The future of agentic coding with Claude Code
- Source
- youtube.com
- Author
- Anthropic
- Date
Context
In an Anthropic video, Claude Code creator Boris Cherny describes coding with AI as moving from typing into an editor with autocomplete toward directing an that edits files for you. Cherny says this was not workable earlier for two reasons: the models were not good enough, and the software around them was not either. Claude Code, from Anthropic, began on an older Sonnet model and was already somewhat useful.
Cherny calls that surrounding software the . In Cherny's description it includes the , how is managed, the tools the model can call, connections to outside services and permissions. Cherny says the model and harness improved together, partly because Anthropic's researchers use Claude Code daily and notice failures such as failed edits. Cherny also says new models are judged by using them for the day's work, because building reliable evals (systematic tests) for such broad use is hard.
- AI agent — An AI system that doesn't just answer once but works toward a goal in a loop — taking actions, reading the results, and deciding what to do next.
- agent harness — The scaffolding around a model that turns it into a working agent — the loop, the tools it can call, and the rules for when to stop.
- system prompt — The standing instructions a model receives before any user input — defining its role, rules, tools, and tone for the whole conversation.
- context window — The maximum amount of text a model can consider at once — its working memory for the current conversation or task.
Checking sign-in…
Loading comments…






