AIE Miami Keynote & Talks ft. OpenCode. Google Deepmind, OpenAI, and more!

2026-04-20 · Source: AI Engineer · Field: Technology & Digital — Artificial Intelligence & Machine Learning, Software Development & Engineering, Robotics & Autonomous Systems · Depth: Advanced, extended

Summary

The AI Engineer Miami conference featured several speakers discussing the evolving landscape of AI, particularly focusing on AI agents and their impact on software development. Key presentations included Gabe Greenberg's announcement of Orchestrator AI, a multi-agent orchestration platform demonstrating significant performance gains over single-agent systems in complex coding tasks. Dax Rods emphasized the importance of "product restraint" in an era of rapid AI-driven development, arguing that increased speed often leads to product bloat and a proliferation of bad ideas. Dex discussed the evolution of RPI (Recursive Prompt Improvement) methodologies, advocating for reading code over plans and smarter token usage through structured workflows. Shashank Goyal of Open Router highlighted market trends, noting a 14x growth in token consumption, cost collapse in older models, and agents becoming the primary workload, with 15% of spend and 40% of workflows being agentic. Nana Andukquay presented a framework for embedding AI code quality gates into the software development lifecycle, stressing early and consistent quality enforcement. Jeff discussed the changing economics of software development, asserting that it now costs less than minimum wage due to AI, and emphasized the need for engineers to adapt and understand AI fundamentals. Philip Kylie demystified quantization, explaining how techniques like NVFP4 can significantly boost inference speed without compromising quality. Finally, Google DeepMind's Alisa Forton and Guiam showcased generative media models (VO, Nano Banana, LIA, Gemini 3.1 Flash) and practical tips for effective prompting and cost-saving. Ben Davis concluded by discussing the evolution of AI SDKs, highlighting the shift from API wrappers to full coding agent SDKs like PI and OpenCode, which enable more complex, script-based agentic workflows.

Key takeaway

For AI Engineers and product managers navigating the rapid evolution of AI-driven development, prioritize architectural foresight and strategic restraint. While AI agents offer unprecedented speed, focus on building robust, governed workflows that integrate quality checks early and often, rather than simply maximizing token output. Your ability to design systems that leverage AI's strengths while mitigating its non-deterministic nature, especially in sensitive production environments, will be crucial for delivering sustainable value and avoiding technical debt.

Key insights

AI agents are rapidly transforming software development, demanding new approaches to quality, efficiency, and architectural design.

Principles

• Match agent authority to workflow risk.
• Define and codify code quality standards.
• Automate everyday tasks for significant value.

Method

Implement a verification layer by defining quality standards, centralizing context, and designing workflows that operationalize these standards early and often, leveraging agent skills for planning, code generation, and robust code review.

In practice

• Use blockwise quantization (e.g., NVFP4) for 30-50% inference speed gains.
• Structure prompts with JSON for consistent generative media outputs.
• Build agents with Langraph for stateful, multi-turn operational workflows.

Topics

• AI Agents
• LLM Context Management
• Model Quantization
• Secure Agent Deployment
• Generative Media

Best for: CTO, VP of Engineering/Data, Director of AI/ML, AI Engineer, Machine Learning Engineer, Software Engineer

Related on AIssential

• AIE Europe Day 2: ft Google Deepmind, Anthropic, Cursor, Factory, Linear, HF, Cerebras & more — AI engineering is rapidly evolving, demanding new approaches to agent orchestration, code quality, and human-AI collaboration.
• Open Source Coding Agents: The Future of AI-Driven Software Development — Open source coding agents enable autonomous, scalable software development by orchestrating multi-agent workflows and offering model flexibility.
• 4 years of observing and living through the AI era — Generative and agentic AI tools offer unprecedented automation and creation capabilities, shifting the focus to human validation and decision-making.
• Agents should interview you — AI agents can significantly accelerate development by interviewing users for requirements and automating code generation and testing.
• Issue #122 - The 12-Step Blueprint for Building an AI Agent. Part I — Effective AI agent development prioritizes systems engineering over prompt engineering for production readiness.
• Agentic Workflows Have Changed EVERYTHING in 2026 (DEATH Of The Senior Dev?) — Agentic AI workflows have fundamentally reshaped software development by 2026, making traditional coding skills obsolete while elevating engineering principles.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by AI Engineer.