Cerstin Mahlow @CerstinMahlow@qoto.org

Bin gespannt, wann/ob auf "jede Suchmaschine wird mit ChatGPT etc. verknüpft" die große Ernüchterung folgt, wenn irgendwann mal auffällt, dass Sprachmodell prinzipbedingt nicht nach wahr/falsch unterscheiden können, sondern nur Muster liefern, die plausibel klingen.

(Vgl. auch: autonomes Fahren ... da sind wir jetzt in der Ernüchterungsphase, wenn ich's richtig sehe ...)

Ein Murrsack mit Zugband für allerhand Nörgeleien. Ein mehrstimmiger Chaor, der nur Lieder mit mehreren Katastrophen singt. USBäh-Stecker zum Aufladen nach besonders miesen Tagen. App zum Erkennen von Sinnvogelstimmen (inkl. Übersetzung). Ein Geh-Dreieck zum Vermessen des schnellsten Auswegs. Knethaken für zu lang gegangene Termine.

#SechsUnmöglicheDinge

This is a big challenge with prompt leak attacks generally: the model just guesses what word should come next, so once it starts spitting out pieces of its own prompt it's perfectly capable of inventing new prompt segments out of thin air

And anything it invents will look convincing, because the whole point of large language models is to generate stuff that looks convincing!

Translation: our half-baked idea to charge for API access broke our internal API access, and we don't know how or why. Captain Birdchan gave us a week to figure it out. Also, here's this other shiny thing that the 3 remaining marketing professionals *think* will be enough to distract the mainstream press from the total clusterfuck that happened today.

@mmasnick

chatbots in search engines are obviously a bad idea because they will confidently lie to you.

but it’s also going to be a massacre for media companies and blogs who are currently publishing good information but rely on traffic from google.

#Stadtwanderung

Auf den Spuren der Ostschweizer Wirtschafts- und Politprominenz in Bern: Stadtführung für die Ostschweizer Regierungskonferenz. Hier das umgebaute Tobler-Areal, wo Jean Tobler, Zuckerbäcker aus Lutzensteig, Appenzell Ausserrhoden, 1908 die weltberühmte Toblerone erfand. Mehr hier:

https://www.stadtwanderer.net/?p=19894

RECEPTIO's version of #receptiogate:

"In December 2022, we were the victim of a vicious defamation campaign via the web, which shut us down for a month. After a legal investigation that established that all the accusations against us were unfounded and the result of the ravings of a sick mind (to which haters gave credence), in February 2023 the centre resumed its cultural activities"
[https://www.receptio.eu/story]

The lack of a full stop after the last sentence suggests that the story isn't quite finished yet...

#whataweekhuh

I’ve noticed that my expectations as a professor are sometimes quite different from those of my postdocs. This is not about research experience, etc., and not about culture, but rather basic work matters; things that go for me without saying. At first, I thought it was just me, but apparently colleagues have encountered similar issues as well.

Therefore I’m starting to wonder whether this is perhaps due to generational differences between #GenX (professors) and #Millenials (postdocs).

So, question for other #GenX professors: Do you have similar experiences? And if yes, any good approaches for dealing with it? #university #research

Im neuen SAGW-Newsletter gelesen: «Universität Zürich schafft unbefristete Stellen im Mittelbau», sog. Lecturer-Positionen in Lehre und Forschung. Die Stellen sind für Postdocs vorgesehen, Verbesserungen gibt es aber auch für Doktorierende und Assistierende. Bleibt nur die Frage der Finanzierung:
https://www.sagw.ch/sagw/aktuell/news/details/news/next-generation-uzh

How hand waving LLMs have been elevated to being „The AI“ by the media is astonishing.

Just because of a being very fluent with perfect syntax. Humans are too easily blinded by people that have high speaking skills.

at this phase of life when I hear things that challenge the very premises I stand upon, I have come to enjoy the free-fall.

https://whatashrinkthinks.substack.com/p/those-that-hear?sd=pf

I wrote a longer-form piece on post.news about the concept of operationalization in AI and the way it disenfranchises communities from making decisions that the systems that govern their lives.

If you want to read it there: https://post.news/article/2LR4k0jt685pzd6uiGeHJlnBwo2

I'll also serialize it, below.

Bet I'm not the first person this week to speculate about Bard in education - but I did write this a while back when Google first talked of LaMDA for learning, and I imagine it won't be long now until it plugs Bard into Workspace and Classroom, and then all across third party edtech products via APIs https://www.fastcompany.com/90641049/google-education-classroom-ai

Hey, remember the first #enshittification of #Google, when every single Google app got #GPlus awkwardly crammed into it because individual googlers’ bonuses were calculated based on how and whether they got people to use G+?

https://www.cnbc.com/2023/02/06/google-ceo-tells-employees-it-needs-all-hands-on-deck-to-test-bard.html

It looks to me like the Bing implementation is the same pattern that a bunch of other people are trying at the moment: you run a traditional search, then dump the search result summaries into a prompt and get the language model to generate text using those summaries

I tried something similar against my blog a few weeks ago! https://simonwillison.net/2023/Jan/13/semantic-search-answers/

My book about keyboards, Shift Happens, is now on Kickstarter!

There is all sorts of new information about it, and a video, and what I think are very cool tiers and rewards on top of a very cool book.

Please back and spread the word!

https://www.kickstarter.com/projects/mwichary/shift-happens?ref=cr61wq