Most of today's quantum computers rely on qubits with Josephson junctions that work for now but likely won't scale as needed for applications. cen.acs.org/materials/el... #chemsky 🧪

And @davidrotman.bsky.social tells us why the AI materials industry is still waiting for its ChatGPT moment. www.technologyreview.com/2025/12/15/1...

Week 5 of our State of AI collaboration with the @financialtimes.com - and this week we have @davidrotman.bsky.social and @richardwaters.bsky.social discussing the impact generative AI is having on the jobs market and the economy as a whole. www.technologyreview.com/2025/12/01/1...

Glad to share my new Substack post on the recent Nature paper by Joonyoung F. Joung and coauthors, introducing FlowER, a generative ML model that predicts reaction mechanisms while strictly conserving mass and electrons. open.substack.com/pub/bravoaba...

“I have no intention of being bullied to step down from my position because of some questions raised in a tweet,” Cook said. www.bloomberg.com/news/article... Amen.

You guys. The Lisa Cook story isn’t about her mortgages. Please please please stop being so credulous.

You can see the productivity boom everywhere but in the manufacturing statistics

Very interesting interview from @davidrotman.bsky.social, exploring how the loss of manufacturing jobs fueled US political divisions, and how we're failing to prevent the next "China shock."

DOE cancels $3.7 billion in demonstration projects aimed at cutting emissions — a couple big carbon capture demos, but also cement, glass, iron, chemicals projects. Calpine and Exxon Mobil were hit, as were startups like Sublime and Brimstone. (Full list in story.) www.nytimes.com/2025/05/30/c...

This is an essential point: For all the talk of nuclear, geothermal or maybe one day CCS, what folks are mostly building now to power AI data centers is new natural gas turbines — and they’ll be operating way past the point when the US power sector needs to hit net zero.

This is the deepest journalism I’ve seen yet on AI’s impact on the US power sector. I’m honored to be included w/ David Victor, Ari Peskoe, Costa Samaras, EPRI, and others. 🔌💡 www.technologyreview.com/2025/05/20/1...

The Trump administration’s broad tariffs won’t bring back manufacturing, and they could stunt the US’s ability to make tomorrow's breakthroughs.

@davidrotman.bsky.social digs into what tariffs will mean for manufacturing:

@caseycrownhart.bsky.social reports on the billions of dollars of climate tech projects that companies have halted since Trump took office. By way of comparison, it's "more than three times the total investments cancelled over the previous 30 months."

It's funny that there's been a prolonged "but what about the optimal case for strategic tariffs that promote critical sectors" debate and meanwhile the actual policy in question is just aiming a rusty shotgun at the kneecaps of the entire manufacturing sector for seemingly no reason

It's a straightforward relationship: high #inequality leads to the triumph of the far right. It will keep winning until governments deliberately seek to make societies more equal. I'm looking at you, Keir Starmer. This week's column. www.theguardian.com/commentisfre...

As this @ddayen.bsky.social interview with @alexsjacquez.bsky.social demonstrates. prospect.org/economy/2025... The one backhanded bright side?

Some much-needed skepticism on the "dire wolves" conversation from @antonioregalado.bsky.social www.technologyreview.com/2025/04/08/1...

Some of the stats in here are really striking. Per one paper, federal investments in R&D account for at least a fifth of all productivity gains in the U.S. economy since World War II. And the societal return for every $1 invested in the sciences is *conservatively* estimated at $5.

1. @alisongopnik.bsky.social, Cosma Shalizi, James Evans and myself have a new piece in Science on "AI" Large Models, pushing back against much of the collective wisdom about what they can and can't do. Official below, unpaywalled at henryfarrell.net/large-ai-mod... . So why this now?

To the rational Yanks out there: this is tha man you need to be reading: @djrothkopf.bsky.social He's been brilliant in seeing what was coming, in commenting on it as it was happening. And he's still brilliant in providing leadership on what to do. davidrothkopf.substack.com/p/how-do-dem...

Studies have shown that tariffs depress productivity in protected industries. U.S. steel is a case in point. (via @bennsteil.bsky.social) www.cfr.org/blog/steel-p...

Taking a cue from @cantlonlab.bsky.social, I have written an op-ed for my local newspaper about the impact of the NIH funding chaos. I'd appreciate it if you read and share. #NIH #Research #Science #Worcester www.telegram.com/story/opinio...

Lopatto Banger Alert (although it was my headline, give me that much) www.theverge.com/2025/1/20/24...

@lauradtyson.bsky.social talking about the productivity implications of AI - the topic du jour #ASSA2025. @smithcollege.bsky.social

Are we up to the moment we are about to face? Our parents and grandparents say yes we are. My latest. open.substack.com/pub/davidrot...

How to build the virtual cell with artificial intelligence: Priorities and opportunities "AI virtual cell, a multi-scale, multi-modal large-neural-network-based model that can represent and simulate the behavior of molecules, cells, and tissues across diverse states." www.cell.com/cell/fulltex...

It’s an idea as popular as it is incorrect: American babies now receive too many vaccines, which overwhelm their immune systems and lead to conditions like autism. www.nytimes.com/2024/12/14/h...

The AI race doesn't have to be zero-sum. International organisations already find natural alignment in specific domains, where competition naturally transforms into collaboration, not through grand agreements. Check out our policy paper

If blackouts occur for existing customers due to data center additions, it’s because the operator is prioritizing service for data centers over existing customers.

Want to increase worker productivity at the national scale? Turns out the answer may lie in funding federal nondefense R&D. I synthesized some of the existing literature on the impact of R&D funding on the economy in a new Research on R&D Funding blog: www.aaas.org/news/researc...

A hard and smart look at what happens next to place-based industrial policy. The word "uncertainty" comes up a lot. Thanks to @markmuro1.bsky.social and his colleagues at @brookings.edu for asking some important questions and doing their homework on this.

The pandemic, big stimulus packages, and other factors played havoc with the labor market and productivity over the past five years. But one thing we haven’t seen is a surge in technology-powered productivity. At least not yet…

In collaboration with Oxford colleagues and led by @zijian-zhang.bsky.social @scao.me we made strides in Agent-driven #selfdrivinglabs ! Furthermore we applied them to #quantumcomputing merging two of my favorite fields! Check our preprint, and code.

Anyone hoping that Musk will blunt Trump's climate denial needs to read this piece: www.washingtonpost.com/technology/2...

Meta's new data center in Louisiana is driving 2.3 gigawatts of new high capacity-factor gas units, according to Entergy. Inflexible data centers will drive massive US gas expansion without smarter planning. #energysky www.prnewswire.com/news-release...

The David Sinclair anti-aging companies story, front page WSJ wsj.com/health/welln... "Follow scientists who say Sinclair tends to exaggerate the findings and implications of age-related research, tarnishing a nascent field vying for credibility"

Zero-sum thinking is a key mindset that shapes how we view the world. A little thread to highlight our work on its roots with @sahilchinoy.bsky.social, @nathannunn.bsky.social, Sandra Sequeira.🧵1/23 scholar.harvard.edu/files/stantc...

Nice explanation from the always insightful @severinborenstein.bsky.social of the complexities and challenges for the grid in meeting exploding demand (or potential demand) from server farms and crypto mining.

My Energy Inst blog today on how much utilities should be obligated to accommodate new large loads: hyperscale server farms and crypto mining. Current paradigm puts risk on incumbent customers and gives new large loads very costly option on utility service. energyathaas.wordpress.com/2024/12/02/i...

I'm not sure how much longer our Mississippi Free Press account will post to X, but for now it makes a good case study—and a good case for why other newsrooms shouldn't fear moving to BlueSky. Look at the difference in engagement for identical posts with the same links. Left: BlueSky, Right: X 🧵

Here's a great starter pack for folks interested in the economics of innovation, competition and IO go.bsky.app/Rchu8QX

I wrote about the death and possible afterlife in of a weird little early college that specialized in teaching weirdos like me. It was called Simon's Rock. little-flying-robots.ghost.io/the-death-an...

"Cars assembled in the US rely heavily on parts from Canada and Mexico. Just 68 of 141 models recorded as having been assembled in the US had engines and transmissions made in the country. For 42 models, parts from Mexico accounted for more than 15% of the total value." www.ft.com/content/3d21...

The core task of democratic governments will be to reclaim the primacy of governance from tech companies, and why Trump’s election challenges this ↘️ www.theguardian.com/technology/2...

"Accurate knowledge in the relevant domain is what matters, but this is not what the industry has been investing in. Chatbots that can write Shakespearean sonnets will not empower electricians to perform sophisticated new tasks." A must read by @dacemoglumit.bsky.social.

Trump has now pledged to impose 25% tariffs on Canadian & Mexican goods his first day in office. That would add billions of dollars to the cost of steel — which we need for every sector of the clean energy economy — from the US's two largest suppliers: www.technologyreview.com/2024/11/25/1...

My latest @ft column: on.ft.com/4eIT8QV

What if AI doesn’t just keep getting better forever? we may already be hitting a plateau in the capabilities of large language models trained with standard methods. arstechnica.com/ai/2024/11/w...