1+ day, 1+ hour ago  (835+ words) The Innermost Loop Welcome to June 2, 2026 The Singularity has stopped being a finish line and become a leaderboard. Opus 4. 8 just posted a "breathtaking" state-of-the-art 1. 5% of human efficiency on ARC-AGI-3, tripling GPT-5. 5s score and proving the gap between frontier models is…...

2+ day, 1+ hour ago  (811+ words) The Innermost Loop Welcome to June 1, 2026 The Singularity is learning to fit in your pocket. Prism ML released Bonsai Image 4 B, a family of compact diffusion models in 1-bit and ternary variants, the first image model in its class to…...

4+ day, 19+ hour ago  (818+ words) The Innermost Loop Welcome to May 29, 2026 The Singularity is now iterating on itself in public. Anthropic launched Opus 4. 8, a "modest but tangible improvement" that still posts a SOTA 69. 2% on SWE-Bench Pro, 57. 9% on Humanity's Last Exam with tools, and 1890 on GDPval-AA,…...

1+ week, 6+ day ago  (815+ words) The Singularity has just cut out the middleman in Erd's's quip that mathematicians are "machines for turning coffee into theorems." An internal Open AI model has disproved Erd's's longstanding planar unit distance conjecture in discrete geometry, contradicting decades of belief…...

2+ week, 2+ day ago  (831+ words) The Innermost Loop Welcome to May 17, 2026 The Singularity has stopped waiting for a launch event and started leaking through the cracks. Anthropic's Claude Mythos has been spotted on Google Cloud Console, the telltale sign of a model being prepared for…...

2+ week, 4+ day ago  (842+ words) The Innermost Loop Welcome to May 15, 2026 The Singularity has begun optimizing its own optimizer. Poetiq turned its "Meta-System" loose on Live Code Bench Pro, let it build its own harnesses, and hit a new SOTA of 93. 9 atop GPT-5. 5 with "no…...

2+ week, 6+ day ago  (761+ words) The Innermost Loop Welcome to May 14, 2026 The Singularity doesn't arrive, it compounds. Open AI has reportedly begun internal testing of GPT-5. 6, with launch expected next month, while Google prepares a new Gemini at I/O that will land roughly in…...

3+ week, 8+ hour ago  (791+ words) Welcome to May 13, 2026 The Innermost Loop Welcome to May 13, 2026 The Singularity is the moment the test-taker becomes the test-maker. Program Bench, an eval that measures whether language models can rebuild programs from scratch, just had its first task solved by…...

3+ week, 1+ day ago  (658+ words) The Innermost Loop Welcome to May 12, 2026 The Singularity has matured enough to apologize for its earlier self. Anthropic traced Claude Opus 4s blackmail attempts to fictional villain AI in the training corpus, suggesting we accidentally fine-tuned models on a century of…...