LMArena, a standard platform ranking the performance of large language model AIs (LLMs), gives preference to models made by giant tech companies, according to a study preprinted on arXiv and not yet peer-reviewed.
Category: TRENDS IN AI – Jun 17 2025
Post
RESEARCHERS RELEASE THE KRAKENCODER
Researchers at Weill Cornell Medicine in New York City have unveiled the Krakencoder, a new AI that weaves together a variety of brain images taken by different devices. The result is a deeper, more integrated view of how a brain affects behavior, thoughts, and how it recovers from injuries.
Post
LEADING AI MODELS “COLLAPSE” UNDER HARD PROBLEMS, STUDY SAYS
When struggling with highly complex problems, “reasoning” AIs from DeepSeek and OpenAI suffer “complete accuracy collapse,” according to a new study released by Apple.
Post
META’S ZUCKERBERG HAND-PICKS AGI SUPERGROUP
Meta CEO Mark Zuckerberg is personally recruiting a team of as many as 50 specialists to speed the company’s progress toward creating artificial general intelligence (AGI), a supercharged synthetic intelligence that can outperform humans across a wide range of tasks.
Post
COAXING AI TO BE SCIENTIFICALLY CREATIVE
Large language model AIs such as ChatGPT or Gemini are good at carrying out rote tasks and copying patterns—writing poems, making pictures—but the folks at Lila Sciences want their AI to think outside the test tube, be wildly original, and make jaw-dropping breakthroughs.