Inference: When Will We Solve Intelligence

2025-12-27 · Source: Turing Post · Field: Technology & Digital — Artificial Intelligence & Machine Learning, Mathematics & Computational Sciences, Software Development & Engineering · Depth: Advanced, extended

Summary

Turing Post's "Inference" series, launched in 2025, is evolving in 2026 to address the question "When Will We Solve Intelligence?" The series, which previously featured interviews with researchers and founders, will now adopt topic-driven interview cycles focusing on areas like robotics, open source, and coding. Each cycle will be accompanied by a "Topic Report" that synthesizes conversations, adds context, and highlights emerging patterns. The article highlights key themes from 2025 interviews, including "What Intelligence Is Missing" (reasoning failures, memory limits, hallucinations), "Coding and the Agentic Web" (agents, tools, developer workflows), "Who Is in Control" (trust, safety, human involvement), and "AI Meets the Real World" (physics, biology, institutions). Karina Hong, co-founder and CEO of Axium Math, discusses building an AI mathematician with a prover system, a knowledge base, and a conjecturer system, aiming for self-improving AI through a hybrid approach combining formal verification with large language models.

Key takeaway

For AI Scientists and Research Scientists focused on advancing mathematical reasoning, you should explore hybrid AI approaches that combine formal verification with large language models. This strategy addresses current AI limitations in complex, multi-step reasoning and offers a path toward verifiable, self-improving systems, particularly in scientific discovery and highly regulated enterprise applications.

Key insights

Solving intelligence requires a hybrid approach combining formal verification with large language models for robust reasoning.

Principles

• Mathematical discoveries have transformative power.
• AI can compress the timeline of scientific breakthroughs.
• Collaboration between humans and AI is crucial.

Method

Axium Math's AI mathematician integrates a prover for verifiable proofs, a knowledge base for new concepts, and a conjecturer for proposing new math problems, forming a self-improving loop.

In practice

• Explore Lean's "natural numbers game" for beginners.
• Formalize math theories in Lean's Mathlib.
• Consider formal verification for regulated industries.

Topics

• AI for Mathematics
• Formal Verification
• Automated Theorem Proving
• AI Research & Discovery

Best for: AI Scientist, Research Scientist, CTO, AI Researcher, AI Engineer, Data Scientist

Related on AIssential

• Artificial Intelligence for Mathematical Reasoning: An Integrated Survey of Language Models, Neuro-symbolic Systems, and Verified Discovery — AI for mathematical reasoning is a rapidly evolving frontier, integrating diverse techniques for verified discovery.
• Deepmind's research AI occasionally solves what humans can't and mostly gets everything else wrong — AI can achieve breakthroughs in specific research tasks but requires careful human oversight and structured collaboration.
• OpenAI cracks an 80-year math belief — A general-purpose AI model achieved novel mathematical discovery, signaling advanced AI capabilities for scientific breakthroughs.
• What it Means to Be a Mathematician When AI Does the Math — AI is shifting mathematics from human intuition and struggle to automated discovery and formal verification, raising existential questions.
• The AI Revolution in Math Has Arrived — AI is rapidly transforming mathematical research by accelerating discovery and proof generation, despite inherent model errors.
• Building Mathematical Superintelligence — AI is rapidly advancing towards mathematical superintelligence, leveraging formal verification for provably correct results.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by Turing Post.