Hey, anything for a buck. BTW, Larry Darryl and Darryl would make a better name for an AI.
· List topics · post a topic · subscribe · Google ·
Posted on 05/26/2026 8:29:32 AM PDT by Twotone
Google Deepmind's new framework AlphaProof Nexus has autonomously solved nine out of 353 open Erdős problems it attempted, including two questions that had gone unanswered for 56 years.
The system also proved 44 out of 492 open conjectures from the Online Encyclopedia of Integer Sequences (OEIS), settled a 15-year-old question about Hilbert functions in algebraic geometry, and improved a known bound in convex optimization. Inference costs ran just a few hundred dollars per problem, according to the research paper.
Unlike (potentially) pure natural-language approaches such as OpenAI's recent solution, the underlying language model in AlphaProof Nexus—in this case Gemini 3.1 Pro—doesn't have to carry the entire logical chain on its own.
Instead, it generates proof steps in Lean's formal language, and the compiler checks each one. Error messages feed directly back into the next attempt. That way, the LLM gets grounded by symbolic feedback, a safety net that offsets the well-known weaknesses of language models when it comes to logical reasoning. Humans only step in at the very end to check the results.
The system consists of four agent variants with increasing complexity. The simplest, Agent (A), deploys independent sub-agents running on Gemini 3.1 Pro in loops: the language model generates proof steps, the Lean compiler checks them, and error messages feed back into the next try.
Agent (B) adds queries to AlphaProof, Google's reinforcement-learning-based system for olympiad math, which can fill in missing proof segments. Agent (C) introduces an evolutionary component. Inspired by AlphaEvolve, sub-agents share a common population of proof sketches. Rating agents built on Gemini 3.0 Flash score these sketches for plausibility and and novelty, then rank them using an Elo system. The fully equipped Agent (D) combines all of these capabilities.
(Excerpt) Read more at the-decoder.com ...
Dear FRiends,
We need your continuing support to keep FR funded. Your donations are our sole source of funding. No sugar daddies, no advertisers, no paid memberships, no commercial sales, no gimmicks, no tax subsidies. No spam, no pop-ups, no ad trackers.
If you enjoy using FR and agree it's a worthwhile endeavor, please consider making a contribution today:
Click here: to donate by Credit Card
Or here: to donate by PayPal
Or by mail to: Free Republic, LLC - PO Box 9771 - Fresno, CA 93794
Thank you very much and God bless you,
Jim
Will solve unsolvable math questions for power plants!
Did they try problems we already know the answer to first ? LOL
I asked Claude:
How many licks does it take to get to the center of a Tootsie Pop?
Claude responded:
The world may never know! 🍭
But science has actually tried to answer this. Various studies have come up with different numbers:
University of Michigan (a licking machine): ~411 licks
New York University: ~364 licks
Purdue University (both human and machine tests): 92–411 licks depending on the method
University of Michigan students (human lickers): ~130 licks on average
The wide variation comes down to factors like tongue size, saliva production, licking pressure, and whether you cheat and bite it (we all do).
The famous 1970 Tootsie Roll commercial with Mr. Owl — who bit at 3 — is probably the most culturally iconic non-answer to the question. It’s one of the longest-running ad campaigns in history precisely because the mystery is more fun than the answer.
Myth Busters should make a come back to check out AI ,LOL
Hey, anything for a buck. BTW, Larry Darryl and Darryl would make a better name for an AI.
There is a book out called something like “solve everything” that Ai will solve all math and then all of the natural sciences within a few short years. At that point we will be ready to solve the unknown unknowns. Basically we will be turning all the science fiction books we have ever read when we were young into science fact.
Disclaimer: Opinions posted on Free Republic are those of the individual posters and do not necessarily represent the opinion of Free Republic or its management. All materials posted herein are protected by copyright law and the exemption for fair use of copyrighted works.