Matthew Roberts

EduHacker / Instructional Designer / Political Scientist

[Matthew Roberts @mmcr]

Thank you to the engineer at Apple who ensured that when you’re flying in an airplane, the Peanuts Apple Watch face will show Snoopy flying in his doghouse.

[Matthew Roberts @mmcr]

https://t.co/xIL20wf8bX offered to turn one of my conference papers into a comic. I was skeptical—but apparently not skeptical enough. #WhatValueCouldThisConceivablyProvide #WhatsWrongWithTheSecondPanel

[Matthew Roberts @mmcr]

More evidence that the conversation about generative AI and education involves fundamental misunderstandings about basic concepts—even if you're a CIO. Yellow bubbles are my annotations.

https://t.co/loqHcvjXn0

[Matthew Roberts @mmcr]

I give a decent number of conference presentations that seem to go well. But the number one question that I almost always get asked afterwards is "how did you control your slides from your phone?"

[Matthew Roberts @mmcr]

https://t.co/rF4mFNZpu9

[Matthew Roberts @mmcr]

In her "Second Breakfast" newsletter today, @audreywatters echoes what I've been saying about generative AI.

[Matthew Roberts @mmcr]

I don't even know what to say.

https://t.co/NnPpnTHSUy

[Neil Renic @NC_Renic]

I am once again pitching my romantic comedy:

- two academics start dating
- discover they are each other's terrible reviewer
- hijinks ensue

Working title: Love is Double-Blind

[Matthew Roberts @mmcr]

I ate when I get the NYT answer correctly but don’t get how it makes sense.

[Matthew Roberts @mmcr]

@Rangerdavie Congrats!

[Gary Marcus @GaryMarcus]

A computer scientist’s perspective on vibe coding:

[Matthew Roberts @mmcr]

For a year and a half I’ve been saying that it’s only a matter of time before students either sue or demand refunds when they learn their instructors have used AI. #TwoEdgedSword

[Matthew Roberts @mmcr]

So far my only complaint about @liznorell’s “the Present Professor” is that no time machine exists so that I could have read the chapter on “academic culture” before I started grad school 26 years ago.

[Matthew Roberts @mmcr]

This explains a lot about why we seem to be on the Darkest Timeline.
https://t.co/JhfeCaCMjf

[Transluce @TransluceAI]

This seems like evidence that the untruthful answers are a more complex learned behavior, and not simply hallucinations due to basic language modeling.

Read the full conversation in ChatGPT here: https://t.co/DalMWJJKwJ
(5/5)

[Transluce @TransluceAI]

We tested a pre-release version of o3 and found that it frequently fabricates actions it never took, and then elaborately justifies these actions when confronted.

We were surprised, so we dug deeper 🔎🧵(1/)

https://t.co/IdBboD7NsP

[Find me on bsky @colin-fraser.net @colin_fraser]

At what point should it be reasonable to expect coherent answers to these? How far beyond PhD-level reasoning must we climb?

[Bryan Alexander @BryanAlexander]

The University of Michigan (my alma mater) is considering this today:
"Resolution to Establish a Mutual Defense Compact for the Universities of the Big Ten Academic Alliance in Defense of Academic Freedom, Institutional Integrity, and the Research Enterprise "

[Matthew Roberts @mmcr]

@RobertTalbert Careful. This sounds like a challenge. I'm sure there's at least one VC who likes jazz…

[Matthew Roberts @mmcr]

A thought based on reading this afternoon's selection of AI-related news: how do we, as a society of overlapping and interlocking communities, adjudicate someone's claim that [insert thing here] is revolutionizing [insert aspect of life here]?