Google DeepMind creates super-advanced AI that can invent new algorithms
Briefly

Google's DeepMind introduces AlphaEvolve, an AI agent designed to address complex math and science problems. Based on the Gemini language models, AlphaEvolve employs an evolutionary approach, enhancing accuracy in algorithmic problem-solving. Distinct from traditional Gemini models that can hallucinate information, AlphaEvolve systematically evaluates and improves multiple proposed solutions. This general-purpose AI promises to assist researchers in various programming tasks, showcasing its potential by already yielding positive outcomes within Google's operations, marking an evolution in AI application.
AlphaEvolve represents a significant leap in AI's capacity, applying an evolutionary method to enhance algorithm handling in diverse research domains.
Unlike previous models confined to specific knowledge areas, AlphaEvolve emerges as a versatile AI agent, designed to tackle a wide array of programming challenges.
Read at Ars Technica
[
|
]