1. Bluesky Feeds /
  2. dame /
  3. Nuanced AI Commentary

Posts from people who have thoughtful and non-binary perspectives on AI that rise above the shallow culture wars. The feed shows newest AI-related posts from a hand-curated list of accounts. Grifters and reactionaries that fall into the anti-AI or pro-AI extremes are disqualified.

Feed on Bluesky

Feeds Stats

  • šŸ’™ Liked by 35 users
  • šŸ“… Updated 26 days ago
  • āš™ļø Provider graze.social

Nuanced AI Commentary Likes over time

Like count prediction
The feed Nuanced AI Commentary has not gained any likes in the last month.

Feed Preview for Nuanced AI Commentary

Ted Underwood
@tedunderwood.com
22 minutes ago
statistical learning that approaches the topic quantitatively. For inst, students should understand why the word ā€œlearningā€ is in ā€œmachine learning,ā€ and what theory of learning is being invoked. They should be able to write a par. about how ANNs are similar to and diff from earlier approaches.
1
0
2
Ted Underwood
@tedunderwood.com
22 minutes ago
I would envision this more as a stand-alone gen-ed requirement (right now) than as Purdue’s plan to integrate it in all courses and majors. You’re right that faculty aren’t ready for that yet. But students should be taking something in this space, whether it’s a history of AI or an intro to +
1
0
0
ae
@aelkus.bsky.social
29 minutes ago
the more depressing alternative is recognizing that politics is a team sport….and people’s social and professional affiliations have a huge weight on what team they choose when they are forced to
2
2
22
ae
@aelkus.bsky.social
30 minutes ago
and its probably the case that over an truly meaningful sample anyone using them as guides would be little better than a random algorithm
1
0
14
words
@words.bsky.social
33 minutes ago
I have been going through lists of old art ideas and poetry widgets and experiment sound pieces and claude is just like yeah we can do that even things I tried to get Claude to build a year ago have shockingly improved results
1
0
1
Ted Underwood
@tedunderwood.com
39 minutes ago
I agree that we don’t already know how to teach ā€œAI literacyā€ well. But I also think literacy is too modest a goal. College students should graduate knowing how ML works; it’s liberal knowledge analogous to RNA or trig, not search-engine skills or Excel. And we do know how to teach it in that way.
Opinion | Stop Pretending You Know How to Teach AI

www.chronicle.com

Opinion | Stop Pretending You Know How to Teach AI

Colleges are racing to make students ā€˜fluent.’ One problem: No one knows what that means.

2
3
18
words
@words.bsky.social
about 1 hour ago
if you actually listen to what people say you would be surprised how rarely humans communicate in syntactically correct sentences I would even say that its harder to get an LLM to make a syntactically incorrect sentence than a human
0
0
2
Benj Edwards
@benjedwards.com
about 2 hours ago
Yeah it looks like junk right now (though I can refine it over time), but if I'm just going to play it myself and have some fun, what does it matter? Making games this way reminds me so much of BASIC on the Apple II as a kid. An interactive hobby experience. The AI coding unlocks a lot of potential
0
0
2
Benj Edwards
@benjedwards.com
about 2 hours ago
I've wanted a Katamari Damacy game with Xmas theme for years. With Claude Code AI coding tool, I just made a playable demo, and it's fun! It's likely no one else will ever play this. We've entered the age of Personal Media: custom entertainment (video, music, games) created for one person to enjoy
Screenshot of Benj's Xmas-themed Katamari Damacy Christmas prototype game
3
0
17
Robin Berjon
@robin.berjon.com
about 5 hours ago
Aside from not getting the tech itself, this also fails to understand distribution channels, chokepoints, how data works, etc. Then you have people who think we're dealing with consumer products and don't see the infrastructure. Consumer products can just be replaced, only ever so harmful.
1
0
3
Robin Berjon
@robin.berjon.com
about 5 hours ago
A bit of everything but a lot of it comes down to not understanding tech in terms of power (and also business strategy). For instance, thinking that AI is general purpose tech, like a whole new internet. And so if we're good at AI, the rest doesn't matter because we then use trade.
1
2
2
Tim Kellogg
@timkellogg.me
about 6 hours ago
when did Claude Code get these elicitation features? super cool. A succinct questionnaire to clarify ambiguity
• Job file names

• Submit →

• List jobs?

* • Job types Should there be a list_jobs tool so the bot can see what's already scheduled? > 1. Yes Add a list_jobs tool to see all scheduled jobs 2. No Bot can just read the jobs/ directory with existing tools 3. Type something.
3
1
27
ponder
@ponder.ooo
about 10 hours ago
imagine how the bynars mustve felt watching riker get oneshotted by the holodeck ai gf
0
0
2
dan layus from augustana
@natalie.sh
about 10 hours ago
this sentiment reminds me of the oxide llm rfd
1
0
5
dan
@danabra.mov
about 10 hours ago
on one hand i use llm’s quite a bit, and sometimes to edit writing on the other hand if i see an llm-ism in something im reading i’m just gonna disengage and scan the rest or close the tab i’m not sure that’s internally consistent
18
2
156
dame
@dame.is
about 12 hours ago
dubious videos of animals doing cute thing have finally reached bluesky which is host to an audience made in a lab for being maximally triggered about such ai slop
1
1
34
post malone ergo propter malone
@proptermalone.bsky.social
about 12 hours ago
fwiw this video was apparently published with clear AI indicators before someone not-the-creator stripped them
1
0
4
henry
@hdevalence.bsky.social
about 12 hours ago
road damage scales as the fourth power of axle weight so a huge amount dynamic bus services aren’t practical without autonomy, which is now solved, that’s what’s new the upside is that it is better run because it not hamstrung by various political machinations and constraints
2
0
0
post malone ergo propter malone
@proptermalone.bsky.social
about 13 hours ago
I don’t think I’m giving up my ability to believe in objective reality when I say ā€œI don’t care whether this particular video is AIā€
2
0
3
post malone ergo propter malone
@proptermalone.bsky.social
about 13 hours ago
would your reaction have been different if there had been an unobtrusive but clear Gemini watermark on it?
1
0
1
post malone ergo propter malone
@proptermalone.bsky.social
about 13 hours ago
I don’t care if the crowsurfing is real. maybe the mechanism there is that I’m approaching it from the same ā€œyeah this is maybe fakeā€ stance I take and have taken with most video slop for years and resent the ask to litigate AI or no on every new ladle of video slop
8
3
58
dame
@dame.is
about 14 hours ago
and this is google search data for the term ā€œaiā€ over the past 20 years
a flat line that jumps aggressively at 2020
0
0
1