AI Comparisons
Every major battle playing out in AI — model vs model, chip vs chip, empire vs empire.23 live · 0 in the works
Model Wars
9 live / 9 totalHead-to-head on the AI products and companies shaping the race.
OpenAI vs Anthropic
Revenue, users, valuation, developer share, and coding tools. The two poles of the same magnet.
OpenAI vs Google
ChatGPT vs Gemini. Search disruption, enterprise adoption, and who is actually winning the consumer.
Claude vs GPT
Street smart vs book smart. Head-to-head on benchmarks, coding, writing, and when to use which.
Open Source vs Closed
The gap is closing. Llama 4, Qwen 3.5, DeepSeek V3 vs GPT, Claude, Gemini — cost, control, and capability.
Price Per 1M Tokens
GPT-4o, Claude, Gemini, Llama, Grok — what intelligence actually costs. Input vs output, frontier vs cheap.
Model Ecosystem
Who is shipping models, how many, and how the landscape shifts week over week. Tracked via OpenRouter.
Model Skills Breakdown
18 models rated across coding, reasoning, vision, writing, context, and speed. Radar charts, heatmap, and side-by-side compare.
God Mode Benchmark
The benchmark that doesn't lie to you. 16 real builder tasks across Execution, Operator, Writing, and Creative. Full test suite public.
The AI Disruption Tracker
Every major Anthropic announcement in 2026 and the market damage it caused. $1T wiped from software stocks and counting.
The Compute Race
7 live / 7 totalWho owns the hardware layer — GPUs, TPUs, custom silicon, and the power behind the models.
Who Owns the GPUs
Microsoft, Google, Meta, xAI, Amazon — the GPU arms race. Who has the most compute and what they paid for it.
NVIDIA vs Custom Silicon
Google TPUs, AWS Trainium, Meta MTIA, Tesla Dojo. Can anyone dethrone NVIDIA? The CUDA moat explained.
Cloud vs Owned Infrastructure
Rent or build. The break-even is ~$500K/yr API spend. Everything above that, owned infra starts winning.
AI in Your Pocket
What compute is actually possible on a phone right now. How the A19 Pro, Snapdragon 8 Elite, and Dimensity 9400 compare — and when your phone stops needing the cloud.
AI Agents Need Wallets
Agents are autonomous economic actors. Stripe requires KYC. A $0.001 API call loses 30,000% to card fees. Here is why crypto payment rails — x402, AP2, USDC on Base — are infrastructure, not speculation.
Run AI at Home
Mac Mini to Mac Studio, Raspberry Pi to Jetson Orin. Every device you can actually own to run models, host OpenClaw, and stop paying per token.
Mac Mini Price History
Every Mac Mini generation since Apple Silicon tracked monthly — M1 to M4 Pro. Base MSRP, specs, and which one to buy for running local AI models.
The Elon Stack
3 live / 3 totalXAI, Tesla, SpaceX — one empire, many moving parts. How it fits together and what it means for AI.
xAI vs OpenAI
Grok vs GPT. Truth-seeking north star vs move fast and ship. The company Elon left vs the company he built.
The Elon AI Empire
xAI, X, Tesla, SpaceX, Neuralink — five companies, three data flywheels, one long game. How it all connects.
Tesla Dojo vs NVIDIA
Dojo is not a NVIDIA competitor. It's a vertical integration play for FSD. Here's what that actually means.
Signal vs Noise
4 live / 4 totalThe meta layer — tracking AI itself. Where it is growing, how fast, and what it is displacing.
Decentralised AI
What is real, what is narrative. Decentralised compute (Akash, io.net) vs token-incentivised models (Bittensor) vs open weights (Llama, DeepSeek). And why Meta did more to decentralise AI with one GitHub repo than all crypto networks combined.
The Gap — AI by Industry
What AI could theoretically do vs what it's actually being used for, across every major occupational sector. The distance between those two lines is the next decade of AI value.
OpenClaw Stats
0 to 295K GitHub stars in 107 days. Star growth, fork trajectory, and the DeepSeek spike that changed everything.
AI vs Real Content
What percentage of social media content is now AI-generated — by platform, by content type, and growing fast.
Request a
Comparison
Which AI battle should we break down next? Drop your suggestion and email — we read every one and build the most-requested first.