MJ Yang

Gemini 2.5 Pro is very solid all around. Claude is better at coding. Last place, for most tasks: ChatGPT-5. However, I have actually been quite impressed with Chinese Open weight models such as Qwen3..

I have been incredibly fortunate to learn both from my amazing coauthors and a fantastic set of HBS alumni. And this work will be foundational for more forthcoming research (by others and myself) on how AI will impact the process of making business strategy going forward.

What is more, we establish causal evidence that business education has a persistent impact on CEO's strategy process -- even decades later. (So yes, business education does actually matter..)

Our basic result is that strategic decision-making styles vary widely across CEOs and that CEOs with more structured practices (proactive, consistent and hypothesis-driven), tend to outperform their (less-structured) peers especially in industries with high degrees of strategic complexity.

Very happy to see this paper in (online) print. Together with Michael Christensen, @raffasadun.bsky.social , @nickbloom.bsky.social and Jan Rivkin, we interviewed hundreds of CEOs to measure how they make business strategy.

pubsonline.informs.org/doi/10.1287/...

Yes, the idea that LLMs pay uniform attention to everything in a paper (or are "better at paying attention that humans") is a widely believed myth.

Financial crisis? "According to Moody’s, structured finance has become a popular way to pay for new data center projects, with more than $9 billion of issuance in the commercial mortgage-backed security and asset-backed security markets during the first four months of 2025." (Ian Frisch, NYTimes)

How the AI Boom Is Leaving Consultants Behind Consultants have a lot to gain helping companies deploy the most transformative technology in decades. Some clients say so far they have overpromised and underdelivered.

"Accenture, (...), in its most recent quarter reported a $100 million increase in new generative AI bookings quarter over quarter. That is down from a $200 million quarter-over-quarter increase the previous two periods."

www.wsj.com/articles/how...

Your next airline ticket could be priced by AI Delta is testing an AI-powered pricing system that could charge two travelers different fares even if they are purchasing at the same moment. Pricing strategy

PSA if you travel to SMS-SF: I used the SMS-provided promo code on United and compared it with exactly the same flight without the promo code. The code made my airfare $40 MORE expensive. Nice anecdote for price discrimination or business ethics?

www.colorado.edu/today/2025/0...

Full press conference: Governor names Charlie Kirk shooting suspect | CNN Utah’s governor Spencer Cox says they have a suspect in the Charlie Kirk shooting in custody. Cox identified the suspect as Tyler Robinson and presented the evidence investigators have uncovered at a ...

On the Charlie Kirk assassination: "Violence is the last refuge of the incompetent." (Isaac Asimov, Foundation)

www.cnn.com/2025/09/12/u...

Faith in God-like large language models is waning That may be good news for AI laggards like Apple

The LLM turning point is coming ( = we may be close to the "peak of inflated expectations" in the Hype Cycle).

www.economist.com/business/202...

Have we learned nothing from conflicts of interest during the subprime housing crisis? "Both authors have a financial interest in
www.expectedparrot.com. Horton is an economic advisor to Anthropic."

Excited to have my research (co-authored with the dream team of @mjyang.bsky.social, Zachary Kroff, and @erikbryn.bsky.social) featured in the New Yorker!

Incredibly excited for my brand new class at Rotman this fall: Progress! Econ history + theory + history of thought + philosophy on why rare orgs at rare times in rare places accomplish new things. Trying to put rigor onto an idea that is very much in the air. kevinbryanecon.com/Bryan-Progre... 1/3

Sam Altman says ‘yes,’ AI is in a bubble Is the bubble about to burst?

Seems Sam Altman thinks "Yes", AI is in a bubble? Yet, he wants to let OpenAI employees convert stock to cash with an implied valuation of $500 Billion? (with a B!). What happend to getting "PhD level" ChatGPT-5 expertise, .. (checks notes) .. two weeks ago?!?

www.theverge.com/ai-artificia...

There is an.....interesting op-ed in today's WSJ from a 22 year old entrepreneur telling everyone that work life balance will "keep you mediocre" It is completly wrong. Let me tell you why with citations. [gift link to the very wrong op-ed] 1/n www.wsj.com/opinion/work...

Odd, or signal that ChatGPT-5 is lower quality, but cheaper to run, because it decides when a high-token test-time compute run is triggered..

This also raises huge issues for replicability of research based on older ChatGPT models..

Urg

Confirmed, geez

There are two Bs!

ChatGPT-5 is here.

Will data centers crash the economy? This time let's think about a financial crisis before it happens.

You should read @noahpinion.blogsky.venki.dev on the data centre investments and whether they will lead to a financial crisis. I think he downplays the risk. It is potentially very high. www.noahpinion.blog/p/will-data-...

Disruption of the traditional MBA is here, whether we want it or not..

Consider: Google has a 90% market share in web and mobile searches. And has Gemini LLM results on top by default on any search. Are 1 trillion processed tokens really a sign of usefulness? (Same question about MSFT forcing LLMs on their developers..)

We see "revealed preference", just not for LLMs.

I have been thinking about this a lot, lately, after finding the troubling productivity impacts of AI on older firms ( conference.nber.org/conf_papers/... ), combined with prior work showing the impacts of digitization on older workers (www.nber.org/papers/w28094).
#EconSky

Importantly, the negative productivity effects are concentrated at old manufacturing establishments and firms. We can narrow down the mechanism as well, by showing that roughly half of the productivity loss is driven by removal of (human-supporting) structured management practices.

We use 2 separate datasets from the US Census Bureau and 3 separate identification strategies (matching, first-difference, IV) to establish credible causal effects. We find as of 2021 strongly negative productivity effects in the short-run and some evidence for potentially positive long-run effects.

SI 2025 Digital Economics and Artificial Intelligence

Naked self-promotion: My fantastic co-author @kmcelheran.bsky.social will present our empirical paper on "Industrial AI" (self-driving forklifts, self-optimizing production lines, autonomous quality control) and its effect on productivity in US manufacturing.

www.nber.org/conferences/...

Its not me, its the LLMs: in domains including coding, math, mapping, logic and now simple physics, they overfit and are unable to generalize. Still waiting for ChatGPT-5 btw..