Carlos T. Estrada Arzamendi

I thought this was about “Attention Is All You Need” for a second and they had the wrong year 😭

Goodreads is the new Ferrari F1 team sponsor!!

This is my make or break year. Time to lock in.

You don’t have to go as far as Australia. These are vendors in Amazon.

The TRM paper feels like a significant AI breakthrough. It destroys the pareto frontier on the ARC AGI 1 and 2 benchmarks (and Sudoku and Maze solving) with an estd < $0.01 cost per task and cost < $500 to train the 7M model on 2 H100s for 2 days. [Training and test specifics] For ARC, it trained on 160 examples from ConceptARC. At test-time, it uses the most common answer of 1000 augmentations at test-time and embeds a fixed shape of the task in the input. [Industry implications] Most AI companies today use general purpose LLMs with prompting for tasks. For specific tasks, smaller models may not just be cheaper, but far higher quality! Startups could (and should) train models for < $1000 for specific "fixed length" subtasks (specific PDF extraction, time series forecasting, etc) and use it as a tool to the general model to not only push performance, but build some meaningful IP at the task they're trying to automate.

The H-1B worker program in the 1990s: • Raised output of IT goods by ~2% • Lowered prices for consumers • Boosted firm profits • But reduced wages & jobs for U.S. computer scientists

We want to get to know more Florida Economists! Join us! #econtwitter

I am going to print this, put it up on my wall, and look at it when any situation looks utterly hopeless.

University Avenue. Gainesville, Florida. April 7, 2025

What a comeback! Congrats to Walter Clayton, Jr. and the Florida Gators on their third national title. This was a fun team to watch.

YOUR 2025 NATIONAL CHAMPIONS.

no shit

So true, Mr. Dragon

Last month we launched our Anthropic Economic Index, to help track the effect of AI on labor markets and the economy. Today, we’re releasing the second research report from the Index, and sharing several more datasets based on anonymized Claude usage data.

AI that empowers decision-makers to utilize knowledge will encourage scientists to pursue more novel research questions, from @joshgans nber.org/papers/w33566

This paper has been in the making for so long that eight children across the co-author group were born between the time we started and now! But it is finally out, and you can check it at arxiv.org/pdf/2503.13323

babe, come to bed i can't, i'm vibe coding

How to educate given AI is a major priority. Tons of folks in education who think AI is a "plagiarism engine" & have no idea of current capabilities makes it hard! But calculators can do anything a 5th grader can; Babelfish translated better than 101 students. We still taught!