This new report from @knightgtown.bsky.social @pfchap.bsky.social @matt-steinberg.bsky.social shows the substantial gap between platform risk assessments and their actual approaches to mitigating risk. (1/13)

I wrote a new brief with @premmtrivedi.bsky.social, AI Agents and Memory: Privacy and Power in the Model Context Protocol (MCP) Era). As AI agents reshape how we interact online, we discuss the new challenges agents pose for privacy, security, and competition. Hope this can be a helpful resource!

Ultimately, i don’t think AI advertising works. Ads are inherently about differentiating your brand, and AI slop makes everything look the same.

solidarity tonight with all my fellow dodger fans on the east coast. stay strong

If you want a lens into what the future of good tech could look like, this is it. I had the chance to help research and write "Making AI Work for the Public", a new report from @newamerica.org’s #ReThinkAI collective. Proud of this one — check it out! www.newamerica.org/technology-a...

Remember the 2010s “Internet of Things” when our toasters suddenly connected to WiFi and became surveillance tools? We should be ready for the 2020s “AI of Things”.

Retweeting this so I can more easily find it again in 3 years when we inevitably need to refer back to it.

10 points to whoever wrote this line

“The problem is that our rules assume technology is either a product (regulated for physical safety), a service (regulated for speech), or an app (governed through platforms). Embodied AI is all three, but no single regulatory agency or statute can oversee the full picture,” writes Matt Steinberg.

Excited to share my latest in @techpolicypress.bsky.social! Remember when the 2010s Internet of Things felt exciting... until we realized our toasters were surveillance devices? I look at what policies were missing back then, and suggest how we can adapt for the AI of Things. tinyurl.com/35ah9sd7

How seriously are people taking the bit of good in Trumps Ai Action Plan? The emphasis on open-source and open-weight models is great for many reasons (no ai platform lock-in!) but i am too cynical to even start to consider possible implications of this

This was a great convo

Hard to stress how bad this is. That's 79M medicaid enrollees ethnicities and home addresses. It's not just some abstract privacy issue. It's the kind of thing that can directly deter immigrants from trying to access basic healthcare

🚨 Excited to share an op-ed I wrote for @fortune.com about OpenAI’s recent attempt to partner with Hollywood -- and what it reveals about the future of AI and creative industries. The good news? AI companies have a chance to see that that the ethical and economic incentives are actually aligned.

So competition, not monopolies, leads to innovation after all! Thanks Bluesky 🙏

They’ll keep hiding their attack on research as “defending free speech” and “fighting censorship”. It’s scary, but very effective, and institutions need to figure out a strong counter asap

As November ends and the holidays approach, it’s officially time to panic about your Spotify Wrapped, and start binge-listening to avoid that one embarrassing artist in your top 5