General-purpose Artificial Intelligence tackles open math problem

OpenAI's GPT-5 reportedly helped mathematician Ernest Ryu solve a long-standing convex optimization problem now under formal verification. The result points to a broader shift from Artificial Intelligence as a math assistant toward Artificial Intelligence as a source of original research.

Mathematician Ernest Ryu used OpenAI’s GPT-5 to tackle a 40-year-old open problem in convex optimization, producing a solution that is now undergoing formal verification. If the proof holds up, it would mark one of the clearest signs yet that general-purpose Artificial Intelligence can contribute original mathematical discoveries rather than only assist human researchers. The significance comes not only from the reported result, but from the kind of system involved. GPT-5 is described as a general-purpose model rather than a narrowly designed mathematics engine.

Recent results described alongside Ryu’s work suggest that mathematical performance by advanced models is moving beyond competitions and into publishable research. OpenAI’s internal reasoning models have produced over 10 new solutions to Erdős-style combinatorics problems, with some reportedly being considered for publication in top journals. At the 2025 International Mathematical Olympiad, models from both OpenAI and Google DeepMind achieved gold-medal performance by solving 5 of 6 problems under standard competition constraints. Google DeepMind’s AlphaEvolve also showed research strength: when tested against 67 research-level mathematical problems, it improved upon the best known results in 23 of them.

Convex optimization has practical importance well beyond pure mathematics. Machine learning training algorithms, financial portfolio construction, signal processing, and logistics planning all rely on it. A 40-year-old open problem in this field is presented as a meaningful constraint on what is computationally practical, not an isolated theoretical curiosity. That makes the reported breakthrough notable both for mathematics and for the many technical domains built on optimization methods.

The report argues that the broader implication is a changing relationship between mathematicians and machine systems. Earlier mathematical successes from companies such as Google DeepMind often relied on systems specifically architected or fine-tuned for mathematical reasoning, while GPT-5 is framed as a general-purpose model able to handle a wide range of tasks. Interest is also growing in autoformalization, the process of translating informal mathematical reasoning into formal, machine-verifiable proofs. That trend could make it easier to turn mathematical intuition into rigorous verified results at far greater speed.

The implications extend into cryptography and digital infrastructure. Cryptography, zero-knowledge proof systems, consensus mechanisms, smart contract auditing, and protocol correctness all depend on hard mathematical assumptions and formal verification. The report does not claim current systems are breaking mainstream cryptography, noting that the problems being solved now are in optimization and combinatorics rather than the number-theoretic foundations of most cryptographic systems. Even so, stronger Artificial Intelligence reasoning in mathematics could accelerate new cryptographic research, improve verification tools, and reshape expectations around protocol security.

68

Impact Score

Suno faces lawsuit from The American Dollar

Suno has been sued by production duo The American Dollar, who allege its generative Artificial Intelligence music service has damaged their sync licensing business. The case could sharpen the fair use fight around whether Artificial Intelligence training causes direct market harm.

Colorado Artificial Intelligence bias law faces federal challenge

The US Department of Justice joined xAI in challenging Colorado’s law on discrimination in high-risk Artificial Intelligence systems, casting consumer protections as ideological overreach. Critics argue the attack weakens accountability for hiring, housing, and healthcare tools that can produce discriminatory outcomes.

Uk debate grows over cross-sector Artificial Intelligence regulation

The UK continues to regulate Artificial Intelligence mainly through existing sector-specific laws and non-statutory principles, while ministers have stepped back from a single overarching bill. Supporters of cross-sector legislation see a need for stronger guardrails, while critics warn broad rules could weigh on innovation.

AMD cuts Ryzen Artificial Intelligence LLM startup time

AMD detailed a two-phase initialization method for on-device large language model inference on Ryzen Artificial Intelligence processors. The approach separates model reading from NPU device setup to reduce cache thrashing and speed startup without affecting correctness.

Microsoft adds FIDES security to Agent Framework

Microsoft has released FIDES in Agent Framework to block prompt injection and data exfiltration with deterministic policy enforcement. The feature labels content by trust and confidentiality, then checks tool calls before sensitive actions can run.

Contact Us

Got questions? Use the form to contact us.

Contact Form

Clicking next sends a verification code to your email. After verifying, you can enter your message.