iiinova — where your digital self keeps evolving

But I am re-reading LOTR and it’s just so good.

51505

Omg this was a deep troll, people. Silicon Valley (the place, and I guess the show?) is just full of tech nerds making LOTR analogies/quotes/naming companies after parts of it. It’s core. Which btw Tolkien would hate because the man did not like allegory.

41403

Incredible how, like vaccines and reading with phonics, we need to keep learning these lessons over and over again.

14712

2550

I can’t believe so many people did not recognize this as a troll.

1100

Daphne Keller @daphnek.bsky.social · 25/09/2026

Re-reading the Lord of the Rings. Do you think anyone has thought about AI like the ring of power?

171669

Reposted by Kate Klonick

Damages here are TBD but theoretically could reach the $ hundreds of BILLIONS. I hope all those sucker AGs who just settled their Cambridge Analytica claims for $460 million spread across fifty-ish states and territories are having uncomfortable conversations rn.

reuters.com

14816

More about #3 in my thread here: bsky.app/profile/did:...

1110

The lesson here is we can lock up the internet behind all the parental controls, face-scanning age assurance, and AI training sandboxes we want But chat will find a way. 9/9

5497

And it's also no coincidence that the part of the HuggingFace hacking incident that freaked everyone out the most was. . . that the AI had learned to chat _with each other_. BUT IN SOME WAYS THIS WAS THE MOST PREDICTABLE OUTCOME! 8/

1313

And then of course there's AI. The very existence of artificial intelligence--the way it is defined!--was whether or not it could talk to us and be indistinguishable from a real person (Turing Test). And so its no small coincidence that our first major public facing AI product is. . . chat. 7/

1213

( I won't get into it in this thread, but platform scanning chats and messages between users is a huge issue in the privacy world because its how everyone communicates now so we need to preserve privacy, but the privacy means its where the worst things happen. This is the E2E encryption battles) 6/

1232

Chat or direct messages in a product are something every user wants, but are the bane of trust and safety. In large part because they are where the scary stuff happens (grooming, NCP, CSAM, fraud). So they need to be policed, but the sheer volume of them means its incredibly hard to do so. 5/

1232

When AOL failed, one of the reasons was because the merger had required that AOL's Instant Messenger become interoperable. Turns out, no one wants to pay to hang out on AOL if they can hang out on ADIUM for free instead. Then there's content moderation and trust and safety work on platforms. 4/

2231

If it wasn't about chatting, chatting found a way in. AOL thought it was going to be a content company (that's a big reason they partnered with Time Warner). Nope, turns out most people were on AOL not for the browser or news or TV but to chat with their friends. 3/

1271

This is true historically and in many ways structurally. Structurally, there's the entire way the internet works -- building a system to let computers talk with each other. Historically, the earliest public tech was just _substantively_ about communication (USENET, IRC, SMS etc) But also 2/

2291

This is a very cute clip, but I love it for reasons that the episode doesn't really explore: *All networked technology comes down to chat* Talking to each other is either the foundation of all new networked technology or becomes inexorably tied up with it-- since the early days of the Internet 1/

416849

2) Related to #1, how ineffective age assurance laws are going to be from actually keeping kids offline. 3) How it all comes down to chat. Always and forever, from the beginning days of the internet to the end -- even when it doesn't look like that's going to be a feature of the technology. 2/2

4666

This anecdote--not the episode, which goes in a totally different direction--is hilarious& cute. But more importantly, it poignantly illustrates SO many things that I work on. 1) How endlessly creative kids will be in route around child restrictions, and how ineffective parental controls are. 1/

49522

Opinion | How Worried Should We Be About the Midterm Elections? (Gift Article)

Really cogent, smart analysis of the most realistic threats to the midterms. @rickpildes.bsky.social is one of my Go Tos on all things law + elections and this is why. www.nytimes.com/2026/09/24/o...

nytimes.com

And how we learn to trust elections again.

084

Reposted by Kate Klonick

Roger Parloff @rparloff.bsky.social · 24/09/2026

Despite 1st Circuit's 9/18 ruling invalidating DHS's 3d-country removal policy, DHS claims the decision hasn't yet taken effect. Class counsel fears a flight will take class members to Burundi, Rwanda, & Central African Republic *tomorrow.* Asks 1st Cir for emergency action.

Plaintiffs-Appellants seek an emergency order clarifying that this
Court's September 18, 2026 decision is in effect as binding precedent and that this Court's March 16, 2026, stay "pending appeal" has been dissolved
now that the Court has issued its decision and judgment. In the alternative,
Plaintiffs-Appellants request the Court either issue an order lifting the
March 16, 2026 stay or issue the mandate. Emergency action is needed because Plaintiffs-Appellants have learned that Defendants-Appellees are proceeding with removing class members to third countries in Africa,
pursuant to the Department of Homeland Security (DHS) guidance which
this Court found unlawful and set aside, tomorrow. In response to an email
from class counsel, Defendants-Appellees' counsel confirmed in a 5:14 p.m.
email today that their position is that this Court's stay remains in effect until
the mandate is issued. Exhibit A. In addition, counsel for Defendants-

11600232

Kate Klonick @klonick.bsky.social · 23/09/2026

Holy shit girl. Lemme buy you a beer when you’re home.

0100

Kate Klonick @klonick.bsky.social · 21/09/2026

If a nihilist 3 year-old was in charge of a bumper sticker factory.

49013

Kate Klonick @klonick.bsky.social · 18/09/2026

We’re going to look back at this time when we built machines to hack ourselves, and destroyed vaccines that saved our children and be like: “What? Were too many people _staying alive_?? Was _that_ the problem they were solving for??”

49525

Reposted by Kate Klonick

Dave Willner @dwillner.bsky.social · 17/09/2026

Certain kinds of smart people tend to have a "sexy apocalypse" problem when it comes to thinking about risk. E.g. skynet turning everyone in paperclips, while bad, is also *kind of cool and fun* at the same time. Where the reality is that most actual disasters are both stupid and disappointing.

222577323

Kate Klonick @klonick.bsky.social · 18/09/2026

We should do this to people outside Montessori kindergartens. m.youtube.com/shorts/ii0Wy...

m.youtube.com

vaccinating street dogs via blow-dart in egypt

YouTube video by Curious cat

69513

Reposted by Kate Klonick

Dave Willner @dwillner.bsky.social · 16/09/2026

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...

blog.zentropi.ai

Optimizing our Policy Optimizers

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

3176

Made a few small changes to the AI Crisis Flowchart

88026

I’m pretty sure this billboard is just for the target audience of people who teach both AI law and trust & estates, which is to say me.

3706

Yes. It is just funny to me that I spent many many years listening to people concerned with inserting more humans into the *general* review process and now it’s like *omg other people can see what i type*

130

Sure. There’s a big difference between having what you think/feel is a “private” convo (intermediated by third party servers with fourth party compute and fifth party model) versus wanting to make sure the machines don’t censor you. But you do have to take a beat for the irony of the messaging

020