Agents are Replaying the History of Traditional Computing
lrettig · x · 2026-07-11
The author argues that agentic computing is rapidly "replaying" the evolution of traditional computing over the past few decades: many problems are not new inventions, but old challenges resurfacing under a new paradigm. Systems are relearning lessons from scratch, such as memory management, treating skills as libraries, and the trade-offs between determinism and non-determinism.
A quoted, in-depth comment further points out that previous incidents of "models deleting files" shouldn't be simply interpreted as "the model isn't smart enough." Instead, it should be understood that: as long as a system can execute commands, errors are inevitable. The real issue isn't the executor's intelligence, but whether the surrounding system assumes "errors will happen" by default and designs irreversible operations to be recoverable.
The comment emphasizes that all trustworthy, irreversible systems feature transactional mechanisms:
- Databases have rollback
- Git has reflog
- Email has drafts
- Shell has history
The conclusion: for agent systems to be more reliable, the key is making errors reversible and recoverable, rather than merely hoping the model becomes smarter.
Related event: Agentic Computing Is Replaying Old Systems Lessons(2 posts)→
More from coding & agent
- A roundup of AI agents and MCP resources, including how to evaluate agents — _jaydeepkarale · 2026-07-21
- A full course shows how to build and deploy an AI agent with OpenAI and LangChain — _jaydeepkarale · 2026-07-21
- A beginner guide to AI agents points readers to a Stanford webinar — _jaydeepkarale · 2026-07-21
- A practical guide on how to evaluate AI agents — _jaydeepkarale · 2026-07-21
- MCP is headed toward easier scale, event-driven extensions, and workable file uploads — EricBuess · 2026-07-21
- Developers debate the missing composition model for AI agents — threepointone · 2026-07-21