.fastcompany.com/90338379/i-wrote-the-book-on-user-friendly-design-what-i-see-today-horrifies-me

tante, to random
@tante@tldr.nettime.org avatar

With Microsoft hiring Sam Altman (after having just fired their "Responsible AI" team) they'll go even more all in into that kind of shit. The hype stays alive.

OpenAI seems pretty much dead now though. Like they never had a real business case that made sense economically given the cost of running all their statistics models, they'll never be able to actually monetize their products to the degree that it would require.

tante, to random
@tante@tldr.nettime.org avatar

"One of the most annoying thing about LLM chatbots is that they all talk like american used car salespersons or people who work in sales departments of software companies"
(Title: Jussi Pakkanen: AI silliness: getting to a no) https://nibblestew.blogspot.com/2023/12/ai-silliness-getting-to-no.html

tante, to random German
@tante@tldr.nettime.org avatar

Auch zauberhaft: Aleph Alpha, die große Hoffnung der deutschen Bundesregierung für "KI" Gedöns generiert (wie alle LLMs) menschenfeindliche Texte (bis hin zur Verharmlosung des Nationalsozialismus).

Antwort des Gründers: Ja das filtern unsere Anwender, wir sind nur Backend, wir wollen da nichts vorgeben. Das sind diese Europäischen Werte von denen alle reden. Schnell ne Standard LLM hinprökeln, Liebling der Regierung werden und die schwierige Arbeit an "Kollege nächste Schicht" verschieben. Je mehr man über Aleph Alpha hört, je mäßiger wird der Eindruck

https://www.zeit.de/digital/2023-09/aleph-alpha-luminous-jonas-andrulis-generative-ki-rassismus

tante, to random
@tante@tldr.nettime.org avatar

The most interesting thing about xAI's 'grok' is how it confirms what me and others have been saying for a while: "AI" construction is less about having new ideas or new tech, it's about just having access to data and resources. If you have money you can spin up a competitive LLM in months.

tante, to random
@tante@tldr.nettime.org avatar

Do I know someone (ideally in Berlin) who has set up and worked on Unity projects with > 10 people? I might have a small consulting gig.

tante, to random
@tante@tldr.nettime.org avatar

I wrote a bit about "generative AI art".

"So we basically put every artists' labor through the meat grinder in order to make a few idea guys feel special without ever having to learn a thing. And that is just so very sad."

(sorry it's a bit rambly, but COVID brain and stuff)

https://tante.cc/2023/11/10/thoughts-on-generative-ai-art/

tante, to random
@tante@tldr.nettime.org avatar

It's kinda neat to have many people realize how bad Gruber's takes are a lot of the time. He's just a tech industry mouthpiece with an Apple-Hardon

tante, to random
@tante@tldr.nettime.org avatar

Proton Mail parties like it's 2021 again and want to store public keys in a blockchain so you can ... verify that a key belongs to a person ... which ... is not really a big improvement over keyservers aside from "slow database magic".

https://fortune.com/crypto/2023/11/16/privacy-proton-mail-key-transparency-blockchain/

tante, to random
@tante@tldr.nettime.org avatar

"Meta is gonna use Threads to monetize Mastodon's userbase" is probably a bit naive. Threads now already has as many accounts as Mastodon and it's not even out in many regions (plus many Masto accounts are zombies). I don't think that putting significant effort towards trying to monetize Masto Users is worth it. Meta just likes the positive PR and the free content.

tante, to random
@tante@tldr.nettime.org avatar

Another study estimating that ChatGPT augmented search leads to probably about 60 times the emissions that current systems create.
(Title: Emissions from ChatGPT are much higher than from conventional search • Wim Vanderbauwhede) https://wimvanderbauwhede.github.io/articles/google-search-vs-chatgpt-emissions/

tante, to random
@tante@tldr.nettime.org avatar

Brian Merchant comes to the same conclusions I outlined in my 2023 Re:Publica talk:

"Generative AI may or may not be a flash in the pan, but it can be a wrecking ball to your job regardless, especially if your boss is looking for an excuse to cut costs or to appear innovative — and you should be ready"

(Original title: Understanding the real threat generative AI poses to our jobs)

https://www.bloodinthemachine.com/p/understanding-the-real-threat-generative

tante, to random
@tante@tldr.nettime.org avatar

“AI” as unregulated space

"AI" companies and "AI" influencers love to argue that the space is so new and unregulated, that we quickly need new regulation to give business a foundation to thrive on. This argument is very dishonest. There is regulation, just not the one they like. Data protection law, copyright law, anti discrimination law, transparency laws all those things do apply, even to systems that you slapped the magic "AI" sticker on.

https://tante.cc/2024/01/19/ai-as-unregulated-space/

tante, (edited )
@tante@tldr.nettime.org avatar

@jollyorc @isotopp On a pure technological level you are right, you can't bar OpenAI from scraping everything and turning it into a blurry JPG of the Web just like you couldn't block Napster/etc. fully.

But I think context and application matters.

There is a difference when a machine "consumes" a piece of media and when a human does it. Let's cast aside all debates about worthless content and shit: Participating in culture (by consumption and creation) is a human right, part of what defines human dignity. Humans and their needs are the end, the reason why we should do things. ChatGPT has no dignity and no right to participate in culture. That's what changes the equation: Locking human out of their own culture (by making it fully pay for play) is inhumane. Locking out OpenAI's crawler from gobbling up everything to sell back as Microsoft Clippy 2.0 is fine. Like, I don't really give a shit what OpenAI wants or needs.

Now on an economic level both things might actually feel the same for artists: Someone takes a thing you put work in without giving you any money to pay rent and eat. Does it matter who does it?

I think it does. (And I am not saying that everything is fine, artists just need to "get good" or whatever, we need to fix shit in that regard with or without "AI".) As an artist or just a writer, illustrator, musician, etc. (explicitly not saying content creator here) you want to reach people with your work. You create because you care. And you'd love to live off of it but if you can't reaching people does matter, forms connections (that might even pay some bills later on). You don't create so some dipshit at Microsoft gets their yearly bonus because they trained a model that can recreate your work in esthetic but stripped of all meaning and connection.

That's why I am tired of talking about these things as tech issues. They are not. They are social and political.

  • 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
    •