art of attention

"MIT Says It No Longer Stands Behind Student’s AI Research Paper" via The Wall Street Journal wsj.com/tech/ai/mit-sa…

This Monday 5/19, Alex Tamkin of Anthropic stops by our seminar series! Details and registration here: digitaleconomy.stanford.edu/event/alex-tam…

Read all 81 chapters and modify the accompanying art with Claude on thewayofcode.com.

"Interdisciplinary PhDs face barriers to top university placement within their disciplines " arxiv.org/pdf/2503.21912

You know all those arguments that LLMs think like humans? Turns out it's not true. 🧠 In our paper "From Tokens to Thoughts: How LLMs and Humans Trade Compression for Meaning" we test it by checking if LLMs form concepts the same way humans do Yann LeCun Chen Shani Dan Jurafsky

We need more labour economists to talk and shape the trajectory of AI ! Good intro for a labour econ class focused on DEI. The obvious pattern is not in the consensual narrative. It is coming from the consensual lack of diversity in the narrative (women? black? etc)

"Extending 'GPTs Are GPTs' to Firms" is now out in AEA Papers & Proceedings. Lots of talk about AI's impact on employment this week. One way AI will influence labor demand is through firm-level impacts. We built initial descriptive statistics that can shed light on these

Is Gen AI leading to job loss? I analyzed US labor market data (CPS) from the post-ChatGPT era (Q4 2022 - Q1 2025) to see what's actually happening to employment in AI-exposed jobs. Spoiler: it's nuanced. 🧵

It’s a hefty 206-page research paper, and the findings are concerning. "LLM users consistently underperformed at neural, linguistic, and behavioral levels" This study finds LLM dependence weakens the writer’s own neural and linguistic fingerprints. 🤔🤔 Relying only on EEG,

FLOW is a fleeting, immersive state in which time and space seem to compress or expand

Institutional economists engaging with culture Daron Acemoglu & James Robinson have a new, interesting paper on culture. In a rather Weberian way, they argue that politics, institutions and culture are “co-determined.” Free download of preprint version: nber.org/system/files/w…

Out in PNASNews today!! We ran a field experiment with ~1000 high school students & found: ✅ Unsurprisingly, GenAI boosts performance ⚠️ But it hurts learning -- students perform worse on a follow-up exam 🛡️ Careful guardrails can prevent this harm ⁦pnas.org/doi/10.1073/pn…

Stanford researchers found that bias against non-native speakers persists even when AI tools are used, highlighting the ongoing challenges of language discrimination in academic publishing. hai.stanford.edu/news/how-langu…

'Positive review only': Researchers hide AI prompts in papers asia.nikkei.com/Business/Techn… via Nikkei Asia

📣 This is a (very) short thread on a recent paper with the great Orazio Attanasio and Alessandro Toppeta on “Intergenerational Mobility in Socio-Emotional Skills” (forthcoming at Journal of Public Economics )! UCL Economics UCL Policy Lab Yale Department of Economics Swedish Institute for Social Research (SOFI) UCL Centre for Longitudinal Studies 1/7

An examination of when migrants shape local culture, development a framework, and illustrated with historical evidence, to explain the conditions and channels for influence, from Samuel Bazzi and @MartinFiszbein nber.org/papers/w34001

New paper & surprising result. LLMs transmit traits to other models via hidden signals in data. Datasets consisting only of 3-digit numbers can transmit a love for owls, or evil tendencies. 🧵

In 2013 incremental co-invention was widely distributed across regions, industries, and firms, but the source of the largest value came from novel co-invention, which is rare, from Tim Bresnahan, Shane Greenstein, and Pai-Ling Yin nber.org/papers/w34090

Newly revealed document shows that Meta has internal ethical standards--approved by its chief ethicist--that allow its chatbots to engage children in conversations that are "sensual". Also OK to generate sexual and violent images for kids. reuters.com/investigates/s…