Korva is a scroll of the internet’s best ideas.
AI could disclose first
Hidden patterns
The data already exists in scattered places: genomic data, medical records, optical sensing systems, and the historical record. A system correlating all of it could spot gaps in that record that make no sense and detect unknown deception operations. It might conclude that a story about some event in history is wrong, dig, and find the event was cover for a crash retrieval.
Why the timing matters
Timing matters, since artificial superintelligence could come online in the next months to years. It has been solving crazy mathematical problems lately, and data efforts like the Stargate AI medical project run by Larry Ellison with the Trump administration collect exactly the kind of genomic data that could reveal a presence. Big tech could inadvertently discover classified US technology this way, and an adversary could use the moment to look like the messianic truth-bearer.
Talking to non-humans
Silicon-based AI may also be the infrastructure humans need to communicate with organic non-human intelligence, since it can learn and figure out new modalities of communication.
How can AI find patterns people miss?
Software can read and cross-reference millions of documents that no person could read in a lifetime. In a 2019 Nature study, Vahe Tshitoyan and colleagues trained a language model on 3.3 million materials science abstracts. It predicted promising thermoelectric materials years before experiments confirmed them, using only the connections it found in text. The same approach can link records scattered across databases.
How has AI done on hard math lately?
In July 2024, Google DeepMind's AlphaProof and AlphaGeometry systems solved four of the six problems at that year's International Mathematical Olympiad, earning a score equal to a silver medal. The problems are written for the world's strongest high school students. Both systems check their proofs with a formal verification language, so the answers are machine-verified. Each system took up to several days on some problems.
Can AI learn to decode animal communication?
Project CETI, launched in 2020, uses machine learning to study the clicks that sperm whales exchange, called codas. Researchers there have recorded thousands of codas and found that patterns in rhythm and tempo carry structure. A related group, the Earth Species Project, applies similar tools to birds, primates, and other animals. Neither has shown that any animal communicates with a language like human speech.
More from The Joe Rogan Experience
