How do people collaborate with o3/deep research when compiling & summarizing references?. It's fiddly to keep having to copy-and-paste to another document, or to tweak prompts and wait for a new response. I've been using a "scrapbook" workflow in cursor:. 1. Get o3 to compile a.

My daughter examined it & says in fact maybe people work more now, because it seems like with laptops me & her mum work a lot when we're at home too. :/.

Except perhaps in Germany:. (qualifiers: womens' participation, non-market work, quadratic scaling of images).

Time spent at work didn't fall as much as we expected in the 1950s:

RT @ellliottt: Call for Papers: . The 4th Annual Zurich Workshop in AI+Economics. to be held Dec 5-6, 2025, hosted by ETH Zurich and Univer….

Clear statement of a common view among AI researchers:. 1. There is a capability gap -- models pass all human exams but can't do real-world tasks. 2. The capability gap is closing fast -- the RL-on-LLM approach can deal with messy tasks (interactive, high-context, long horizon).

Macroeconomics in four lines:. Tonight the baker eats just what she baked.‍‍‍‍‍‍‍‍ the fisherman eats just what he caught.Because this morning at the market.‍‍‍‍‍‍ everyone tried to sell more than they bought.

The closest thing I'm aware of is @davidautor 's "Applying Ai to Rebuild Middle Class Jobs" -- though I don't think there's an explicit model in there, & as I recall he doesn't talk about some of the points above.

One nice model of LLMs is that they make private knowledge public. I haven't seen an economic model written down in quite this form, is there a reference I'm missing?. Suppose each person has a given level of productivity at each task, which reflects their knowledge, and.

Driving down the 5 from SF to LA and I keep checking the car's estimate of how long until we arrive, hoping with each glance for it to go down quicker. At some point I stop paying attention to the time remaining and start looking at the time elapsed. I realize I'm now wishing.

I think the effects of AI on the offense-defense balance across many domains is really important topic & I'm surprised there aren't more people working on it, it seems a perfect fit for econ theorists.

How does AI change the balance of power in content moderation & communication?. I wrote something on this (pre-OpenAI) with a simple prediction:. 1. Where the ground truth is human judgment, AI favors defense. 2. Where the ground truth is facts in the world, AI favors offense.