You ever really enjoy an apéritif and then read the fine-print…?

I need funny stupid internet videos. The more vintage the better.

There are two wolves. One always lies. One controls the lever of trolley track. They're in a state of superposition inside a box. On the trolley are infinite monkeys at infinite typewriters going to an infinite hotel. The trolley must first travel halfway, but before that a quarterway, and so on...

This is many years old (perhaps 2021?), but I thought maybe I should post my Shrek With A Pearl Onion Earring painting on this account

Some personal news werd.io/my-next-expe...

My next experiment

I'm spending a year at Stanford to explore community platforms in news.

werd.io

Can we stop AI from flooding scientific journals? cen.acs.org/policy/publi... "...the only lasting solution is changing how scientists are evaluated: away from volume and toward quality." Yes! This ☝️ a major conclusion in my last book. 🚨 Quality > Quantity 🚨

Can we stop AI from flooding scientific journals?

AI detectors may never beat fraud. Changing how science rewards its researchers might be the only fix that does

cen.acs.org

------> "Running a similar classifier task on Appiah’s own academic CV helps show the absurdity of this procedure. When I asked an LLM to classify the published titles and book descriptions in Appiah’s CV... I found that Appiah would seem to be part of the trend he wants to critique." 8/n

One of many things I appreciate in this post is an understanding that the medium/message determinant binary is unhelpful. Infrastructure and human living shape each other reciprocally. Before we technodeterminist, we should unwrap another layer.

Erin Kissane@kissane.myatproto.social · 3w ago

Well, good morning. Would you like to also look at some beautiful seaweed and think about knowledge? www.wrecka.ge/holdfast-0-s...

Live: @baratunde.com and New_ Public Co-Director @deeptidoshi.bsky.social open.substack.com/live-stream/...

LIVE NOW: Building Digital Public Parks with Deepti Doshi of New Public

Life With Machines on Substack

open.substack.com

There were two big lapses in the Hugging Face / OpenAI hacking story: the lack of security protections on OpenAI's end, and Hugging Face's vulnerable production database. That should worry anyone who uses AI with sensitive data. werd.io/ai-vendors-c...

AI vendors can't be trusted to secure their systems. Newsrooms need to act accordingly

There were two big lapses in the Hugging Face / OpenAI hacking story: the lack of security protections on OpenAI's end, and Hugging Face's vulnerable production database. That should worry anyone who...

werd.io

Bullet from MIT Sloan: "Experts said the five risks with highest expected severity are AI possessing dangerous capabilities, competitive dynamics, weapons and cyberattacks, power centralization, and AI spreading false or misleading information."

Hypervisible @hypervisible.blacksky.app · 2w ago

“Researchers asked 272 experts to evaluate 24 AI risks based on their likelihood and severity of harm between 2025 and 2030.”

too soon?

too soon?

He has the face of someone called Esteban.

Ironically Dos Passos got sucked into AI training twice

Mark Sample@samplereality.bsky.social · 4w ago

Pretty sure if you ran White Noise through an AI checker, it'd be flagged as AI-generated. DeLillo didn't invent staccato compound noun phrases or liberal use of words like "humming" or "fluorescent" (try John Dos Passos)—but he surely influenced the training data.

This is a huge deal - New York City is getting it's own Public Interest Technology Crew! (PIT CREW) www.nyc.gov/mayors-offic...

Mayor Mamdani Launches “Public Interest Technology (PIT) Crew” to Rapidly Build Digital Solutions to Public Problems

nyc.gov

It is something which is being done to us by a long, purposeful dismantling of institutions that support mass literacy. I have essay coming later this year, but basics are covered in this talk. Deliteracy is not symptom of algorithmic tech, the tech is being developed to assist deliteracy.

After OpenAI (Vandal Live at Wake Forest Humanities Institute)

Listen now (83 mins) | Apple Podcasts | Spotify

theamericanvandal.com

Ironically Dos Passos got sucked into AI training twice

Mark Sample@samplereality.bsky.social · 4w ago

Pretty sure if you ran White Noise through an AI checker, it'd be flagged as AI-generated. DeLillo didn't invent staccato compound noun phrases or liberal use of words like "humming" or "fluorescent" (try John Dos Passos)—but he surely influenced the training data.

Luke makes so many awesome clocks, but this is one of my faves yet. Extrapolating, a broken clock that is only right twice a day could be right all the time if it moved around carefully... #ItsFiveOclockSomewhere

Luke Steuber@coolhand.bsky.social · 4w ago

Five o’clock clock!

Wrote a short piece arguing that higher ed must help steer AI. TLDR: If we outsource this to tech, we outsource our whole business. But rejectionism is basically stalling. If we want to survive, schools themselves must proactively shape AI for education & research. [1/6, unpaywalled at 5/6] +

Opinion | AI Is the Future. Higher Ed Should Shape It.

If we want to stay at the forefront of knowledge production, we must fit technology to our needs.

chronicle.com

The future of AI shouldn’t be decided by a small number of companies alone. Mark Surman's piece in Fortune about why open-source AI is emerging as a critical path toward a more transparent, accountable, and human-centered future. fortune.com/2026/06/29/o...

Mozilla President: meet the open source 'rebel alliance' that could break Big Tech's grip on AI | Fortune

Middle powers like Canada and the EU are quietly building an alternative AI future — one that's open, accountable, and not controlled by a handful of labs.

fortune.com

Any scholar who is concerned about the politicization of academia and wastes energy on a moral panic about “too much woke” when fascists are taking over whole institutions and gutting research funding should be ashamed of themselves.

Chronicle of Higher Education@chronicle.com · 4w ago

Opinion | Ashley T. Rubin, a sociologist who was among the authors of the recent report on the state of the humanities and humanistic social sciences, offers some hypotheses about activist scholarship. https://chroni.cl/4vZhpf2

Good short essay on the successes the Democratic Party has when it has engaged to the left. https://campaign-trails.ghost.io/punching-left/

Punching Left

Over the last month, there's been a flurry of panic from self-styled moderates and centrists in the Democratic Party about the success of a few socialist candidates in party primaries. Some of this grumbling has simply been the self-interested reaction of inert incumbents who are (rightly) worried

campaign-trails.ghost.io

A delightfully imperfect handmade flyer is infinitely more compelling

We Are Living in a ‘ChatGPT Flyer Pandemic’

"Hey if this is your flyer, I’m not going, I’m not donating, I’m not sharing. Don’t ask me."

404media.co

Shannon Mattern@shannonmattern.bsky.social · 10mo ago

A favorite flyer + a good LA pole that has held lots of flyers :) A have a chapter abt publicly-posted writing — incldg flyers — in my 2017 book; I was inspired by Johanna Drucker's work on public signs, Lisa Gitelman's writing about poles, + Jacob Riis's photo of Chinatown pole postings :) ...

GenAI can check its own output, but it can't tell you what it doesn't know it's missing. That gap matters most in K-12 edtech, where FERPA, accessibility, and student data are on the line. We call it "comprehension debt." Learn more:👉 buff.ly/YgGvUDr

The MIT Data + Feminism Lab is hiring a one year postdoc! We are working with Andrea Parker (GA Tech) and Brooke Foucault-Welles (Northeastern) on a project related to youth as co-designers of civic technologies to support neighborhood community wellness. careers.peopleclick.com/careerscp/cl...

Postdoctoral Associate

MIT - Postdoctoral Associate - Cambridge MA 02139

careers.peopleclick.com

This essay was very good.

Frank Pasquale@frankpasquale.bsky.social · last mo.

“The most obvious reason that culture brims with the melting-together of the goofy and the horrible is because undeniably stupid stuff really is crashing into the very serious news in undeniably disturbing ways” news.artnet.com/art-world/ma...

Does anyone have any advice for a parent whose daughter goes around all day singing random songs in the style of Michael McDonald? #AskingForAFriend

obviously the best way to hold up Both Sides™️ is with a strong floor

Sharon 🪳🌹@sharonk.bsky.social · last mo.

currently reading through the "Strong Floor, No Ceiling" book that Jeffries and a couple of other Dems have been using as a tagline for policies: it's so funny because it has lines like these

Was just reading @willkurt.bsky.social interesting post on manipulating LLM text outputs via...IIUC...an entropy crossfader. Wondering @eryk.bsky.social saw this too. www.countbayesie.com/blog/2026/7/...

Making LLMs Better at Creative Writing using Entropy — Count Bayesie

Tired of boring LLM writing? In this post we explore how using entropy to modify the sampler can dramatically improve the quality of LLM writing!

countbayesie.com