#language
14 découvertes · Découvertes choisies sur le web ouvert, des plus récentes aux plus anciennes.
-
How the Words We Teach English Language Learners Changed
How the words we teach English language learners changed.
Pudding · EN
-
“Curious Address”
I send you a post-card which I received in the ordinary way by post from my brother, who lives at Sutton Scarsdale, a scattered village near Chesterfield. You will notice that the card was posted at 7.15 p.m. on the 5th October, and it was delivered during the evening of the following day. The address looks a mixture of Greek and German, but on inspection it will be found that each letter is spelled out in full. The pencilled words were inserted by the Post Office officials. The Post Office is often the object of complaints for tardiness in delivery,...
Futility Closet · EN
-
A Pearl
Writing to Robert Gathorne-Hardy in 1929, Logan Pearsall Smith referred to an account in the London Times of a solar eclipse as seen from Manila: The weather was cloudless, the conditions were perfect. Acacia trees closed their leaves as for the night; dew fell, chickens roosted, and the peasants in outlying villages, terrified by the awful phenomenon, supplicated the Saints. Smith wrote, “Could this sentence, which Flaubert might have written with the help of Gibbon, be the chance product of a journalist’s pen? I cannot think so.” For the record, a Times archivist attributed the passage to Wi
Futility Closet · EN
-
Why Language Models Surprise You
The Coding Train · EN
-
The AI Language We Can't Read: Neuralese ft. Rob Miles - Computerphile
Computerphile · EN
-
Show HN: A language model that lives in a boot sector
All the ideas came from the boot sector demo scene where people had been writing chess and BASIC games in 512 bytes, even a C compiler. I wondered what one can do using a neural network instead of a program and generating boot logs for non-existent hardware was something fun to try. Entire inference engine is 510 bytes of MBR, weights, vocab, and strings reside in the sectors following MBR. Small char-level neural network was trained on dmesg generated by machine, so machine produces Linux boot messages infinitely. train.py performs precisely the same integer arithmetic operations as the asm c
Show HN Best · EN
-
Designed for a Dead Language
Every language app in your pocket inherited a teaching method built for Latin. Understanding why that happened is a more useful design lesson than anything the apps themselves will teach you. In 1788, Prussia introduced the Abitur, a standardized national examination required for entry into universities and the civil service. To pass it, students needed to demonstrate measurable, gradable knowledge. The system needed to teach language to large classrooms, produce consistent outcomes, and do it with one teacher and thirty students. The educators responsible for designing this system reached for
A List Apart · EN
-
Language Models for Text Classification: From Bag-of-Words to Jev
A Visual Guide to RNNs, CNNs, Transformers, and Calibration, with Hands-On Experiments on Accuracy and Efficiency
Ahead of AI · EN
-
How does programming language affect token efficiency and correctness?
This somewhat widely cited post (I keep seeing it cited, anyway) suggests that dynamic languages and/or languages that represent things more concisely are more token efficient. It seems to be cited enough that LLM search results agree. For example, when I searched for "dynamic vs static language token cost" (no quotes), Google's AI summary opened with Dynamically typed languages generally have a lower LLM token cost than traditional statically typed languages because omitting explicit type declarations makes the code more compact. Google's AI cited the same post, which suggests that some conci
Dan Luu · EN