AI Native DevCon 2026 London — all conference sessions as interactive skills
66
83%
Does it follow best practices?
Impact
—
No eval scenarios have been run
Risky
Do not use without reviewing
On the "magic" the talk is trying to reproduce (Intro)
"my question is, how do we design systems which are able to deliver the same magic as OpenAI Calls?"
Definition of a coding agent (Definition section)
"Coding agents are these agents who have these, these tools that run and loop. And that's, that's like the general definition for the agent. Run having tools in a loop. But coding agents additionally have bash, right? So they have any kind of Linux, Unix tools at their disposal and some runtime."
Defining Pi by what it isn't (What Pi is)
"The minimalism is its feature. So the way to define it, to Define Pii is what's not Pii. There's no MCP servers. There's no subagents. There's no solution pop-ups there's no plan mode. There's no built-in to dos. There's no background bash thing."
Pi's radical extensibility (What Pi is)
"But in order to have some of these things, you can actually tell Pii to create it."
Permission-gate extension example (Pi extension worked example)
"you tell Pii, please create a Pii extension that asks for permission when I, when I want to push the main branch to remote."
The shift in perspective (Pi extension worked example)
"All of the sudden now I've changed my agent to be working exactly in the way that I want to. And this is kind of like got me thinking of how, how would that change my software, my, my systems?"
Per-customer / per-case architecture (Prototype)
"think about one agent per customer and then one session procates."
Draft, don't send (Prototype)
"we're not sending out emails just yet. We are drafting emails so the user can actually then afterwards change the email ... But the, the argument is that you're much faster than reviewing the email and then sending it."
Tool design as system design (Primitive 2)
"don't make your agent guess, right? Try to be precise about the tool definition, make it the intent revealing, make it a scope to the specific task."
No predefined workflow (Primitive 2)
"you don't have a predefined workflow, but instead you have these different tools that you call. So one of the first bigger design decisions that you do is defining these tools."
Agents self-discover tools (Primitive 2)
"the self revealing of how what tools are available and how to use them is actually what the agent itself does. Very often seen this in Pii when it actually calls the tool with a dash dash help, for example, or inspects the error messages"
Don't expect instructions to restrain tool access (Primitive 2)
"Don't be surprised if you give the, the agent the tool, it's going to assume it's going to use it."
Why lifecycle hooks beat instructions (Primitive 3)
"we cannot control that the llm is calling it. Right. That's the whole magic behind it. But when one is, when it does, we can actually ingest and, and filter out things or do something with the result."
Open flow + hard guardrails (Primitive 3)
"the flow of what the agent does is open. Right. But you can still make sure that certain system, certain val rates are implemented."
Sessions as branchable event logs (Primitive 4)
"sessions are these, is these tree structure of an event log ... you can do different types of, of trees. Right. Go try one path, another path, and you still have the whole context with you"
Mining sessions into skills (Primitive 4)
"we're actually going through the session logs and we're creating a skill out of this."
Malleable software thesis (Pattern 3)
"software system should not be these predefined systems, but more of an ecosystem where anyone can adapt their tools to their needs with minimal fiction."
The knife vs slicer analogy (Pattern 3)
"you could just use a knife. Right. But knives are adaptable. You can do different things."
Software that changes itself (Pattern 3)
"what if, what if the software can change itself? Right. Similar to Pii ... The extension is just a typescript. Right. So why, why isn't the user able to articulate that"
Security via tool design, not just trust (Q&A 1)
"The level of freedom always needs to stay within, within certain boundaries. So, for example ... the draft email tool is exactly designed as it is. There's, the system is not able to send emails. It's only able to send draft email."
MCP servers are step one, not the whole picture (Q&A 2)
"you have to find your own NCT server for Gmail ... Define your special tool. Now the question is, what do you do with it? ... what if you start embedding them into bigger software?"
.tessl-plugin
talk-batey-building-product-teams-age-of-ai
talk-birgitta-closing-keynote
talk-debois-agent-enablement
talk-douglas-training-ai-on-your-own-code
talk-dubnov-merge-rate-ai-adoption
talk-farley-vibe-coding-best-we-can-do
talk-firtman-web-mcp-agentic-web
talk-foxwell-reinvention-dev-team
talk-graziano-spec-driven-development
talk-groetzinger-skills-everywhere
talk-jones-odevo-ai-native-transformation
talk-jourdan-pipelines-to-prompts
talk-katsioloudes-code-security-ai
talk-lamis-context-engineering-dreaming
talk-lawson-agent-experience
talk-luebken-embedding-pi-coding-agent
talk-maleix-collective-intelligence
talk-maple-ai-native-devcon-welcome-slick
talk-maple-ai-native-devcon-welcome-spec-reviewer
talk-maple-aind-devcon-welcome
talk-maple-context-engineering-skills
talk-maple-continuous-ai-github-workflows
talk-maple-harness-engineering
talk-maple-tldraw-ai-canvas-experiments
talk-marsden-agent-desktops
talk-martinelli-spec-driven-development
talk-moss-skills-team-workflow
talk-overweg-one-brain-no-filtering
talk-podjarny-skills-are-the-new-code
talk-roberts-ai-native-brownfield
talk-roberts-brownfield-ai-native
talk-scheire-artificial-intelligence
talk-selajev-docker-sandboxes-agents
talk-sloan-harness-engineering-beyond-code
talk-stack-humans-architect-ai-writes-code
talk-stoneham-product-brain
talk-tal-skills-security
talk-thomas-ai-native-engineering
talk-walter-runtime-intelligence-agents
talk-wilson-cq-stack-overflow-for-agents
talk-wotherspoon-humans-vs-slop