Your Agent Thinks Too Much. 🤖🍗 | Crazy Thursday @ SF Tech Week

Search, Decisions, Memory, Tools & Runtime for the Next Generation of Agents

About the Event

Agents are taking on longer tasks, working across apps, and returning to projects over time. Every step depends on the stack around them: the evidence they retrieve, the decisions they make, the context they remember, and the systems that let them act.

TypeSafe AI’s recent launch of Jev, a “System One” model designed for fast, structured decisions, adds a timely question to the conversation: how should specialized decision models fit into an agent stack alongside reasoning models, search, memory, tools, and runtime?

Join SEAMATE × CORESPEED in downtown San Francisco for a founder and builder mixer on the infrastructure behind the next generation of agents. Through a keynote, short partner talks, a panel, and live demos, we'll explore the choices that shape useful, reliable agent products—and where the next opportunities are emerging.

Come to compare architectures, share lessons from what you're building, and meet potential collaborators across the agent ecosystem.

What We'll Explore

• Search: How should agents retrieve fresh, relevant evidence from the web and company data? Where do retrieval quality, source verification, and context engineering change the outcome?

• Decisions: How should an agent plan, choose tools, route work, and know when to ask a human? When does adding more agents help?

• Memory: What should persist across sessions, how should it be updated, and how do teams prevent stale context from becoming a recurring mistake?

• Tools: How do MCP and APIs connect agents to real workflows? What do permissions, identity, and reliable tool execution need to look like?

• Runtime: Where should agents run, and how do sandboxes, persistent state, retries, and observability support longer tasks?

Across all five layers, we'll discuss evaluation, latency, and cost per successful task: how teams tell whether a system is improving, and which parts of the stack they should build or buy.

The Evening

Opening Keynote — A practical map of the emerging agent stack and the product opportunities around it.

Sponsor & Partner Spotlights — Focused, 10-minute introductions to the teams and technologies building different layers of the stack.

Panel + Audience Q&A — A discussion of architecture choices, production lessons, and the tradeoffs behind agents that can complete real work.

Live Demos — 10-minute slots to show agent products and workflows in action, explain how they work, and share what the team learned.

Founder & Builder Networking — Time to continue the conversation and meet people building complementary products and infrastructure.

Tentative Agenda

All times are local to San Francisco (Pacific Time).

• 5:00–5:15 PM — Check-in & Welcome

• 5:15–5:30 PM — Opening Keynote: What’s the New Agent Stack?

• 5:30–6:15 PM — Panel Discussion + Audience Q&A

• 6:15–7:00 PM — Agent Open Mic · 5-minute builder showcases

• 7:00–8:00 PM — Founder & Builder NetworkingSpeaker, partner-session, and demo details will be announced. The agenda may be updated as the lineup is finalized.

Who Should Attend

• Founders building agent products, developer tools, and AI infrastructure

• Engineers working on retrieval, memory, orchestration, tools, and runtime systems

• Product and technical leaders bringing agents into real workflows

• Investors exploring opportunities across the agent stack

About Seamate

Seamate is a community for founders, builders, and investors exploring AI, startups, and global markets. 40k+ members across 10+ countries. We run events where the emphasis is on who's in the room and what's actually being built — not on filling seats.