BLUE
Profile banner
TU
Ted Underwood 🦋
@tedunderwood.me
Uses machine learning to study literary imagination, and vice-versa. Likely to share news about AI & computational social science / ciência social computacional / 計算社会科学. Information Sciences and English, UIUC. Author of Distant Horizons (Chicago, 2019).
4.9k followers3k following8.7k posts
TUtedunderwood.me

A very short blog post reflecting on an experience I've had several times lately: where students show up with a fully-formed solution to a problem I thought was still too difficult for them. #machinelearning#AI

Will AI make us overconfident?
Will AI make us overconfident?

Like the internet or a magical sidekick, chatbots are reorganizing knowledge to be more interactive and more accessible.

13

LTnolauren.bsky.social

Really helpful framing! Thank you!

0
NLlagernina.bsky.social

Thanks! This is a really useful take that I will take with me in upcoming encounters with AI-assisted student work.

1
CTcarlbergstrom.com

“As in any fairytale, accepting magical assistance comes with risks.” Great line. And as a genAI skeptic I appreciate thoughtful “pro” arguments such as yours.

0
Zzhuuuuu.bsky.social

AI is killing itself, so yes it will probably make us overconfident.

0
JCjonathancheng.bsky.social

Reminds me of 5 years ago when suddenly every other undergrad knew how to write attn networks —or at least the transformers statements. No LLMs then, but the internet was already there to give students the initial courage to ask daunting questions like, “what is a subword tokenizer?”

1
BBbrettbobley.bsky.social

Thanks Ted -- great post.

0
Mmaniagnosis.bsky.social

Javascript is actually a nice language. It just has 3 problems. 1. MS and Netscape used it as the browser wars' battleground. 2. Many decided it needed to be superficially more like other langs. 3. It provided an easy on ramp to web programming, resulting in a legacy of really bad code.

1
EDeverydayhuman.bsky.social

I wonder what would happen if you simulated a social network. With AI that was trained to teach and engage people if they would learn automatically with the ai itself learning at the same time. Verses traditional social media engagement. 1. the students social network blind to ai. 2Control 🤷‍♀️

1
MNmm-jj-nn.bsky.social

Got around to reading this now and it's very good! "[A]n extension of a project that the internet started thirty years ago: a project to reorganize knowledge interactively" is a nice way of putting something I've been struggling to articulate.

1
WSwenyishang.bsky.social

I think the "it’s easy with ChatGPT" feeling is especially valuable in DH. I believe many of us have encountered humanities students who feel daunted by coding. A bit of confidence or overconfidence can be crucial in helping with "emotional framing" (quoting @dmimno.bsky.social's words at DH2024)

0
Profile banner
TU
Ted Underwood 🦋
@tedunderwood.me
Uses machine learning to study literary imagination, and vice-versa. Likely to share news about AI & computational social science / ciência social computacional / 計算社会科学. Information Sciences and English, UIUC. Author of Distant Horizons (Chicago, 2019).
4.9k followers3k following8.7k posts