Prof Dr Computer Science—Artificial Intelligence, Software Engineering, AI ethics—Head of #CS Dept, Co-Director https://xitter.cf/t.co/B7kXgCwcuG. Read me on Linked+In.
Berlin, Germany
Joined May 2010
- Tweets15.6K
- Following3.1K
- Followers7.8K
- Likes32.6K
Oh... well... I told you so. Sooner than I expected, but no surprise, as I tweeted *checking my timeline* yesterday!
(Too late, @jack. How does it feel to witness the destruction of ... Twitter?)
xitter.cf/dmonett/status/1604062…
"The more adept LLMs become at mimicking human language, the more vulnerable we become to #anthropomorphism, to seeing the systems in which they are embedded as more human-like than they really are [which] is amplified by the natural tendency to use philosophically loaded terms"
My new paper is a critique of the way we talk about large language models.
arxiv.org/abs/2212.03551
I argue that LLM-based systems are "simultaneously so very different from humans in their construction, yet ... so human-like in their behaviour, that
Dagmar Monett retweeted
Finally got around to reading this one, and yeah, the NYT continues to be trash at covering AI:
"the existence of a highly capable linguistic superbrain"
"could make it possible to create personalized therapy bots"
#AIhype
>>
nytimes.com/2022/12/05/techn…
Oral exams would be a good choice in that case. Or writing part of the essay in a 1-3 h in-person test.
If mastering or good command of a topic is a goal, then education must rethink its assessment tools (and probably also teaching).
Worth noting: when I refer to oral exams or any other kind of assessment tool, I'm not referring to using #AI or any other tech to automate them (it might be confusing at first read; some comments reveal that).
"The tyrant individual dictates his morals to the rest of society and makes use of new technologies and communication strategies to determine behavior patterns for the rest of the people."
@Eric_Sadin, any plan editions in English of your books? More people need to read them!
"La era del individuo tirano" (@cajanegraedit) parmi "les dix livres de l'année qui diagnostiquent et analysent les défis auxquels nous sommes confrontés en ce nouveau siècle", par @LaVanguardia
lavanguardia.com/cultura/cul…
"ChatGPT is more about bullshitting than creativity, which serves as a neat metaphor for what has happened with our technology sector and how it 'feels like a giant organ for bullshittery―for upscaling human access to speech and for amplifying lies.'"
vice.com/en/article/bvmk9m/e…
It reminds me every single day of the dictatorship I lived in for 30+ years, in a country surrounded by water. Melon Usk shows all characteristics of a laughable dictator, even the capacity for attracting bootlickers en masse.
This quoted post is unavailable.
- a proliferation of individual, little tyrants that would give whatever possible to be in the position of the dictator
- banning of certain words
- rebranding of other words in order to accommodate the dictator's discourse
"The veracity of news or of a scientific article is not measured by the coherence of the text, but by the credibility of the sources, by the strength of the evidence and the solidity of the arguments. This is where we find a big problem w/ ChatGPT."
(🇪🇦)
agenciasinc.es/Opinion/Las-f…
Dagmar Monett retweeted
I am THRILLED to announce that the #ConceptArtAssociation has released a #gofundmecampaign to fund multiple actions, so we artists can stand up for our rights and be heard in the public and in DC! Art friends please donate, SHARE this link. Let's do this!
gofund.me/2df3dc07
Dagmar Monett retweeted
#Pressefreiheit darf nicht nach Belieben ein- und ausgeschaltet werden. Unten stehende Journalisten können auch uns ab heute nicht mehr folgen, kommentieren und kritisieren. Damit haben wir ein Problem @Twitter.
Dagmar Monett retweeted
ChatGPT creates new work for good programmers. What work? I'm glad you asked. Good programmers will have to fix the mess left behind by bad programmers that relied on ChatGPT. (Inspired by @giskard23)
Dagmar Monett retweeted
ALT a Venn diagram showing things that are actually slowing down development teams (constant reprioritizing, burnout, pointless meetings, technical debt, generating subtle devastating bugs) - and things that an AI coding assistant speeds up (generating subtle devastating bugs, googling boilerplate code). The intersection is generating subtle devastating bugs