KidsWriteJokes Their were 2 fish in a bowl and 1 fish looked at the other and said fish don’t talk
MisterCinecal Every British movie that isn't about WW2 is about an auld lad getting his groove back and they're all called something like Major Eccleston's Barmy Bank Holiday pic.twitter.com/u6jGeEXKra
MisterABK This may have been an April Fools, but the fact that it was so believable speaks volumes about our society. twitter.com/MisterABK/stat…
rob_heighton I feel like there's a way to kill two birds with one stone here twitter.com/delosrusso/sta… pic.twitter.com/IqbOPYZeVb
Leute, die ihr Mastodon nur mit einer handverlesenen Untermenge ihres Twitters befüllen. Weiß man da schon Genaueres?
random_walker AI isn't monolithic, AI risks aren't monolithic, and interventions can't be monolithic. Defenders of the letter who ask "if not this, then what?" are ignoring like 30 different tailored responses already in play. Just one example, from consumer protection: ftc.gov/business-guida…
Eilenbrat Wann evolutionshistorisch akkurates Jurassic-Park-Remake mit Federn und Schmollmund? twitter.com/guardiannews/s…
asayeed ... the argument on the idea that the model of the world derived entirely from text could *never ever* be a sufficient model to represent the kind of autonomous-intelligence "danger" that the longtermists fear is, like I said, just never going to be that convincing.
asayeed Outside of linguistics and NLP, this argument about models vs. the real thing is never going to be that convincing.
There's lots of reasons to dislike longtermism and the "Future of Life" sketchiness and to reject that petition, but constantly hanging ...
asayeed It's not even a sufficient argument that you couldn't blur the map/territory distinction with text alone ("learning from form"). It has all the flaws of the Searlean "Chinese Room" argument.
asayeed Nothing in the "octopus" argument and similar claims is a kind of constructive proof of a strong "map/territory" distinction, i.e., one in which the map is never a sufficiently acceptable approximation of the territory until it literally is the territory.
asayeed The "strong AI" claim is that there is a point at which a *working model* of cognition is sufficiently representative of the "real thing" that you cannot distinguish it behaviourally from the "real thing". And that level is not a full replication of the biological form.
KulturGeschwtz Auf der Suche nach einer neuen edgy Persönlichkeit dachte ich mir, vielleicht probiere ich es mal mit „konservativ“. Zur Einstimmung in mein neues Ich lese ich folgende Titel: pic.twitter.com/TncVZYg8UB
texttheater Everything Everywhere All at Once – Kung Fu im Finanzamt #DeutscheZusatztitel
DK_Pitchbot Wer in seine Tweets anglophone Lexeme interlaziert, exkludiert systematisch Personen, die derart assumtionsreplenierte rhetorische Ludibrien und Kodizes nicht qua privilegierter Provenienz internalisiert haben. Es ist ableistischer, klassistischer "Humor". Meine Eltern haben ein
tacosconference Did you know TaCoS has two sister conferences?
StuTS is a general linguistics student conference open to all subdisciplines. It takes place May 25th - 29th in Frankfurt am Main. Registration and CfP are open until April 24th!
(1/2) pic.twitter.com/HpGjc6wK8i
derwahremawa Beispiel: wenn ein Schiff hupt, ist es gleich 4x lustiger, wenn man sich „LE HÖNK“ dabei denkt
derwahremawa Kleine Humorkunde:
– jedes Geräusch wird doppelt so lustig, wenn man es sich als Klangwort vorstellt
– jeder Umlaut macht ein Wort wiederum doppelt so lustig
– dito ein alberner französischer Artikel davor
– dito Großbuchstaben
hordcare @texttheater kucke stechend bewundernd
Namensidee … titanic-magazin.de/fachmann/2023/…
pandemic_de könnte eine Liste machen von Samen die ganz besonders nicht in Schottergärten geworfen werden sollten, weil sie viel zu gut anwachsen würden twitter.com/diepartei_dadi…