Interconnects
What ChatGPT, Claude, Gemini & Grok actually say · July 2026
The verdict
Interconnects appears in 1 AI-ranked category — best position #3 for ai newsletter.
Exceptional firsthand analysis of frontier-model training, post-training, evaluations, and open models from Nathan Lambert; nearly tied with Latent Space and stronger for research-oriented practitioners.
Gemini Offers unmatched insider insights and critical analysis of model alignment, RLHF, open-weights evaluation, and post-training dynamics written by a frontier machine learning researcher.
Claude The sharpest ongoing analysis of open models, RLHF/post-training, and lab dynamics — Lambert's Ai2 vantage point yields technically grounded takes you can't get from aggregators, and it earned outsized influence during the open-weights surge; effectively tied with Import AI, ranked below only for narrower scope.
Where Interconnects falls short, per the models
- GPT Some of the best material is paid, and its technical, open-model-heavy focus is too specialized for general AI users.
- Claude Deliberately opinionated and centered on model training and open-vs-closed dynamics — thin coverage of applications, tooling, and anything outside the model layer, with the best essays behind a paid tier.
- Gemini Highly specialized in post-training and model evaluations, making it less useful for developers seeking general software application design or simple API wrapper tutorials.
Top alternatives per the models: Latent Space · The Batch · Ahead of AI · Import AI
Rankings are computed from what the models answer, re-polled continuously · raw reasoning shown verbatim · methodology