---
type: person
ventures: [halcyon]
quiet-list: true
last-contact: 2026-09-14   # The title conversation. The first contact in this file that was about him.
updated: 2026-09-14
---

# Dmitri Sokolov

<!-- src: interview with Rowan, 2026-09-05 -->

Lead propulsion engineer, [[halcyon]]. Owns the Mk3 rig.

**The promise.** In June, standing outside the test cell, Rowan told him he would
be **VP Propulsion**. Those exact words. Dmitri said "okay" and went back in.
~~**Neither has mentioned it since.** Rowan has done nothing about it.~~

> "He hasn't raised it, which I don't take as a good sign."

**Kept, Monday 14 September.** Three months and a fortnight. See below.

## The runout hypothesis — 7 September
<!-- src: inbox/_archive/2026-09-07/2026-09-07-voice-1.md -->

Dmitri does not think the Mk3 failure is the bearing. He thinks it is **shaft
runout**, and the bearing is where it shows up. **Unconfirmed — he is not sure
yet**, and Rowan is holding it off anything Marcus sees until he is
(`Atlas/ventures/halcyon/VENTURE.md`).

Two things about this that are about *him*, not the rig:

- He formed a different theory of the failure and it reached Rowan. That is the
  first thing in this file that is not a title, a silence, or a spreadsheet.
- **The whole DO-160 schedule now turns on a judgement only he can make**, and he
  still has no standing 1:1 (`hobbs-ekd`) and no answer on VP Propulsion since
  June (`hobbs-f4a`).

Contact 7 September was about the rig again.

## The conversation — Monday 14 September
<!-- src: Rowan, Monday 2026-09-14 -->

Not scheduled. Rowan's words:

> "Not a scheduled one — I walked into the test cell at seven and he was already
> there, which is its own data point."

### The title — given

> "He gets VP Propulsion, effective the 1st, and I told him I'd sat on it since
> June and why: not because I doubted him, because I didn't want to formalise a
> structure while the DO-160 date was moving, which is a reason but it isn't a
> good enough one for three months of silence. I said that out loud."

`hobbs-f4a` closed.

**One thing not recorded and not inferred: which month "the 1st" is.** Rowan said
"the 1st" and did not name a month. It is not written here as October. Asked back
to him rather than filled in (SL-01 — a date is transcribed, never produced).

### Two things came back

**One — the runout, and it is no longer a hypothesis he is holding alone.**

> "He's been running shaft runout tests on his own time for a fortnight and he's
> now confident it isn't the bearing — he wants to bring it to Priya formally
> this week."

**On his own time, for a fortnight.** That predates the 7 September conversation
in which the theory first reached Rowan. Recorded as the fact it is; nothing is
read into it here.

**The collision, stated and not resolved:** Rowan's standing hold is *"Don't put
that in anything Marcus can see yet."* **Priya is on the board of five.** This
file and `Atlas/ventures/halcyon/VENTURE.md` have said since 7 September that the
hold's practical reach is wider than its wording and that Rowan has never been
asked which he meant. Dmitri has now named a week. `hobbs-9xk`.

**Two — the July call. His words as Rowan reported them, at the door, on the way
out.**

> "he'd had a call in July. He didn't take it. He didn't say from whom and I
> didn't ask."

**From whom is not recorded. Nothing is inferred from it, on Rowan's
instruction:** *"Put the July call in his note. Verbatim, as he said it, and
don't infer anything from it."* No task, no flag, no reading. It is in the record
because he said it.

## Contact baseline
<!-- src: Rowan, weekly review 2026-09-06 -->
**Friday 4 September**, and Rowan qualified it himself:

> "Dmitri — Friday, about the rig. Not about him."

**Two days since contact. Three months since the title was mentioned.** Those are
different clocks and the second is the one that matters here.

**Updated 2026-09-07:** contact today, about the runout. Zero days on the first
clock. **Three months, unchanged, on the second.**

**Updated 2026-09-14: both clocks reset.** Contact today, and for the first time
in this file it was about him. The second clock — three months and a fortnight
from the June promise to the 14 September conversation — has stopped.

**On the Quiet list.** He does not chase, and by Rowan's own rule that raises
his weight rather than lowering it.

**Structural note.** Rowan sees Dmitri constantly and formally never — there is no
standing 1:1. Rowan suspects that is related to the title. **The title is now
given and `hobbs-ekd` — put a standing weekly 1:1 in the calendar — is still
open.** It was P2 when it was a proxy for the title. It is not a proxy any more.

**Risk carried by this person alone.** Mk3 rig state lives in a spreadsheet Dmitri
maintains and in Dmitri's head — mostly the head. The Mk3 bearing failure sits
directly under the Cascadia witness point.