TikTok Served Nazi Propaganda, Jan. 6 Committee Found

The platform has largely escaped notice in the public battles over content moderation, but it’s not immune from hateful extremism

Why Not Mars (Idle Words)

The Platform Accountability and Transparency Act, Take Two

New legislation would give researchers in the U.S. access to study data from the largest social media companies, writes John Perrino.

AI Artwork Will Change Everything: Peter Mohrbacher Shares Thoughts on Midjourney

What do professional artists think of AI art? Peter Mohrbacher, the word-renowned artist, shared his thoughts about AI Artwork and the platform Midjourney. Below are highlights of Peter’s thoughts…

My Reflections on the Release of the January 6 Committee Report on Trump's Attempted Election Subversion and the Expected Passage This Week of Electoral Count Act Reform: Gratitude, Awe, and Partial Relief #ELB

We are just a few weeks shy of the two-year anniversary of the January 6, 2021 insurrection in the United States Capitol, the culmination of a series of events engineered by Donald Trump and his allies to disrupt the peaceful … Continue reading My Reflections on the Release of the January 6 Committee Report on Trump’s Attempted Election Subversion and the Expected Passage This Week of Electoral Count Act Reform: Gratitude, Awe, and Partial Relief →

Madison Square Garden Uses Facial Recognition to Ban Its Owner’s Enemies

MSG Entertainment, the owner of the arena and Radio City Music Hall, has put lawyers who represent people suing it on an “exclusion list” to keep them out of concerts and sporting events.

Language models are better than humans at next-token prediction

Current language models are considered to have sub-human capabilities at natural language tasks like question-answering or writing code. However, language models are not trained to perform well at these tasks, they are trained to accurately predict the next token given previous tokes in tokenized text. It is not clear whether language models are better or worse than humans at next token prediction. To try to answer this question, we performed two distinct experiments to directly compare humans and language models on this front: one measuring top-1 accuracy and the other measuring perplexity. In both experiments, we find humans to be consistently \emph{worse} than even relatively small language models like GPT3-Ada at next-token prediction.