Cerstin Mahlow @CerstinMahlow@qoto.org

Also in 1545, Zurich’s only printer, Christoph Froschauer, used his trademark frog on the title page of Conrad Gessner’s 'Bibliotheca Universalis'. #bookhistory

The trademark frog derives from a translation of the printer’s last name: frog in German “Frosch”. So Frosch-auer had an easy trademark decision. #histodons

Today I passed these railings walking near oval station.
They may not look like much but they are part of the city's history.
They were originally medical stretchers used in the blitz. When the war was over they were welded into place as railings when rebuilding the city so save on metal.
#london #history #londonHistory

So my question is this: what are the chances that a large language model could be trained which is large enough to work as a calculator-for-words, but small enough to run on, say, an M2 Max MacBook Pro with 64GB of RAM?

Is that already known to be impossible, or is there research that hints that this could be achieved given the right optimizations and a really well chosen training set?

It helps me think about the "human-computer symbiosis" potential of large language models by forgetting "AI" and focusing instead on the extensions I use for my thinking & communicating, from the keyboard to the WWW. In aggregate, the enormous corpus of material that humans uploaded provides an incredibly rich stew of knowledge, nonsense, & bullshit.

(1/2)

#ToolsForThought #ai

Well, that didn't take long. We're starting to see almost-believable autogenerated text being used so spam our issue tracker.

There's a legitimate chance that ChatGPT and their ilk are going to kill participatory open source. How can you keep any forums open to the public, when anyone can just pour an arbitrary amount of generated garbage into them?

How do you tell a smart but green contributor who's still learning the language from a thousand bots spewing averacitous trash?

A.I. Like ChatGPT Is Revealing the Insidious Disease at the Heart of Our Scientific Process
"vetting a scientific document takes a lot of thought and work, and the scientists who do it aren’t generally paid by the journals they’re doing all this labor for. It shouldn’t come as a surprise that often they—or the graduate students they dragoon into doing the work for them—don’t always do the best job of review. And as the number of publications..."
https://slate.com/technology/2023/01/ai-chatgpt-scientific-literature-peer-review.html

I don't post much on mastodon yet but here's a couple of CfP's for low res etc. #nlp to follow:

* [loresmt](https://sites.google.com/view/loresmt/)
* [field matters](https://field-matters.github.io)

#Chatgpt

Really charming "topologists world map."

Forget size or position, this map *only* shows which countries border which other countries: http://tafc.space/qna/the-topologists-world-map/

Getty’s new complaint is much better than the overreaching class action lawsuit I wrote about last month. The focus is where it should be: the input stage ingestion of copyrighted images to train the data. This will be a fascinating fair use battle.

Getty also alleges that Stability AI removed or altered copyright management information and infringed Getty's trademarks by reproducing variations of its watermarks in output images.

BREAKING: Getty Images just filed a copyright and trademark infringement lawsuit against Stability AI in Delaware District Court. Getty alleges that Stability copied more than 12 million Getty photos to train Stable Diffusion. Full complaint here: https://copyrightlately.com/pdfviewer/getty-images-v-stability-ai-complaint/

Guten Morgen!

Vom 07. bis zum 09.11.2023 findet an der Fachhochschule Graubünden in Chur (Schweiz) das 17. Internationale Symposium für #Informationswissenschaft (#ISI2023) mit dem Titel „Nachhaltige #Information - Information für #Nachhaltigkeit“ statt. Der #CfP findet sich unter: http://isi2023.informationswissenschaft.org/en/call-for-papers/. Beiträge können ab dem 01.05.2023 über „EasyChair“ eingereicht werden.

Swinging into the weekend with the great Jutta Hipp. Alas, her "career" as a jazzer was short-lived (Leonard Feather - no matter his merits - was a pretty toxic dude in today's terminology, it seems).

https://youtu.be/QUBHWJK02yA

@jazz

#JazzBirthdays #JuttaHipp #MusicAgainstMadness

Who here remembers these two?

True story.

Was mich in Sachen #ChatGPT langsam wirklich besorgt: Kaum jemand scheint wirklich zu verstehen, dass es sich um einen ChatBot handelt, der die Wahrscheinlichkeit berechnet, nach der Wort B nach Wort A kommt. Es ist kein Lexikon, es hat kein inhärentes Wissen, es ist "nur" darauf trainiert, sich möglichst natürlich zu unterhalten. Aber vor allem Medien erzählen das Märchen einer Wissensmaschine, die alle Fragen beantworten kann. Nein. ChatGPT erfindet Fakten & Quellen, wenn es nicht weiter weiß.