NewsHive
CONTACT USANALYST PORTAL →
aiBACKGROUND

Spent 3 months on an Erdős 993 problem with Claude Code + Codex. Found out today

BACKGROUND
1 SIGNALFIRST DETECTED 5 October 2026
Narrative synthesis pending — 1 signal collected.
How the story developed
Other feeds and industry voices
6 Oct
For 31 years, mathematicians couldn’t finish this proof. The new paper says ChatGPT found most of the ideas and proofs behind a problem tied to efficient computer search and sampling, while the authors mostly worked to understand and explain it.

<table> <tr><td> <a href="https://www.reddit.com/r/singularity/comments/1wyxdtu/for_31_years_mathematicians_couldnt_finish_this/"> <img alt="For 31 years, mathematicians couldn’t

OPINION · WHO · singularityOPEN STORY
8 Sept
OpenAI LLMs has resolved a roughly 90 years old mathematical problem

<table> <tr><td> <a href="https://www.reddit.com/r/OpenAI/comments/1waylo4/openai_llms_has_resolved_a_roughly_90_years_old/"> <img alt="OpenAI LLMs has resolved a roughly 90 years

OPINION · WHAT · OpenAIOPEN STORY
14 Oct
GPT-6 Astra autonomously solves 49 out of 50 unpublished theoretical physics research problems, scoring 98% at under $3 per problem. Researchers initially marked several answers wrong, only to discover the AI had caught errors in the calculations of world-class physicists.

<table> <tr><td> <a href="https://www.reddit.com/r/singularity/comments/1x5y9ue/gpt6_astra_autonomously_solves_49_out_of_50/"> <img alt="GPT-6 Astra autonomously solves 49 out of

OPINION · WHAT · singularityOPEN STORY
28 May
Using the results of erdos unit distance problem,mathematicians disprove a long standing mathematics conjecture.

<table> <tr><td> <a href="https://www.reddit.com/r/singularity/comments/1tq57dw/using_the_results_of_erdos_unit_distance/"> <img alt="Using the results of erdos unit distance

OPINION · HOW · singularityOPEN STORY
3 Jul
I Pointed My AI Research Engine at Goldbach's Conjecture — It Found a Hidden Bias

<p>As a developer building AI for scientific discovery, I wanted to test if autonomous research actually works. So I built <a href="https://github.com/subhansh-dev" rel="noopener

REPORT · HOW · Dev.toOPEN STORY
27 Mar
RYS Part 3: LLMs think in geometry, not language — new results across 4 models, including code and math

<table> <tr><td> <a href="https://www.reddit.com/r/LocalLLaMA/comments/1s59rct/rys_part_3_llms_think_in_geometry_not_language/"> <img alt="RYS Part 3: LLMs think in geometry, not

OPINION · WHAT · LocalLLaMAOPEN STORY
16 Apr
Claude Opus 4.7 Just Dropped – Better Long Tasks, Self-Verification, 3x Better Vision .

<!-- SC_OFF --><div class="md"><p>Anthropic just pushed out <strong>Claude Opus 4.7</strong> and it’s live right now on <a href="http://claude.ai">claude.ai</a>, the Platform, and

OPINION · WHAT · ArtificialInteligenceOPEN STORY
23 May
Golfing and stylistically aligning a proof using Claude Code | Another Certified Hood Classic by Terrance Tao and Claude

<table> <tr><td> <a href="https://www.reddit.com/r/singularity/comments/1tlw3l1/golfing_and_stylistically_aligning_a_proof_using/"> <img alt="Golfing and stylistically aligning a

OPINION · HOW · singularityOPEN STORY
Sources
ArtificialInteligence

NewsHive monitors these sources continuously. All signal titles above link to the original reporting.

Intelligence by NewsHive. Need help navigating what this means for your business? Contact GeekyBee →