Videos DqtmZE6Hl0g
The Prompt is the Platform - Dominik Tornow, Resonate HQ
Scene timeline
37 shot(s).
keyframes kept every frame deduplicated
What was stored
- cues
- 190
- whisperx 190
- chunks
- 32
- from 190 cues
- keyframes
- 15
- kept of 37 captured
- frames with text
- 15
- 99 lines read
- chapters
- 0
- from the source metadata
- keyframe bytes
- 1.8 MB
- word timings on 190 cues
Provenance
| stage | state | model | started | took |
|---|---|---|---|---|
fetch |
done | — | 2026-08-11 05:45 | 1m 17s |
stt |
done | — | 2026-08-11 05:47 | 16s |
chunk |
done | — | 2026-08-11 05:47 | 0s |
text_embed |
done | — | 2026-08-11 05:47 | 0s |
keyframe |
done | — | 2026-08-11 05:47 | 49s |
ocr |
done | — | 2026-08-11 05:48 | 3s |
frame_embed |
done | — | 2026-08-11 05:48 | 3s |
Frames, and what the machine read
-
- The Prompt is the Platform1.00
- * Resonating...0.96
- Al Engineer0.98
-
- Dominik Tornow1.00
- Resonate HQ1.00
-
- general1.00
- bespoke1.00
-
- implementation1.00
- specification1.00
- implementation1.00
- implementa0.99
-
- Resonate Server1.00
- Resonate Server0.99
- Resonate Protocol1.00
- Resonate Se1.00
-
- Resonate Durable Execution1.00
- on NATS.IO1.00
-
- Target1.00
- Spec1.00
- Abstract1.00
- Concrete1.00
- Spec1.00
- Impl1.00
-
- Agent1.00
- Agent1.00
- Agent1.00
- Abstract1.00
- Simulation1.00
- Concrete1.00
- Concrete1.00
- Spec1.00
- Impl0.99
- Spec1.00
- Impl1.00
-
- >>> kvs = KVStore(...)0.98
- >>> kvs = KVStore(...)0.94
- >>> t = kvs.create("key", "foo")0.99
- >>> t = kvs.create("key", "foo")0.99
- >>> t = kvs.update("key", "bar", t)0.97
- >>> t = kvs.update("key", "bar", t)0.97
- >>> t = kvs.update("key", "baz", t)0.99
- >>> t = kvs.update("key", "baz", t)0.99
- >>> kvs.get("key")0.98
- >>> kvs.get("key")0.99
- (“baz", 2)#FRESH0.90
- ("foo", 0)#STALE0.93
Transcript
190 cues· 2,081 words· 12,329 chars
- 0:02 In 2026, coding agents will quietly retire their first software platform.
- 0:09 Not because it's bad, simply because the platform is unnecessary.
- 0:15 I am Dominik Thurnow.
- 0:16 I am founder and CEO of Resonate.
- 0:19 Resonate is a durable execution platform built with minimalism and simplicity as its core technical values.
- 0:27 And these properties will play a central role in this talk.
- 0:31 At Resonate, we have a working theory where software engineering is headed.
- 0:37 General purpose implementations will increasingly be replaced by bespoke implementations generated on demand, not as a new library, a new framework, or a new platform, but as a minimal extension of the infrastructure that is already in place.
- 0:57 If this theory holds true,
- 1:00 reuse will move upstream.
- 1:03 Instead of reusing a general purpose implementation, we will reuse a specification, and we will derive a bespoke implementation from it.
- 1:16 In fact, we can build many bespoke implementations, tailor-made for the infrastructure that is already in place.
- 1:24 We just have to ask the agent.
- 1:27 At this point, the prompt is the platform.
- 1:33 Resonate is a dual execution platform.
- 1:35 We have an implementation of the Resonate server.
- 1:38 We have implementations of the Resonate SDK for TypeScript, Python, Rust, Go, and Java.
- 1:45 So we have to ask, what does this new reality mean for us?
- 1:51 If implementations become generatable, where does our value live?
- 1:57 And our answer, our value moves from implementation to specification.
- 2:05 Now this changes how we think about Resonate.
- 2:09 The product is no longer the implementation.
- 2:12 The product is the specification, the protocol.
- 2:15 And from that protocol, we want to derive multiple server implementations.
- 2:22 One is a general purpose Resonate server, our reference implementation.
- 2:26 Others are implementations built with infrastructure partners.
- 2:31 For customers and partners, this means durable execution right on top of their existing infrastructure with minimal additional dependencies.
- 2:40 So the question is no longer, can we build a server?
- 2:44 The question is, can we repeatedly synthesize trusted servers from the same specification?
- 2:51 And if so, how?
- 2:58 When we talk about agentic engineering, we focus all of our attention on verification.
- 3:04 How do we know the result is correct?
- 3:07 But today, I want to focus on the specification instead.
- 3:12 And more importantly, how can agents participate in specifying the system, not just building or verifying it?
- 3:23 Now, Resonate is partnering with multiple infrastructure providers to bring durable executions natively to their technology stack.
- 3:31 One of them is Sunadia, the company behind NATSIO, an open source messaging system designed for building modern distributed systems.
- 3:41 For the rest of this presentation, we will use Resonate on NATSIO to explore our agentic engineering practices.
- 3:49 How do we go from specification to implementation?
- 3:53 First, we need to level set our mental model.
- 3:58 This picture is a common view of agent decoding.
- 4:02 There's an agent, there's a specification, and then there's an implementation.
- 4:08 And for many applications, that is enough.
- 4:11 But it is not enough for what we are trying to do, because we are not trying to generate one implementation from a specification.
- 4:23 We are trying to generate multiple target-specific implementations from the specification.
- 4:29 So the specification must not take any aspect of an implementation into account.
- 4:36 The specification must not assume a concrete database schema or concrete indices.
- 4:42 The specification must not even assume a relational database with tables and transactions at all.
- 4:48 It must not assume a key value store.
- 4:50 It must not assume weak consistency.
- 4:52 It must not assume strong consistency.
loading