I thought this op-ed about the upside of #AI in government transformation was very good. Among other things, it gave me an unexpectedly positive view of Palantir via its role in Operation Warp Speed and its attempts to battle bloated Pentagon procurements. washingtonpost.com/opinions/2024/…

My latest op-ed for @TheInformation highlights the way that in today's Silicon Valley, a few very deep-pocketed investors act as a kind of Central Committee, choosing those who will get the most money and have a head start in a race for monopoly. theinformation.com/articles/ai-ha…

Looking forward to this. I'll be talking about the need for tech platforms to support their ecosystems and to create more value than they capture if they want to stick around for the long haul.

How do dominant platforms use their #algorithms to shape markets & extract #rents from suppliers & users?

With applications to emerging #AI systems, here are key findings from 3 new papers & 2 years of research with Tim O'Reilly, Mariana Mazzucato (PI), & Rufus at UCL Institute for Innovation and Public Purpose 🧵1/

Remarkable, comprehensive thread from @ilanstrauss outlining our work on how internet platforms extract 'attention rents' from their users and their suppliers. If you can't read the papers, you'll get the gist here!

Virtually all other AI risks flow from this one. The race for monopoly breeds secrecy, which makes AI harder to govern. This is one reason why we need to commit to open source for AI. Licenses will require some rethinking, but it's the way to go.
techpolicy.press/monopoly-power…

This is high praise coming from one of the leading authorities on antitrust. Chuffed!

My new op-ed in The Washington Post, in which we discuss the deep state, kludgeocracy, PRA, NEPA, lawyers, Schedule F, hiring reform, and trusting public servants. Gets a little spicy. LMK what you think!
wapo.st/49y6j5T

So why is everyone saying the Biden administration is doing such a bad job with the economy? Evidence of our current post-truth media ecosystem, where hot takes driving traffic matter more than seeking out truth, or just a delay in catching up with the evidence?

So great to see our new UCL Institute for Innovation and Public Purpose algorithmic rents work covered by the incredible Rana Foroohar - paper cited can be found here: ucl.ac.uk/bartlett/publi… Financial Times Tim O'Reilly Mariana Mazzucato

I've tried to explain why Amazon's marketplace advertising business represents an abuse of market power, and the mechanisms by which tech platforms manipulate the attention of their users in order to extract money from their ecosystem of suppliers.
ucl.ac.uk/bartlett/publi…

Thanks to immigrants, the US won 75% of Nobel prizes in 2023. Without them, we only would have won 25%.

High skilled immigration is a policy lever for more innovation unlike anything else. Oh and its free.

I LOVE Watch Duty. Now that I'm spending a lot of my time living in a fire-prone area of the Sierra, it's my go-to resource for alerts about nearby wildfires and keeping up with location, acreage, direction of movement, containment, and evacuations.

When tackling the 'All we need is STEM!' attitude, my response: Developing the vaccine was the STEM problem; distribution & getting shots in arms was the Social Science problem; getting people to trust it & combatting misinformation was the Humanities problem--which did we fail?

Great stuff! I love the idea of open access journals devoted to cutting edge scientific research on global issues.

A fascinating AI watermarking proposal (giving AI its own unicode fonts that could look just like existing fonts but have a different code value) from Alistair Croll that should be explored posthaste.

wired.com/story/to-water…

Philip Guo's new article about programming with ChatGPT is one of the few that actually describes the experience of doing something non-trivial with ChatGPT, and talks about what works, what doesn't work, and what could be improved.
oreilly.com/radar/real-rea…

I am encouraged by continuing actions from the The White House to address AI discrimination, bias, harms, and risks. The voluntary commitments announced today must be a precursor to sorely needed legislation. The companies making these commitments cannot be allowed to check their…

I got a lot out of this episode of The Realignment Podcast with Richard Kahlenberg, especially as it relates to the truly key role of housing policy in all equity discussions. Our zoning rules really must change, and now. bit.ly/3Y3aP76

We may underestimate the value of AI because we overestimate the accuracy of humans

We likely wouldn’t pass the Turing test ourselves: “Almost half of our participants (42%) decided that their conversational partner (that was in every case a human being) was a computer program.”