Chapter 2How does it work?
The desk
The model can only "see" what fits on its desk right now. The software around it has to decide what stays.
The desk. It only "sees" what fits in its context window right now. It has no memory between conversations unless software feeds earlier material back in.1
Go deeper: training vs. use
Training
Happens before release, at the company that makes the model. It adjusts the parameters, takes enormous computing power, and produces a fixed model.45
Use
Happens every time you send a message. The parameters don't change; the model reads what's on its desk and predicts a reply.1
Your chat doesn't retrain the model on the spot. Whether your conversations are later used to train a future model depends on your account and settings.67 See Where does my data go?.
Models use information at the start and end of a long input more reliably than information buried in the middle.2
Notes (7)
- 1Anthropic, “Effective context engineering for AI agents,” Anthropic Engineering, September 29, 2025, https://www.anthropic.com/engineering/effective-context-engineering-for-ai-agents (accessed October 26, 2026).
- 2Nelson F. Liu, Kevin Lin, John Hewitt, Ashwin Paranjape, Michele Bevilacqua, Fabio Petroni, and Percy Liang, “Lost in the Middle: How Language Models Use Long Contexts,” Transactions of the Association for Computational Linguistics 12 (February 23, 2024): 157–173, https://doi.org/10.1162/tacl_a_00638.
- 3Anthropic, “Models overview,” Claude Platform Docs, https://platform.claude.com/docs/en/models/overview (accessed October 26, 2026).
- 4Jared Kaplan, Sam McCandlish, Tom Henighan, Tom B. Brown, Benjamin Chess, Rewon Child, Scott Gray, Alec Radford, Jeffrey Wu, and Dario Amodei, “Scaling Laws for Neural Language Models” (preprint, arXiv, January 23, 2020), https://doi.org/10.48550/arXiv.2001.08361.
- 5Tom B. Brown, Benjamin Mann, Nick Ryder, Melanie Subbiah, Jared Kaplan, Prafulla Dhariwal, Arvind Neelakantan, et al., “Language Models are Few-Shot Learners” (preprint, arXiv, May 28, 2020), https://doi.org/10.48550/arXiv.2005.14165.
- 6Anthropic, “Updates to our consumer terms,” August 28, 2025, https://www.anthropic.com/news/updates-to-our-consumer-terms (accessed October 26, 2026).
- 7Anthropic, “Is my data used for model training? (commercial),” Anthropic Privacy Center, August 18, 2026, https://privacy.claude.com/en/articles/7996868-is-my-data-used-for-training (accessed October 26, 2026).
Chapter 2 · page 3 of 3
Turn pages with the ← and → keys.