The Write Last, Read First Rule How to guarantee strict serializability across database systems A special guest post from @dominiktornow.bsky.social tigerbeetle.com/blog/2025-11...

Partial Failure & Partial Retry RPCs create recovery boundaries: When an execution crashes, only the executions within its recovery boundaries restart Executions in different recovery boundaries (across RPCs) continue running.

Partial Failure but Total Retry?! If we follow the familiar model of chained HTTP calls then a failure in the chain forces a restart from the failure point This works for short-lived, cheap executions, but collapses under long-lived, expensive ones

From the terminal to the cloud, from prototype to production Learn how agents scale beyond a single machine and master identity, coordination, and recovery Build a recursive, distributed Deep Research Agent with @resonatehqio.bsky.social + OpenAI—in under 120 lines Chapter 3 is live

A Promise doesn’t just represent eventual completion A Promise also represents eventual invocation We just tend to forget the latter

I'm experimenting with animations for the 3rd chapter of Systems Engineering for Agentic Applications This shows the call graph of the recursive, distributed Research Agent unfolding over time, highlighting durable promises I dig the result but they are a lot of work 🥲

Weekend Treat ... sorry ... Read 🎃 Formal models are magic. Not only as a verification tool but as a thinking tool, lifting the fog of uncertainty and assumption Their magic lies in rigorous, ruthless, relentless clarity

Two sentences explain I/O better than any CS course If an expression yields an effect, it percolates through the program to the top, with the context stored as a continuation. An external authority handles the effect and passes the result back to the continuation.

Uncertainty, formalized: Possible Worlds The fewer worlds an agent considers possible, the less his uncertainty, and the more he knows In Reasoning about Knowledge

Trampolined Style If you want to understand Event Loops, you need to read this paper from 1999: A trampolined program is organized as a single loop in which computations are scheduled and allowed to proceed in discrete steps. Grokking Event Loops = Grokking Continuations

Join me in 30 min. I'll be in the live chat

Building a recursive, distributed Deep Research Agent, scaling to hundreds of agents across dozens of nodes Chapter 3 covers: ➡️ Agent Identity ➡️ Coordination ➡️ Recovery When agents leave the terminal, you need to become a distributed systems engineer

Roll up your newspapers, we’re squashing Heisenbugs with Deterministic Simulation Testing 🗞️ Join my talk tomorrow 11/22, 11:20 AM PDT I'll be in the live chat for Q&A Sign up below ⬇️

T-2d

Run your app for minutes, hours, days, weeks, or month On FaaS like AWS Lambda or GCP Functions Serverless @resonatehqio.bsky.social

Introducing Serverless @resonatehqio.bsky.social Build distributed, durable applications with Resonate's dead simple programming model, Distributed Asynchronous Await Now on serverless

Did the dreaded Heisenbug infect your system?! Then don't miss my talk on Deterministic Simulation Testing See you next week

The 1st Table Mountain Database Management Seminar Oct 9, 2025 An overnight hike — DBMS philosopher walk and talks, card games, South African braai under starry skies — with TigerBeetle, professors Philippe Bonnet and @peterabcz.bsky.social, and @dominiktornow.bsky.social #TMDBMS

From brainstorming transactional applications with the @tigerbeetle.com team to drafting blog posts with Table Mountain as backdrop Cape Town is a whole vibe

In the next chapter of Systems Engineering for Agentic Applications: When agents leave the terminal, agentic applications become distributed applications. We'll cover ➡️ Coordination ➡️ Recovery ➡️ Progress And why you should never hang your agent off a TCP connection 🔪

Distributed Async Await ✅ Simulation tested ❌ Formally verified Interested in some live coding ... wait, no ... specifying a distributed protocol?! Who’s up for a webinar to formalize @resonatehqio.bsky.social protocol stack. From scratch. In quint. Let me see hands

Chapter 2 gets a more formal treatment of Agents, environments, autonomy, alignment, etc Helpful or heavy?!

MCP: real innovation or literally just a function call?! Where do you stand in the protocol debate?

The second chapter of Systems Engineering for Agentic Applications just dropped This chapter covers ➡️ Agents & Environments ➡️ Goals & Plans ➡️ Context Engineering + We build an agent to organize our desktop Bonus: Is MCP just calling functions by name?! bsky.app/profile/domi...

The second chapter of Systems Engineering for Agentic Applications will drop tomorrow We'll cover ➡️ Agents and Environments ➡️ Goals and Plans ➡️ Autonomy and Alignment ➡️ Context Engineering + We build an agent to organize our desktop agenticapplications.substack.com

You cannot step into the same river twice －Heraclitus, Greek Philosopher An agent is a tuple (model, tools, system prompt, history) A = (M, T, s, h) After each interaction, the history changes A' = (M, T, s, h+i) Is A' still A? When is A' still A #PhilosophyFootnote bsky.app/profile/domi...

Agents don't have a plan Often, we prompt agents to plan out loud. But agents do not commit to a plan. The proposed plan becomes part of the history, the context, biasing but not binding future decisions. Why? Agents are turn based. Every step is a new decision

The #P99CONF speaker gift just dropped. Box says: Get swag, not lag I say: 10/10 swag box Come see me squash Heisenbugs with deterministic simulation testing bsky.app/profile/domi...

From pixels to paper, from Manning to mailbox, Think Distributed Systems is finally in my hands. Thank you everybody for your encouragement and being on this journey together ❤️

Prompt Engineering, formalized Prompt Engineering has felt intangible to me. Hype but no precision In Chapter 2, we define Prompt Engineering as a search problem that maximizes a fitness function on an agent’s trajectory Too abstract?! Too dense?! Don't worry. We build this from the ground up

Engineering without a dependable mental model is just guesswork In Chapter 2, we fuse theory with practice and formalize goals, tasks, trajectories, alignment, and even prompt engineering. Timeless mental models you'll use forever

In Chapter 2 of Systems Engineering for Agentic Applications, we build our own MCP. Not another MCP server but our own protocol We'll discover - What is MCP - What does MCP do - Why would we even need MCP

Durable Execution, distilled Distributed Async Await is so simple, the API fits on @resonatehqio.bsky.social's landing page That's not just a feature, that's a flex

Hot take MCP's Achilles heel? Requiring a persistent TCP connection How do you handle long running tool calls?!

Chapter 2 covers a lot of ground: → Goals and subgoals → Tasks and subtasks → Autonomy → Alignment → Open loop agents → Closed loop agents → Agent equivalence → Agent compatibility The foundations for building systems, not toys

Your desktop is a mess? In the next chapter of Systems Engineering for Agentic Applications we build an agent to organize our desktop A simple task and a perfect lens for understanding how agents work and the systems engineering challenges they present

One of the hardest problems in agentic applications is alignment: An agent is aligned with the user if the agent ranks possible states the same as the user

No frameworks In the 2nd chapter of Systems Engineering for Agentic Applications, we’ll build an agent from the ground up Why? To study agents from first principles and identify the systems engineering challenges we’ll face at scale

The first chapter of Systems Engineering for Agentic Applications just dropped This chapter covers 1⃣ AI foundations 2⃣ AI agents and agentic applications 3⃣ Code samples exploring AI APIs and simple agents systems-engineering-for-agentic-applications.resonatehq.io/chapters/cha...

The first chapter of Systems Engineering for Agentic Applications will drop tomorrow We will cover ➡️ AI foundations ➡️ AI agents and agentic applications ➡️ First contact with AI APIs and AI agents

Virtualization, an abstraction where the logical resource resembles the physical resource ➡️ Virtual machines are still machines ➡️ Virtual networks are still networks ➡️ Virtual memory is still memory Abstractions hide. Virtualization impersonates.

Virtualization ≠ Virtual Machines So what is virtualization then?! What's your mental model?!

Systems Distributed is hands down the best distributed systems conference today Thea Klaeboe’s presentation on how CERN powers physics discovery at massive scale and speed is absolutely riveting One of the best talks I’ve ever seen Premiers in 2 hours

The biggest lie in Software Engineering Undo There’s no undo Only a new Do that resembles the old state The past is immutable We just pile on a new present

I'm thinking about publishing Systems Engineering for Agentic Applications in multiple languages AI can handle the first drafts, but I can't validate the translations If you like to collaborate on validating and refining a translation in your language, please DM me bsky.app/profile/domi...

In Event Sourcing, what’s the right order of operations?! 1️⃣ Receive → Process → Persist → Apply → Return 2️⃣ Receive → Process → Apply → Persist → Return Do you persist before apply or apply before persist?

"Traditional applications execute predetermined flows, agentic applications dynamically orchestrate actions" Predetermined flows sounds sharp, but no non-trivial program unfolds in a predetermined way So what does dynamically orchestrate actions really mean?!

One chapter turned out too long for Substack. So I will publish the chapter announcements on Substack and the chapter on Docusaurus Bonus Easier to collaborate :)

Even simple examples hide big systems engineering lessons: Chat Interface Know what’s ephemeral vs durable Streaming chunks → ephemeral transport Complete responses → durable storage Same origin. Different lifetimes. Different infrastructure.

The DST in @resonatehqio's TS SDK injects failures by pseudo-randomly replacing a character in a message, simulating: • Invalid format • Valid format, invalid schema • Valid format and schema, invalid payload Simple yet packs a punch bsky.app/profile/domi...