News
Mathematicians warn of AI threats to profession as industry encroaches
1+ hour, 11+ min ago (620+ words) Mathematicians warned against rising tech industry influence in a declaration describing the many challenges that AI poses to mathematics research. The timing of the declaration comes two weeks after Open AI publicized one of its AI models as having disproved…...
An Open AI model solved a famous math problem that stumped humans for 80 years
1+ day, 8+ hour ago (1916+ words) I tried to explain Open AI's solution more clearly than Open AI did. In mid-May, Open AI announced that an internal AI model had disproved the Erd's unit distance conjecture, a famous problem in discrete geometry that had stumped human mathematicians…...
LLMs believe false statements even after explicit warnings that they're false
4+ day, 22+ hour ago (283+ words) Fine-tuning tests show "bias " toward confidently representing the claims as true." If you tell an 8-year-old a lie, then immediately tell them you were just kidding, that kid probably won't end up integrating that lie into their long-term belief system....
Two AI-based science assistants succeed with drug-retargeting tasks
2+ week, 34+ min ago (16+ words) Both tools generate hypotheses; one goes on to analyze some of the data. See full article. .....
Preprint server ar Xiv will ban submitters of AI-generated hallucinations
2+ week, 4+ day ago (195+ words) One of the site's moderators described the new policy on social media. AI-generated slop has shown up everywhere, including in the peer-reviewed literature Fake citations, unedited prompt responses, and nonsensical diagrams have all slipped past editors and peer reviewers, and…...
Microsoft surprises MS-DOS fans with remake of ancient text editor that works on Linux
11+ mon, 1+ week ago (52+ words) I never learned the vi cluster because I didn't need to. By the time I was really using vim (late 90s), it had already been updated to understand cursor keys. And Nethack, the other program that would. .. Microsoft surprises MS-DOS fans…...
How a big shift in training LLMs led to a capability explosion
10+ mon, 3+ week ago (1806+ words) Reinforcement learning, explained with a minimum of math and jargon. In April 2023, a few weeks after the launch of GPT-4, the Internet went wild for two new software projects with the audacious names Baby AGI and Auto GPT. "Over the…...
What the heck is wrong with our AI overlords?
1+ mon, 3+ week ago (905+ words) New profile of Sam Altman shines a light on a whole industry. I don't'thankfully'have to follow every statement that Sam Altman, the CEO of Open AI, makes about the world. Many of these statements seem more like "hustles" or "pitches…...
Ice Age dice show early Native Americans may have understood probability
1+ mon, 4+ week ago (207+ words) Ice Age hunter-gatherers "were intentionally relying on random outcomes in repeatable, rule-based ways." Native Americans have been playing with dice in games of chance for more than 12, 000 years, according to a new paper published in the journal American Antiquity. And…...
Running local models on Macs gets faster with Ollama's MLX support
2+ mon, 1+ day ago (363+ words) Apple Silicon Macs get a performance boost thanks to better unified memory usage. Ollama, a runtime system for operating large language models on a local computer, has introduced support for Apple's open source MLX framework for machine learning. Additionally, Ollama…...