First, on magic states: In quantum information theory, magic is the resource that separates a quantum system from anything a classical computer can efficiently simulate. It's what makes quantum computation powerful — and elusive. A magic state is a quantum states that possesses magic. It enables quantum computation.
Quantum computers are arriving but we don't know what to run on them.
In 1994, Peter Shor published an algorithm that could factor large integers exponentially faster than any classical method — rendering most of the internet's cryptographic infrastructure theoretically vulnerable. It was a landmark result that demonstrated what quantum computers were capable of. Yet, thirty years on, it remains essentially the only quantum algorithm with a clear, transformative advantage that survives contact with realistic hardware and practical problem sizes.
This is the central embarrassment of quantum computing: advances in quantum algorithms have been rare and lackluster.
The progress in artificial intelligence towards solving complex mathematical problems thus presents and opportunity. Can we accelerate quantum algorithms discovery using AI tools?
A new kind of laboratory
We are building at the intersection of two converging developments. The first is fault-tolerant quantum hardware — devices capable of running deep circuits with error correction, making large-scale quantum computation a question of when, not if. The second is AI that can reason about hard scientific problems autonomously, continuously, and in parallel.
Together, these change the structure of algorithm discovery. A laboratory of powerful AI agents — each reasoning independently about quantum circuit structure, complexity-theoretic constraints, and algorithmic opportunity — can search a space no human team can cover. That is a qualitatively different research instrument than anything the field has had before.
We are building that instrument.
We are not claiming a specific algorithm today. We are not predicting timelines. We are establishing the infrastructure — computational, theoretical, and organizational — to conduct this search seriously and at scale. The goal is a portfolio of quantum algorithms with genuine, practical advantage: in optimization, simulation, cryptography, or domains not yet identified.
We are not claiming quantum supremacy in any application domain. We are not asserting that AI will trivially solve hard open problems. We are asserting something more specific: that the throughput of algorithm search is the binding constraint on quantum utility, and that AI systems are already capable enough to materially increase that throughput — and will become more so. The field is leaving significant discovery on the table by relying exclusively on human researchers working at human speed.
If you think this is wrong, we'd like to convince you otherwise. If you think it is right, and you are a researcher, investor, or institution who wants to work at this intersection, we should talk.
hello@magicstatelabs.com