Social media encourages the worst of AI boosterism
4 months ago
- #AI hype
- #mathematics
- #social media
- Social media amplifies hype around AI breakthroughs, often without proper verification.
- Example: OpenAI's GPT-5 falsely claimed to solve 10 unsolved Erdős math problems, but it only found existing solutions.
- Mathematicians like Thomas Bloom criticized the misrepresentation, highlighting the need for cautious claims.
- AI's ability to trawl through vast literature is impressive but overshadowed by exaggerated claims.
- LLMs struggle with consistency in fields like medicine and law, despite hype about their capabilities.
- AI startup AxiomProver solved open Erdős problems and excelled in the Putnam competition, sparking debates on social media.
- Judging AI's true capabilities requires deeper analysis beyond social media buzz.