I don't know why.

The non-Anthropic employee, Tristan Buckmaster, is the one paying for OpenAI models and presumably the one who chose to use them. The Anthropic employee, Levent Alpöge, was collaborating in his personal capacity, and obviously it wouldn't make sense for him to cut off their work together just because his employer's competitor's tool was used.

Could be a copy/paste error, or could be an LLM trying to cite its own work by just looking at a list of sources used and guessing which one the claim came from.

It's possible the company would have failed on its own, but we'll never know, and it wouldn't be because their product was too good. If that were the case then all the recommended alternatives in the Worse on Purpose article would also be out of business. This company looks pretty comparable, and has been making basically one thing very well for 100 years: https://www.allamerican1930.com/pages/about. How often do people need to buy a new high quality pressure canner? They seem to be making it work despite durable products in a relatively niche market.

There are a lot of citations in the paper backing up the idea that handwriting promotes learning, though (or at least spelling, memory, recall, and letter recognition/understanding). It reads to me like an attempt to explain an observed phenomenon: "given that handwriting appears to promote learning, what could be a mechanism for that?" It isn't in a vacuum just claiming that "these brain signals clearly prove that learning is happening".

The "general discussion" section also covers why the specific connectivity they looked at might be related to learning.

> I can say that, when writing by hand, one is continually solving a series of small problems having to do with the spacing of words, how letters are connected, the crossing of the letter t (sometimes more than one in the same word) and the dotting of the letters i and j, and how to accomplish all of those things through coordinated movements not just of the fingers but of the whole arm.

Do you think he needs to be a neuroscientist to make this claim?

Comments URL: https://news.ycombinator.com/item?id=48674000

Points: 7

# Comments: 0