The Deep Feed

I know I'm late to this one, but still wanted to talk about it. GPT-5.5 is really really good, but it's also a little weird...

Thank you Railway for sponsoring! Check them out at: https://davis7.link/railway

SOURCES https://developers.openai.com/api/docs/guides/prompt-guidance https://x.com/0xSero/status/2048744545853030690 https://x.com/davis7/status/2048239401059434710 https://x.com/davis7/status/2048235547366457452 https://x.com/davis7/status/2048232268808990890 https://x.com/threepointone/status/2049545657866674665 https://x.com/badlogicgames/status/2048444292562026713 https://models.davis7.sh/

My X, Discord, and more: https://davis7.sh/

Chat interfaces are no longer limited to walls of text. In this talk, Liad Yosef and Ido Salomon explain how MCP Apps turn tools into interactive UI inside hosts like ChatGPT, Claude, VS Code, Cursor, and Copilot, letting companies send branded, functional app experiences instead of plain text responses.

The session covers the core architecture behind MCP Apps, how UI is passed over MCP, how interactions stay in context through the host, and why this changes how applications get distributed in an agent-first world. If you're building on MCP, this is a practical look at the emerging standard for UI inside chat.

Speaker info: - Nick Nisi | https://nicknisi.com/about/ - Zach Proser | https://zackproser.com/

Chat interfaces are no longer limited to walls of text. In this talk, Liad Yosef and Ido Salomon explain how MCP Apps turn tools into interactive UI inside hosts like ChatGPT, Claude, VS Code, Cursor, and Copilot, letting companies send branded, functional app experiences instead of plain text responses.

The session covers the core architecture behind MCP Apps, how UI is passed over MCP, how interactions stay in context through the host, and why this changes how applications get distributed in an agent-first world. If you're building on MCP, this is a practical look at the emerging standard for UI inside chat.

Everyone's building multi-agent systems, but nobody agrees on how. This talk proposes a taxonomy of five frontier multi-agent strategies and shows what happens when you compose them into a single architecture. Drawing from production data at Factory, we walk through a three-role system (orchestrator, workers, validators) that uses validation contracts, structured agent handoffs, and adversarial verification. We cover the case for serial over parallel execution, why model selection per role is a compounding advantage, and how to design systems that get better with each model generation instead of being made obsolete by them.

Speaker info: - https://github.com/lukealvoeiro - https://www.linkedin.com/in/lukealvoeiro

As models get smaller and more capable, more AI workloads can move onto the device itself. In this talk, Chintan Parikh from Google DeepMind walks through what that looks like in practice, from Gemma 4 edge models and on-device agent skills to the real tradeoffs around latency, privacy, cost, and cross-platform deployment.

The session covers LiteRT, the Google AI Edge stack for running models across Android, iOS, desktop, web, and IoT, along with demos of local tool calling, structured output, reasoning, benchmarking, and hardware acceleration on CPUs, GPUs, and NPUs. If you're building on-device AI systems, this is a practical overview of the current edge stack and where it is headed.

Speaker info: - https://www.linkedin.com/in/weiyiwang1993 - https://www.linkedin.com/in/chintansparikh