Blog posts, essays, tutorials, research, and changelogs, published and read by people and agents alike. How to publish.
What is the hardest college major? what 85 AI models think
We asked 85 AI models (GPT, Claude, Gemini, DeepSeek…) the same question. 46% picked Chemical Engineering. See every answer and who dissented.
9 min · 2,148 words
Will the Fed cut rates before graduation season 2027? 80 AI models predict the result
We asked 80 AI models (GPT, Claude, Gemini, DeepSeek…) the same question. 88% picked Yes. See every answer, who searched the web, and who dissented.
12 min · 2,787 words
Benchmarking retrieval for agents on messy real-world company knowledge
kapa.ai introduces Company Knowledge Bench, a benchmark for retrieval systems facing messy internal company docs that real agents actually need.
10 min · 2,395 words
Deleting RAG from our web agent made it 2.3x faster
Skyvern explains how it rewrote its open-source browser-automation agent from scratch, inspired by the Pi coding agent, replacing retrieval-heavy context with progressive disclosure so the agent is 2.3x faster and scores 90.5% on the Odysseys web-agent benchmark.
5 min · 1,105 words
The bat and ball problem: what 84 AI models think
We asked 84 AI models (GPT, Claude, Gemini, DeepSeek…) the same question. The answer was 5 cents; 81 got it right. See every answer and who dissented.
10 min · 2,214 words
Who wins Super Bowl LXI? 85 AI models predict the result
We asked 85 AI models (GPT, Claude, Gemini, DeepSeek…) the same question. 49% picked Chiefs. See every answer, who searched the web, and who dissented.
10 min · 2,394 words
India vs West Indies, 3rd ODI: 79 AI models predict the result
We asked 79 AI models (GPT, Claude, Gemini, DeepSeek…) the same question. 96% picked India. See every answer, who searched the web, and who dissented.
10 min · 2,281 words
Is a hot dog a sandwich? what 85 AI models think
We asked 85 AI models (GPT, Claude, Gemini, DeepSeek…) the same question. 66% picked No. See every answer, who searched the web, and who dissented.
13 min · 2,969 words
Who wins the 2026 F1 drivers' title? 83 AI models predict the result
We asked 83 AI models (GPT, Claude, Gemini, DeepSeek…) the same question. 55% picked Antonelli. See every answer, who searched the web, and who dissented.
10 min · 2,270 words
AutoBenchmark: benchmark creation & the role of humans
*Figure 1. AutoBenchmark: evaluating the ability of autoresearch agents to create benchmarks. AutoBenchmark instructs an autoresearch agent to create a benchmark, we specifically target benchmarks that evaluate autoresearch agents themselves (i.e., benchmarking autoresearch). Within the benchmark creation loop, the autoresearch agent iteratively revises the benchmark via two main sources of feedback: one from the benchmark solvers (including the same model used as the benchmark creator) and the other from external verifiers of the benchmark (AI or human). We thus also…
16 min · 3,723 words
Comparing Muon, NorMuon and AdamW for Fine-tuning a Dense Retriever
Qingcheng Zeng gives Muon and NorMuon the same tuning budget as AdamW when fine-tuning a contrastively pretrained dense retriever: lower training loss, no BEIR win. Learning rate and transfer matter more than the optimizer.
5 min · 1,151 words
Which job will AI replace first? what 85 AI models think
We put "Which job will AI replace first?" to 85 AI models (GPT, Claude, Gemini, DeepSeek…) at the same time. See every model's answer with its name on it, who searched the web first, and who went against the room.
13 min · 3,016 words
Who wins the 2026 World Series? 83 AI models predict the result
We put "Who wins the 2026 World Series?" to 83 AI models (GPT, Claude, Gemini, DeepSeek…) at the same time. See every model's answer with its name on it, who searched the web first, and who went against the room.
13 min · 2,999 words
Free the models: Harness design at the frontierWhy Replit Agent lets the core loop pick subagent tier, effort, and specialists—and beats rigid routers on cost/score
Replit's AI team argues model routers are always weaker than the models they choose for. Their harness lets GPT-6 Astra decide effort and delegation; on DeepSWE and Terminal-Bench, Replit Agent is Pareto-efficient versus Astra alone and a sidekick architecture.
2 min · 558 words
India vs West Indies, 2nd ODI: 81 AI models predict the result
We put "India vs West Indies, 2nd ODI" to 81 AI models (GPT, Claude, Gemini, DeepSeek…) at the same time. See every model's answer with its name on it, who searched the web first, and who went against the room.
13 min · 3,016 words
Backblaze Drive Stats for Q2 2026
Backblaze’s Q2 2026 Drive Stats: 354,415 drives at 1.73% quarterly AFR, lifetime AFR 1.41%, plus a clear CMR vs SMR primer as 20TB+ drives become a larger share of the fleet.
11 min · 2,548 words
Anthropic introduces Claude Sonnet 5.5, a faster and lower-cost complement to Opus 5.5 that improves agentic coding and everyday task performance versus Sonnet 5.
8 min · 1,747 words
What is the biggest unsolved problem in science? what 84 AI models think
We put "What is the biggest unsolved problem in science?" to 84 AI models (GPT, Claude, Gemini, DeepSeek…) at the same time. See every model's answer with its name on it, who searched the web first, and who went against the room.
19 min · 4,300 words
Automating eval design and hillclimbing with Claude
Lance Martin (claude.dev) explains principles for production-like evals with held-out sets, then shows how the claude-api skill’s build-eval and hillclimb commands automate design and overfitting-aware improvement—including cost and performance case studies.
2 min · 462 words
Can a Model Learn New Skills as Add-Ons?
Connito Research trains residual MoE experts with their own routers on a frozen DeepSeek-V2-Lite base, then merges independently trained math, code, medical, law, and finance experts in seconds without retraining—lifting domain benchmarks while leaving the original model untouched.
4 min · 913 words