Melanie Mitchell

Analogy and metaphor galore!

I was interviewed by the @nightsciencepod.bsky.social podcast about metaphor and analogy in science:

open.spotify.com/episode/1hqU...

Pairs well with @sflusberg.bsky.social's interview on metaphor on the @manymindspod.bsky.social podcast
disi.org/the-aura-of-...

Phil, thanks for your clarifications on this. Yes, there were many scientists in Epstein's cabal, but it would be tragic if that fact leads to even more anti-science sentiment in the public (as well as conspiracy thinking about even simple algorithms).

Mood:

(h/t @lettersofnote.com)

P.S. For people who listened to the show: It turns out that Claude's "funnier" (and NSFW) joke was taken almost verbatim (and without citation) from a three-year-old Reddit post. So my guess that it was in the training data was correct!

Brainwaves: Is AI actually thinking? The rate at which artificial intelligence is able to replicate human behavior has increased in recent years. Does that mean it's thinking like us? In the third episode of "Brainwaves," what artificial...

I enjoyed this fun discussion with Meghna Chakrabarti and @kmahowald.bsky.social about thinking, human or machine, on WBUR's On Point show.

(Top of page is condensed version, scroll down to "On Point Full Broadcast" for full version.)

www.wbur.org/onpoint/2026...

Thoughts on this?

bsky.app/profile/been...

How can I convince Chris Hayes to read my Substack?

On Brian Cantwell Smith and the Promise of AI Today I had the bittersweet pleasure of participating in a symposium honoring the late philosopher Brian Cantwell Smith, a good friend whom I’d known for over 30 years.

I wrote about the late philosopher Brian Cantwell Smith, and his profound thinking about AI and the nature of intelligence.

aiguide.substack.com/p/on-brian-c...

(4/4)

Brian distinguishes between "reckoning" -- the kind of computational prowess that enables AI systems to win gold medals at math competitions, and to write complex code -- and "judgment": a deep, caring commitment to the world that enables ethical action. (3/4)

The Promise of Artificial Intelligence: Reckoning and Judgment An argument that—despite dramatic advances in the field—artificial intelligence is nowhere near developing systems that are genuinely intelligent.In this p

This book is amazingly prescient, and should be read by anyone who claims to know what "intelligence" means, and claims we are close to "AGI". (2/4)

direct.mit.edu/books/book/4...

Festschrift Honouring the Late Professor Brian Cantwell Smith - Faculty of Information Morning 9:00 AM: Welcome Coffee & Registration Coffee and pastries will be served. 9:40 AM: Opening Remarks Host: Kelly Lyons, Professor, Faculty of Information, cross-appointed with the Department of...

On the eve of the University of Toronto's festschrift honoring the late philosopher (and overall wonderful human) Brian Cantwell Smith, I'm re-reading his 2019 book, "The Promise of Artificial Intelligence". (1/4)

ischool.utoronto.ca/festschrift-...

The #ACMBookClub is back!

We’re kicking things off on Sunday, Feb 22, w/ a deep dive into Melanie Mitchell’s acclaimed Artificial Intelligence: A Guide for Thinking Humans, and a finale conversation w/ the author on March 24.

Register today: buff.ly/a8L7qzA

Bibliography cleanup is known to be harder than AGI. Same with successfully connecting a laptop with a projector. It's what we humans will be doing long after the robot takeover. 😇

I'm so sorry to hear you were laid off -- what a loss for all of us! I hope you will find another great job soon. So sad what is happening to WaPo.

Large Language Models and Emergence: A Complex Systems Perspective Emergence is a concept in complexity science that describes how many-body systems manifest novel higher-level properties, properties that can be described by replacing high-dimensional mechanisms with...

Hi Eli, Here's a paper some colleagues and I wrote on this: arxiv.org/abs/2506.11135

Don’t forget that he held the title of “Chief Learning Officer” at Trump University.

An Artificial Intelligence Developed Its Own Non-Human Language When Facebook trained chatbots to negotiate with one another, the bots made up their own way of communicating.

Anybody remember this from 2017?

www.theatlantic.com/technology/a...

Plus ça change, plus c'est la même BS.

AI agents now have their own Reddit-style social network, and it's getting weird fast Moltbook lets 32,000 AI bots trade jokes, tips, and complaints about humans.

Good take on the "Moltbook" story from @benjedwards.com

arstechnica.com/information-...

Another good take I saw on the other site: "more of a fanfic for AGI than actual AGI". This could describe a whole lot of AI misalignment red-teaming....

Another good metaphor I saw over on the other site: "more of a fanfic for AGI than actual AGI"

Apt metaphor.

This was a lot of fun and @melaniemitchell.bsky.social's newsletter is a rare island of coherence & measured discussion on AI

aiguide.substack.com

#ACMBookClub 2026 Webpage

I'm excited that the ACM book club will be reading and discussing my book on AI! I'll be joining on March 24 for the final discussion.

www.acm.org/articles/edu...

harcx Verify .bib file citations against academic databases (Semantic Scholar, DBLP, Open Library)

My UCSB colleague Gurusha Juneja wrote a python tool to auto-check your bibtex for hallucinated citations: pypi.org/project/harcx/

It supports papers, books, and
URLs.

Usage:
pip install harcx
harcx references.bib

⬇️

The Waymos mistook them for stopped ostriches.

(ML in-joke, sorry)

Opinion | Trump’s Norway Letter Proves This Isn’t Sustainable

Gift link: www.nytimes.com/2026/01/21/o...

Opinion | Trump’s Norway Letter Proves This Isn’t Sustainable

"There is no need to go on about the lunacy of this letter. Not even the president’s most eager apologists could spin this as anything other than what it is: a disgrace." @jamellebouie.net knocks this one out of the park. www.nytimes.com/2026/01/21/o...

Has anyone tested current LLMs' ability to do this?