Yeah man, people just being like crazy condescending is part of why I gave up on mastodon after a relatively short trial run

I dunno why that guy thinks bullying is a way to encourage people to use mastodon

Or I dunno, maybe an AI agent blocked me based on sentiment analysis which is lame but whatever

Huh, guess I'm also on the anti-anti-AI blocklist.

I don't subscribe to any blocklists to get on there in an automated way, so must have been an artisanally crafted block based on something I said, which I can respect lol

But will it emit an ominous hum? That's essential

Significant differences are indicated with 🐦‍⬛

Yeah, tbh though AI is around in higher ed and people are talking about it, it's not at all like Bluesky in real life.

For the most part as a prof, it is just a minor inconvenience at my job when I have to clean up the occasional AI slop, and move around assessments to do less online.

Rip to a real one over at island greek, what a killer donair

Guess who booked a vacation to Spain in June 💀

I'll turn the car engine into a power source for bitcoin! I'll rent out our hotel room for datacenters to make NFTs!

The beach? Coolant for data centers!

We tried with moose first, but they were too powerful to control

Calvin and Hobbes is timeless, all time greatest newspaper strip

Yeah, and like even for us already in faculty positions, even if we do manage to get one in our department they will be basically pure research, so they won't really help with teaching or service, so more workload creep

I guess I should be glad they gave education anything at all yet...

As is typical, I thought I could solve my R data formatting problem with a regular expression, so I flailed around for like an hour, despaired at my ignorance, then solved the problem a different way much faster once I gave up.

The prompt said *wrong* answers

When I told an AI that the resource it linked me to was written by me, and thus I knew it did not contain the information it claimed it did, it said that was "a real 'I am the Senate' moment". It's not! It was an 'I am Pagliacci' moment, at best. Damn thing couldn't even get that right.

For example, Table A1 for Likert is pretty equal:
Strongly disagree:
Disagree: 15.32
Neutral:49.58
Agree: 82.62
Strongly Agree: 98.59

Table A3 suggests is ordinal, but not quite so equally spaced:

Never: 0.29
Rarely: 13.30
Sometimes: 57.06
Often: 72.49
Always: 99.08

I like Caspar et al. (2020) to get a sense of what kinds of anchors constitute equally weighted anchors. They take a bunch of qualitative descriptors, and re-scale them on a 100pt scale for comparability. Really useful for survey design (the appendix is a goldmine)

psycnet.apa.org/buy/2019-477...

Things they could have done would be: (a) cap Allen where it links to oxford; (b) remove parking on one side of the street so there can actually be a bike lane; or (c) make the bike lane on duncan or lawrence with the present divider in place.

Allen is totally not traffic-calmed, it's congested imo

This divider doesn't really seem to accomplish its goal. Right now, Allen St. seems to have more traffic, and with people parking on both sides of the street only a single lane for bikes and cars in places. It essentially funnels all traffic down Allen with no dedicated bike lane, so everyone loses.

Hahaha, no stones cast here, I'm no hypocrite

I think everyone has to learn that lesson a few times when doing academic writing, it's definitely been me some days too.

Sometimes you think you know how the world works and yet...

A great learning moment in mentoring a student.

She wrote a part in the paper, and just left a blank intending to find a citation to support the point. Then after hours of reading, discovered that the point she was trying to make was actually wrong!

Lesson is: Always read first, then write!

Also, living in this neighborhood I actually find that Allen street has more traffic now than ever, and Chebucto lane also sees way more traffic. Duncan and Lawrence are traffic calmed by this moreso, imo

I had heard that the reason for this is that the barricade cuts off so much of the neighborhood street access, that they need to have them movable for emergency vehicles if it arises (and also for snow plow access).

This is why I'm not a fan of the people putting rocks in the road here.

To me, genAI is a statistical model and its results can be imperfectly reproduced if the prompts are given. So I think of it more like how we'd treat proprietary stats software than a collaborator.

The difficulty is the very large number of use cases for genAI, which resists standardization

In those cases, it would be normative to put her in the acknowledgements section of the paper, at least in my field's norms.

I still hold the values of transparency and reproducibilty in research, so while a researcher using AI without declaring precisely how it was used isn't "cheating" per se, failing to fully cite sources and describe methods used is still poor quality research on these metrics, imo

I agree with a lot in the thread! For researchers, I would really like to see the norm of citing AI and providing the prompts used if people are going to use it.

In stats consulting, I am seeing PhD folks ask me to review what seem to be AI analysis plans and it is a huge time sink.

Ok, both legitimately awesome.

Does ... that mean you could in theory do a chloropleth for middle earth?

Ohh, hahaha I love that

Just once, I wish the government could like, function as a minority like we voted for.

Really not a fan of all this floor walking changing the outcome post hoc, not a good look at all.