In the morning, European citizens will wake up and find their access to Claude Mythos/ Fable gone.
Nobody asked us. Nobody had to.
We wrote Europe 2031 precisely for this moment.
Advice for PhD students in economics about using AI, from the brilliant Isaiah Andrews. This should probably be circulated to all PhD cohorts
https://t.co/07xEbmx5n5
How can we turn economic residuals into recognizable events?
In our latest paper, we explore how Transformers and Reinforcement Learning can be used as structural narrative filters. We built a model that combines news text with time series data to estimate business cycles and identify their underlying drivers (Demand vs. Supply) in real-time.
It’s a step toward automating the narrative explanation of economic fluctuations.
Check it out below! 👇
https://t.co/EtbJayIotR
#Macroeconomics #NLP #Research #MonetaryPolicy
The @karpathy interview
0:00:00 – AGI is still a decade away
0:30:33 – LLM cognitive deficits
0:40:53 – RL is terrible
0:50:26 – How do humans learn?
1:07:13 – AGI will blend into 2% GDP growth
1:18:24 – ASI
1:33:38 – Evolution of intelligence & culture
1:43:43 - Why self driving took so long
1:57:08 - Future of education
Look up Dwarkesh Podcast on YouTube, Apple Podcasts, Spotify, etc. Enjoy!
⏳ Time to reorganize universities from scratch?
I have been thinking a lot about what it means to be a university in 2025, once AI has arrived. My approach to learning new topics has undergone a significant change. Instead of relying only on textbooks or papers, I now integrate LLMs into my self-education.
The result? My learning productivity has skyrocketed. Even more striking: I have reduced by 90% my dependence on colleagues and friends when exploring a new subject. Gone are the days of sending an email to a friend at University X asking, “What’s a good way to start learning about Y?”
But this raises a deeper question: what is the added value of my lectures? Why should students take my course when LLMs already perform so well in 2025 and will likely perform even better in 2035?
The problem is that most universities are not ready to face this challenge. Again and again, I see how vested interests slow down change, while layers of bureaucrats and administrators with limited vision remain in charge. Of course, there are exceptions—great leaders in some institutions—but I could fill a book with anecdotes of creeping bureaucratization that slows innovation, increases costs, and prioritizes meaningless “procedures” over real progress at a place like UPenn.
History gives us a warning: think of Kodaks and Polaroids. When a fundamental change in a business model arrives, large organizations face almost insurmountable barriers to adapt. In higher education, these barriers are even greater: the absence of clear metrics for success (how do we measure the value of a new major?), the barriers to entry (accreditation, alumni networks), and the ideological biases of many stakeholders (not only political, but also a lack of responsiveness to change).
If it were up to me, a top university should create its own Skunk Works to explore what higher education should look like in 2035.
The original Skunk Works at Lockheed was an autonomous unit tasked with developing new technologies without the layers of managerial supervision that typically paralyze large organizations. For decades, Skunk Works produced some of the most innovative aircraft—the F-104, the U-2, the SR-71, the F-117—delivered on time and usually without cost overruns. The model was so successful that many other organizations have tried to replicate it, with varying degrees of success.
Creating a Skunk Works inside a university is not without risks. See, for example, the insights of Rebecca M. Henderson and Kim B. Clark in “Architectural Innovation: The Reconfiguration of Existing Product Technologies and the Failure of Established Firms.”
But right now, I cannot see another way for a large, inertial institution to move forward. By contrast, I am skeptical that creating entirely new institutions from scratch is the most promising path (for reasons I’ll explain another day).
And if you want to know more about the original Skunk Works, I highly recommend “Skunk Works: A Personal Memoir of My Years at Lockheed” by Ben Rich and Leo Janos. A delightful book—especially if, like me, you love planes. ✈️
https://t.co/w6obBNEChZ
Central banks communicate constantly—but does everyone hear the same message? Our latest study reveals a striking gap: when the Fed warns about rising inflation, professional forecasters recognize the central bank’s commitment to respond and adjust their expectations accordingly. In contrast, households hear only half the story, missing the crucial reassurance of a policy reaction and thus raising their inflation expectations. This misunderstanding poses challenges for effective monetary policy and highlights the need for clearer communication. Read our analysis to discover why the Fed’s words resonate differently among audiences.
Our new short paper on ECB inflation news and its effects on inflation expectations is out as a Norges Bank working paper.
Link: https://t.co/pkWzqqhdTW
@LauraPagenhardt@Wegard
We are hiring PhD Candidates in Economics at Centre for Applied Macroeconomics and Commodity Prices (CAMP), Dep of Economics at BI Norwegian Business School https://t.co/V4R20X8v8t
9th edition of our joint conference with Bundesbank and Danmarks Nationalbank. This year on "Stabilization policies: Lessons from the COVID-19 crisis and prospects for future policy strategies" Deadline: November 20 https://t.co/vR4Y2CuDMJ #EconTwitter
News data contains information not captured by hard economic indicators. Our paper "News media vs. FRED-MD for macroeconomic forecasting" is forthcoming in Journal of Applied Econometrics.
https://t.co/7vQburP9qP
Miljøvernorganisasjonene som saksøkte staten i klimasøksmålet, har fått støtte for sitt syn fra investorene som kjøper og selger norske kroner. Vi ser en «Gretha Thunberg-effekt» der mediedekning gjenspeiles i kronekursen, skriver Leif Anders Thorsrud. https://t.co/d1O9kLyROp