We're training 100,000 Americans to defeat Trump's next coup attempt before, during, and after the midterms. Join us:

Hands Off Our Vote in 2026! · Indivisible

**Hands Off Our Vote** is Indivisible’s national program to ensure all eligible voters get to cast their ballot, every ballot is counted, and the results are honored. **By signing up you’ll become a part of the biggest election protection program in Indivisible’s history. And even if the elections are secure where you live, you can help protect at-risk voters in targeted areas during rapid-response moments and be a trusted messenger against the flood of disinformation.** We know that Trump and his Big Lie cronies are plotting to do everything they can to undermine our elections in 2026. Whether it’s trying to end mail-in voting, targeting registration programs, blocking vote-counting, even trying to ban states from encouraging people to vote – it’s going to be no-holds-barred from them this year as they try desperately to block the free and fair elections. But we’re going to be ready for whatever dirty tricks they try. **And we’re going to do it with community power and preparation, not panic.** You'll get updates on threats to our voting rights and learn how to counter them, and be ready to mobilize in defense of the vote, both in your area and nationally. We have tremendous power to protect our elections in 2026. **Sign up here to get started and join the Hands Off Our Vote movement.** **If you're in Alaska, Georgia, Iowa, Michigan, North Carolina, Ohio, or Texas, sign up [here](https://www.mobilize.us/indivisible/event/916480) as well.**

mobilize.us

It's not too late to plan a No Kings Day Voter Drive! #RegisterDemocrats the @FieldTeam6.org way - and help swing elections! We'll teach you all the secrets & coach you through the entire process.💙 📆 Various days & times #Voterizer #NoKings social.fieldteam6.org/s/Crf3ErAz

Sign up here:

social.fieldteam6.org

More voters are turning to AI for guidance on how & for whom to vote. This report looks at how well 126 campaign websites are set up for AI search and lays out what campaigns can do ahead of Election Day to improve their AI visibility & accuracy. innovatingforpublicgood.substack.com/p/the-ai-imp...

The AI Impact on Midterms – Part 1

AI crawler exposure and AI platform accuracy across crucial 2026 battleground races for House, Senate and Governor

innovatingforpublicgood.substack.com

Almost no 2026 campaign website is ready for AI search. That's the headline from a first-of-its-kind study sponsored by @ifpg.bsky.social, which scanned every House, Senate & Gubernatorial race rated toss-up — 63 races in 24 states, 126 campaigns. innovatingforpublicgood.substack.com/p/the-ai-imp...

The AI Impact on Midterms – Part 1

AI crawler exposure and AI platform accuracy across crucial 2026 battleground races for House, Senate and Governor

innovatingforpublicgood.substack.com

Written by @willrobinson1776.bsky.social and myself... My team studied every competitive House, Senate and Gov camapiagn sites: neither side is even trying on AI. But there is still time to change the one thing they have direct control on to inform AI Answers. substack.com/home/post/p-...

On AI: Nobody Is Minding the One Thing a Campaign Actually Controls

Tim scanned every campaign website in the 2026 battleground. Both parties are equally unprepared for the AI systems now answering voters’ questions - but this a problem you can fix in an afternoon

substack.com

Join the No Kings coalition for a nationwide election protection training to level up your election defense skills. We will focus on the role of sustained mobilization and rapid response work in our movement.

No Kings: Vote Early: Election Protection 201 · No Kings

Join the No Kings coalition for this training to prepare for engaging in sustained mobilization and results protection work depending on opposition response to election results. *A core principle beh...

mobilize.us

Tremendously excited to hang out with the folks doing the most in 2026 to keep the web healthy. If you're in London, pencil this in for after work next Thurs.

Open Web Advocacy@open-web-advocacy.org · 3w ago

🇬🇧 OWA is having a special meetup in London on the 1st of October 2026 together with London Web Standards. 🌐 Join us for talks on browser competition, web apps, browser choice and data portability followed by Q&A and a chance to meet people from OWA and others working on these issues. 🔗 More info:

The FBI is investigating a major hack of its own network that could affect thousands of federal employees.

Keystone Kash’s FBI Investigating Jaw-Dropping Breach Claims

A notorious hacking group has claimed responsibility for the massive cyberattack.

thedailybeast.com

Combining Signal-level UX with mesh networks like Reticulum for off-grid communication is the ultimate dream for digital resistance. What do you think is the biggest technical hurdle preventing a unified app like this right now?

How can organizations counter misinformation ahead of the 2026 midterms? In this clip, @timothychambers.net shares key findings from "Countering Disinformation: Strategies for 2026 and Beyond," a new report from IFPG & @deweysquare.bsky.social. www.psgconsulting.com/videos/v/cou...

Today we are releasing the next generation of our policy optimization tools for content classifiers. It is our hope this can be another step towards helping shore up human control over AI-powered systems. blog.zentropi.ai/optimizing-o...

Optimizing our Policy Optimizers

Today, we are releasing our next-generation policy refinement tools: policy-only correction, label-only correction, and auto-optimization.

blog.zentropi.ai

The safety team at RunwayML has been doing some of the most innovative work in the industry. It's exciting to see how they use Zentropi to ensure the safety of their videos in real-time **as they are being generated**. Details on their blog: runway.com/news/safety/...

Runway News | Moderation in Real Time

Runway built a synchronous moderation system for real-time video generation to scan streaming frames for harmful content in under half a second — layered on Runway's existing safety defenses.

runway.com

Runway is doing some very cool stuff using CoPE/Zentropi: "We have designed and tested a new fast moderation system that will run synchronously once a user’s input clears moderation.... The synchronous moderation through Zentropi narrows the exposure window to less than half a second."

Runway News | Moderation in Real Time

Runway built a synchronous moderation system for real-time video generation to scan streaming frames for harmful content in under half a second — layered on Runway's existing safety defenses.

runway.com

Curiious @davewilner.bsky.social and @samidh.bsky.social curious how you see the possiblity of #Jev and #Zentropi and COPE working together? Is there anyone looking at workdflows that combine Jev / COPE Zentropi (and maybe other larger LLM's based on the right tool for the right job?

I’ve also taken the time to consult with professionals in the content classification and trust & safety spaces, including @julietshen.bsky.social at ROOST and @dwillner.bsky.social & @samidh.bsky.social at Zentropi. But above all, I have a clear understanding of what Charcoal has to do.

The system is available now on zentropi.ai usable directly on the website or through the Zentropi Agent Skill for coding agents. Normally, full use is subscriber-only because of how compute-intensive it is, but for launch we’re giving anyone a limited quota of free so folks can try it.

Wooo, first Coop release I've directly been a part of — hoping you'll see more of that soon! I've been grinding on reducing tech debt and figuring out how to make the code more maintainable long term

ROOST@roost.tools · 3w ago

The community of ROOST contributors has been busy over the past several weeks working on the upcoming 1.1.0 release of Coop. We have a pre-release available for testing, and would love your feedback if you’re currently deploying Coop or are interested in trying it out.

This is super interesting. Would love to talk about what it'd look like to give @graze.social feed operators access to this

All of which is a roundabout way of saying zentropi.ai is now open for everyone! Free tier for testing, enterprise tier includes optimization. We've believed our system could take a 3-month process and reduce it to 3 days. Seeing it proved out with 50 experts? Worth the whole conference. 🧵 9/9

Watching people's faces shift from skepticism to excitement as they realized we'd built something that could finally help them was one of the more cathartic moments of my career. We can finally focus on what to do, instead of constantly struggling to make a way to do it. 🧵 8/9

@samidh.bsky.social and I built Zentropi because we thought we had something that could help. Instead of training many models for specific harms, we trained one model (CoPE) to read and follow written policies. And we built a product to help people quickly write measurably better policies. 🧵 2/9

There have been many attempts to farm out content moderation to AI that have been...not great. So I'm interested in this thread about Zentropi at TrustCon.

Dave Willner@dwillner.bsky.social · last yr.

For 17 years working in trust and safety, I've watched talented people burn out on impossible tasks. The problem isn't the people, it's the systems. Traditional moderation requires months of retraining for every policy change. Only big companies can afford it, and even then it works poorly. 🧵 1/9

Some more cool stuff from @dwillner.bsky.social @samidh.bsky.social and the Zentropi folks!

Dave Willner@dwillner.bsky.social · 4w ago

At TrustCon this year I talked about a technique we’ve developed for automatically optimizing content-moderation policies, using an inversion of the binocular labeling approach Zentropi had already pioneered. Today we're shipping the tool that technique became. blog.zentropi.ai/optimizing-o...

What Do You Allow AI to Act in Your Name? Agentic Identity, Delegated Commerce, and the Road Toward Conditional Access open.substack.com/pub/openchri...

What Do You Allow AI to Act in Your Name? Agentic Identity, Delegated Commerce, and the Road Toward Conditional Access

For years, the public debate about artificial intelligence centered on what a machine could generate: words, images, computer code, recommendations, forecasts, or deceptive media.

open.substack.com

“On Thursday, Mr. Gorton said that “the success of the polio vaccine was a massive medical fraud” and that widespread vaccination was a “public health disaster.” He received a standing ovation… Mr. Kennedy spoke next, linking arms with the group even as measles cases continue to rise”

RFK Jr. Assures Vaccine Skeptics They Have a ‘Friend at the White House’

The health secretary spoke at a conference featuring misinformation about vaccines, including a claim that the success of the polio vaccines was a “fraud.”

nytimes.com

Bari Weiss has plunged CBS News into ratings calamity. —"Evening News" has posted worst ratings of 21st century —"CBS Mornings" has posted worst ratings of 21st century —And now "60 Minutes" has posted worst ratings of 21st century More in @status.news: www.status.news/p/60-minutes...

Bari’s ‘60’ Bust

Bari Weiss blew up “60 Minutes," ostensibly to create a better show. Instead, her revamped premiere just suffered its worst debut ratings of the 21st century—what one longtime TV executive told Status...

status.news

The rule of law held the line again yesterday, w/ @democracydefendersfund.org & partners in the fight Before SCOTUS ruled, another fed. judge stopped Trump's absurd USPS rule nationwide & barred anyone from following it Every voter: vote by mail, vote early, VOTE @morningjoe-msnow.bsky.social👇 -TN

You can’t debunk what you never see. @timchambers.bsky.social and I wrote a Substack on prebunking - helping people spot manipulation before the next lie lands. Don’t just chase the last lie. Prepare for the next one. Read and share: willrobinson.substack.com/p/you-cant-d...

You Can't Debunk What You Never See

Scott Goodstein is right that we won't spot the next attack in time. There is a defense that doesn't require spotting it.

willrobinson.substack.com

Good perspective from Dave Winer about the news that Matt Mullenweg is on leave from Automattic: "I didn’t want to be the CEO of a tech company, and I expect that being the CEO wasn’t really a good fit for Matt either. A CEO... https://www.manton.org/2026/09/10/good-perspective-from-dave-winer.html

A little late to the party, but wowzers. Best wishes to everyone at Automattic: it's an important company filled with brilliant people, and they deserve stable leadership and an environment that supports them. techcrunch.com/2026/09/09/a...

Automattic's board forces CEO Matt Mullenweg into leave of absence | TechCrunch

Mullenweg said in a company Slack message that it was against his will.

techcrunch.com