← back
Why Agent Engineering — swyx
Original: Why Agent Engineering — swyx
Takeaway
Agent engineering is emerging as its own discipline at the intersection of MLE and software engineering, driven by cheaper intelligence, better tools, and clear PMF in coding/support.
Summary
- swyx explains AI Engineer Summit's pivot to an agent-engineering-only conference, banning vendor pitches and excluding RAG/open models/GPUs.
- Catalogs 6+ rival agent definitions from Simon Willison's crowdsourced list plus OpenAI's newly published definition.
- Why now: better reasoning + tool use, MCP, model diversity (OpenAI's share fell from 95% to 50%), 1000× drop in GPT-4-level token cost in 18 months, RL fine-tuning.
- Predicts agent + X (RAG, search, code) as the dominant 2025 business pattern; coding and support agents have clearest PMF.
- Highlights Brett Taylor's outcome-based pricing and multi-agent + faster inference as undercovered tailwinds.
agent-engineeringai-engineerindustry-trends
Original description
Swyx explains why we are all in on Agent Engineering in 2025. Slides: https://docs.google.com/presentation/d/1SWoBIvTQu__uNEvSawmNcROiUx-n86O_fP0arZcTGb8/edit?usp=sharing Recorded live at the Agent Engineering Session Day from the AI Engineer Summit 2025 in New York. Learn more at https://ai.engineer and purchase tickets to our next event, the AI Engineer World's Fair, in SF June 3 - 5 here: https://ti.to/software-3/ai-engineer-worlds-fair-2025 About swyx swyx (Shawn Wang) is Editor of Latent.Space. As a developer experience leader and angel investor, swyx led developer tooling at AWS, Two Sigma, and three devtool unicorns (Netlify, Temporal, Airbyte). He is also the founder of Smol AI, the LLM data pipeline company that produces AI News, the widely-read AI industry newsletter that was 99% created by customizable research agents.