This is not a blanket defense of using LLMs in education, they have to be used properly. We know from this well-powered RCT that just having the AI give you answers lowers test scores. https://
papers.ssrn.com/sol3/papers.cf
m?abstract_id=4895486
…
RESEARCH
-
LLMs in Education: RCT Study Shows Answer-Giving Lowers Test Scores
By
–
-
LLM Users Show Cognitive Decline Study Methodology Questioned
By
–
This line from the abstract is very misleading: “Over four months, LLM users consistently underperformed at neural, linguistic, and behavioral levels.” It does not test LLM users over 4 months, it tests people who had an LLM help write an essay about that essay 4 months later.
-
LLM Use in Education: Engagement and Learning, Not Brain Rot
By
–
This study is being massively misinterpreted. College students who wrote an essay with LLM help engaged less with the essay & thus were less engaged when (a total of 9 people) were asked to do similar work weeks later. LLMs do not rot your brain. Being lazy & not learning does.
-

Chinese Version Uses Different Task Counting Character Occurrences
By
–
The Chinese version is doing a fairly different task. It's using the character for "ha" (as in the English "haha" laughing sound) and asking how many times that appears in a sentence. Chinese doesn't really have many-character words, so the task doesn't quite fit here.
-
Real-time Generative Movies: Exciting Milestones Ahead
By
–
Now: 8s videos with audio in about 3 minutes
.
.
.
.
Future: real time generative movies, episodes, lectures; they generate as you watch them, with steerable mechanisms during viewing. They are captivating and worth watching. What are the exciting intermediate milestones? -
Carbon Atoms as Patterns: Semantics Through Relational Change
By
–
A carbon atom is a pattern, its semantics are defined by how its presence changes other patterns.
-
Should humanity have evolved? Ethical implications explored
By
–
Do you think it would have been better if humans had not evolved?
-

ChatGPT o3 solves task requiring web search and multi-hop deduction in 2m54s
By
–
* = the answer to the next question A task meant to require web and computational search guided by multi-hop deduction starting from a partial SHA1, answered in 2m 54s by ChatGPT o3 featuring Tom Cruise and Addison Rae
-
NVIDIA DeepMind Disney Partner to Train Humanoid Robots in Simulation
By
–
🇫🇷À Vivatech, Jensen Huang a annoncé un partenariat entre NVIDIA, DeepMind et Disney.
— VISION IA (@vision_ia) 19 juin 2025
Objectif : créer la simulation physique la plus réaliste du monde… pour entraîner des robots humanoïdes.
L’IA apprend désormais dans des mondes virtuels soumis aux lois de la physique.
C’est… pic.twitter.com/o5aUkbb7prAt VivaTech, Jensen Huang announced a partnership between NVIDIA, DeepMind, and Disney. Objective: to create the world's most realistic physical simulation… to train humanoid robots. AI now learns in virtual worlds subject to the laws of physics. That's where it truly
-
Computers and Brains as Mathematical Computing Systems
By
–
Computers don't imitate brains at all. The computer is a very general mathematical idea: a system that can execute functions via arbitrary sequences of state transitions. Both brains and GPUs are in this category.