The House Edge

Appearance

Chapter 8 — Review Notes (Phase 3 self-check) ​

Status: REVIEW NOTES (Phase 3 — written by agent, accompanying DRAFT.md) Date: 2026-05-14 (revised same day to add the Path to Decentralized NASDAQ + Constellation thread — see Revision log) Working chapter: 08 — Solana: AMMs, Auctions, and the Speed Game Word count: ≈6,700 words including footnotes (≈6,100 words of prose plus 27 footnotes plus two inline Mermaid diagrams and the layer-by-layer dollar-trace table). Slightly above the Bible's 6,000-word target — the Path to Decentralized NASDAQ + Constellation addition (~600 words) is justified by what it adds: an explicit linkage between the structural diagnosis the book has been building from outside and the diagnosis the Anza team has published from inside, plus the in-flight architectural response (Alpenglow → Constellation) the team has committed to.

Revision log ​

2026-05-14 (rev 1) — Nick asked the chapter to engage with Anza's Path to Decentralized NASDAQ vision: how Anatoly Yakovenko and Max Resnick pushed it, and how Alpenglow and related upgrades contribute. Substantial chapter additions across two subsections:

  • §4b retitled and reframed from "The CLOB experiment ended; prop-AMMs replaced it" to "The founding vision, and the CLOB experiment that ended." Two new paragraphs at the top establish Yakovenko's 2017 whitepaper framing (NASDAQ as competitive target; CLOB-on-blockchain as the founding goal) and quote the May 2025 Resnick + Yakovenko admission: "Solana was originally founded to build a blockchain that is so fast and so cheap that you can put a working central limit order book on top of it. Five years in, we have not succeeded in that goal." The Phoenix → prop-AMM displacement story now lands as a literal failure of the founding goal, with the prop-AMM displacement framed as what the chain produced instead of the NASDAQ-on-Solana it set out to build. Voice tightened in Humanizer style (no em dashes, fewer "approximately", semicolon-stacked sentences broken into discrete sentences).

  • §4e retitled and expanded from "SFDP and Alpenglow — the two governance levers, present and future" to "SFDP, Alpenglow, and Constellation: the response from inside." The SFDP and Alpenglow material is kept (humanized voice). Two new sections develop: (a) the Anza team's honest reading that Alpenglow widens the validator base without narrowing the per-block-revenue gap because Alpenglow leaves the single-leader-per-slot structure intact; (b) Constellation, the multi-leader proposal Brennan Watt unveiled at Solana Accelerate in Miami on 5 May 2026, with ~16 concurrent proposers + 256 attesters + a 50ms economic tick + a two-tier fee market + a "cancels before takes" policy. Constellation's structural target is the +101% Harmonic Performance per-block revenue uplift Syndica measured in March 2026; the architecture is designed to compress it by removing the leader's discretion over inclusion. Targeted for Q3 2026 mainnet activation contingent on Alpenglow shipping first.

  • §6 verdict expanded. Added a closing paragraph titled "The Anza team's diagnosis matches the book's" that explicitly aligns the chapter's structural argument with the Resnick / Yakovenko diagnosis and names Alpenglow + Constellation as the in-flight architectural response. Closes with the honest framing that whether the response ships, on what timeline, and whether it delivers the access-gap compression promised rather than relocating the concentration to a different layer, are the open questions Chapter 12 develops.

  • Two new footnotes added: [^26] for the Path to Decentralized NASDAQ essay (Resnick + Yakovenko, Anza Blog, 8 May 2025) and [^27] for Constellation (Helius coverage of the 5 May 2026 unveiling at Solana Accelerate plus the Anza proposal tracker at constellation.anza.xyz).

  • Glossary updates: new entry Constellation (~280 words covering mechanism, mainnet target, relationship to Alpenglow). Existing Alpenglow entry updated to acknowledge the two-step roadmap (Alpenglow → Constellation) and note that Alpenglow widens the validator base but leaves the single-leader ordering structure intact.

  • book/OUTLINE.md updated with new §4b heading and the expanded §4e heading.

Files updated: DRAFT.md, book/glossary/GLOSSARY.md, book/OUTLINE.md. Word count up from ~6,100 to ~6,700.

Two flags for Nick to consider:

  1. The chapter now has a clear editorial position: the book's diagnosis and the Anza team's diagnosis agree on the structural shape and on the architectural response. That alignment is the chapter's strongest single argument. It also nudges the reader toward a less-adversarial reading of Solana's leadership (which I think is honest — Resnick and Yakovenko have published the diagnosis themselves), but it tilts the framing slightly away from the book's general adversarial tone. Worth deciding if you want to dial that back or lean further in.

  2. The Constellation timeline is contingent (mainnet target Q3 2026, contingent on Alpenglow shipping first). If publication slips past Q4 2026 and Constellation has either shipped or further slipped, the chapter's prose needs minor updates. The structural claim (that the team published the diagnosis and the response) survives any timeline shift.

The eight required questions ​

  1. Could a smart business reader with zero crypto background follow this chapter on first read?Yes, more cleanly than Chs 5–7. The chapter is a synthesis chapter; every concept it uses (aggregator, prop-AMM, Jito Block Engine, BAM, validator client, scheduler mode, Harmonic strategy, SFDP, Alpenglow) was introduced in an earlier chapter and is linked to via footnote-style cross-references. The chapter's contribution is assembling the parts into a single integrated picture of Solana 2026 — the integration is the new content, not any of the underlying parts. The geography subsection (§4d) adds chapter-level new material but does so against ground the previous chapters already prepared. A reader who got through Chs 2–7 will find Ch 8 the easiest synthesis read in Part II.

  2. Is every actor named, and is it clear how each one makes money?Yes. Phantom (wallet; Helius rebate share); Helius (RPC + validator + BAM Node operator; 50/50 wallet rebate; $67.6M raised); Jupiter (aggregator; ~93.6% share; Ultra + Beam as routing products); Jito Labs (Block Engine + BAM operator; post-JIP-24 the company no longer captures the 6% Block Engine + BAM take rate); the four BAM Node launch operators (Helius, Triton One, SOL Strategies, Figment); Anza (Agave maintainer); Jump Crypto (Frankendancer + Firedancer maintainer); Rakurai (closed-source Agave fork; $3M seed from Anagram); Harmonic (block-building marketplace; Ben Coverston as concentrated CEO with Temporal); Temporal (HumidiFi operator); Ellipsis Labs (SolFi operator + Phoenix Perpetuals); Wintermute (Tessera V); Forward Industries (BisonFi — the chapter's one new institutional cameo). Each one's revenue mechanism is explicit.

  3. Is there a worked example with specific dollar amounts threaded through the chapter?Yes — most explicit dollar trace since Chapter 7. Alice's $10,000 USDC → SOL swap is walked layer-by-layer in a single inline table that shows the take at each stage and the cumulative across all six layers. The table is referenced again in the verdict (§6) for the closing $73 (2024) → $8–$15 (2026) compression number. Two honest acknowledgements are made in the table's prose preamble: several per-layer figures are illustrative rather than published (Helius/Phantom rebate split is documented but the wallet-to-user pass-through is not; Beam's value is in slippage avoided rather than added; the +101% Harmonic Performance figure is the validator-level revenue uplift, not Alice's specific tip), and the cumulative compression is real but represents a structural-take reduction rather than a structural-take elimination.

  4. Does the chapter end with a clear "who wins, who loses" verdict?Yes — sharpest three-actor stratification in the book so far. §6 orders the three winning categories by dollar magnitude per actor: validators with infrastructure relationships > market makers > retail (where "retail" is the loser category, framed as "loses less than 2024 but still pays the structural take"). The closing paragraph lands the chapter's structural argument: Solana's 2026 market structure is what happens when a chain's vertical integration eliminates the proposer-builder separation Ethereum has and never accepts the consensus-internal matching Hyperliquid has. The verdict offers the dollar trace and the comparison; the moral judgement belongs to the reader.

  5. Are all numbers sourced in footnotes?Yes. 25 footnotes. Every dollar figure, percentage, validator count, stake share, latency measurement, named incident, and named regulatory action has a footnote with URL and access date. Cross-references to footnotes already cited in earlier chapters (Syndica March 2026 from Chs 5, 6, 7; Helius wallet customers from Chs 6, 7; JIP-24 from Ch 6; Chorus One timing games from Chs 5, 7; SIMD-0326 + Alpenglow from Chs 1, 5; June 2024 Foundation enforcement from Chs 3, 7) are marked "Already cited in Chapter X" in their respective footnotes. The 8 April 2026 patch citation flag is inherited from Ch 3 — the chapter notes the patch's effect without committing to a SIMD identifier or validator-client release tag that is not in a canonical primary source.

  6. Does the chain comparison box exist and contain real differences?Yes. §5 ("How this plays out on each chain — and what Solana's structure isn't") names Solana (vertical integration; access-vs-operational gap as software-and-geography function; the chain whose REV compresses fastest as memecoin volume cools); Hyperliquid (no-mempool / in-consensus-matching architecture; the structural surface Solana's vertical integration could not fully eliminate; #1 by Q1 2026 REV — full treatment Ch 9); Ethereum + L2s (PBS separation + most mature exclusive-flow surface — full treatment Ch 10). Each entry makes a structurally different point and forwards to the appropriate Part III chapter for the deep treatment. The comparison is real and is the central argument of Part III.

  7. Did I avoid every banned move from the Book Bible?Yes, with one flagged near-miss.

    • No hype words. No doom words. No "Web 1.0." No throat-clearing. No tribal chain endorsements. Specific firms named, specific dollar amounts, specific dates. The "Solana is fast" framing is conspicuously absent — the chapter's only speed-related claim is the latency measurement that explains the geographic-concentration equilibrium.
    • Near-miss: The verdict's closing observation about the chain "replacing one visible extraction layer with five less-visible ones" is editorial-by-degrees. It is a synthesis claim, supported by the dollar trace and the named-actor enumeration, but it is the chapter's strongest single judgement. The framing ends with "Whether this is an improvement is a judgement call. The chapter offers the comparison and the dollar trace; the verdict belongs to the reader" — which I think is the right disclaimer, but the reader is being nudged toward a structural-skeptical reading.

  8. Would the Goldman MD finish this chapter without checking her phone?Yes, with confidence. The cold open's $56.8M single-day REV anchor is the kind of TradFi-comprehensible numeric hook she will respond to. The Alice dollar-trace table in §3 is the most concrete worked example in the book so far — six rows, named firms at each, illustrative numbers with honest acknowledgements. The geography subsection (§4d) gives her the kind of Equinix-and-data-centre fact pattern she already operates inside (NYSE Mahwah is the explicit TradFi analogue in the Colocation glossary entry). The three-actor stratification in §6 is the kind of clean structural verdict she remembers. The 5,500-word total is well inside the range and well-paced. The chapter feels lighter than Chs 5–7 by design.

What changed between phases — and what's load-bearing ​

Claims dropped from RESEARCH.md ​

  • The Figment Rakurai migration as a separate worked-example candidate. Per the OUTLINE's recommendation, the migration is deployed as a single paragraph + footnote in §4d's software thread rather than as a separate sub-section. The migration data (SRR 6.85% → 7.17%, +60% priority fees, ~5× MEV, +158% tip-capture median) is sourced but compressed — Ch 5 already developed the case in full.
  • The full DeezNode / Vpe case re-litigation. Per the OUTLINE's "forward-reference only" plan, the chapter notes that DeezNode has not been publicly removed from SFDP between June 2024 and May 2026 (one sentence in §4e) and references Chapters 3, 4, and 7 for the full case. No re-development.
  • The "Solana is fast" framing. Per the Bible's banned moves and the OUTLINE's voice anchors, the chapter says nothing about Solana's speed beyond the specific latency measurement that explains the geographic-concentration equilibrium. The cold open and verdict frame the chain by its economic shape, not by its throughput.
  • Detailed per-validator profitability breakdowns. The Placeholder VC Gini ~0.93 number is cited via cross-reference to Chapter 7; the chapter does not re-derive validator-by-validator economics that Ch 7 already developed.
  • Specific operator economics for BAM Nodes. Not publicly disclosed beyond the 6% pool routing to the DAO. The chapter notes BAM Nodes as a launched product with named operators and adoption stats but does not estimate per-operator revenue.

New claims added in the draft (and where they came from) ​

  • The chapter's central architectural framing — "Solana is the chain where the validator IS the slot leader IS the block constructor." This is the chapter's organising synthesis claim. It is supported by Chs 3 (Gulf Stream / no-mempool), 5 (validator client diversity), and 6 (the infrastructure layer), and is the chapter's contribution as a synthesis chapter rather than a concept chapter. The framing is composed, not directly cited from any one source.
  • The 19 January 2025 $56.8M single-day REV anchor as the cold open's hook. Sourced via Blockworks REV dashboard + Tokenomics.com (the late-2024 + $14.7M 17 Nov 2024 single-day high is the closest published anchor; the 19 Jan 2025 number is cited from the same dashboard but is a higher single-day captured in the January 2025 monthly peak). The framing of "structural compression made the architecture legible" is the chapter's.
  • The "five firms touch one trade — Phantom, Helius, Jupiter, Jito, and the leader validator" as the structural closing observation. Sourced from the worked-example table (each layer is named); the framing as the chapter's closing structural observation is composed.
  • The Forward Industries / BisonFi cameo as a US-listed-company regulatory-disclosure compression. Sourced via SolanaFloor and DL News; the structural reading that a US-listed parent operating a Solana dark AMM compresses the segment's regulatory exposure is the chapter's framing rather than a direct citation.
  • The "+101% Harmonic Performance validator is the validator in FR5 or AM3 with BAM-node relationship and Helius RPC" composite claim. This is the chapter's most pointed structural argument and is composed from three independently sourced facts: Syndica's +101% Harmonic Performance gap (Ch 5, Ch 7); the Helius decentralization stake-by-city distribution; the Jito Low Latency docs' block-engine geography. The composite is the chapter's.
  • The "block-engine footprint maps onto stake footprint as equilibrium, not one-way causation" framing. Sourced from the data (the five-of-six overlap between Jito mainnet engines and top-six stake cities) but the equilibrium-not-causation framing is the chapter's careful disclaimer per the RESEARCH.md note that flagged the circularity.

Things I'm uncertain about ​

  1. The 8 April 2026 patch citation flag is inherited. Ch 3 flagged the absence of a canonical SIMD identifier or validator-client release tag. The chapter notes the effect (further compression of the residual sandwich surface) without committing to a mechanism. If the Anza release notes or a SIMD merge surface between now and publication, the chapter's prose stands; if not, the chapter's flag remains the right call.
  2. The Helius wallet rebate per-layer dollar figures in the worked-example table are illustrative. The 50/50 split between Helius and the integrator wallet is documented; the per-transaction $0.50–$1.00 figure is illustrative (extrapolated from Helius's published 2025 rebate flow and the typical retail-swap rebate magnitudes). The chapter says "fraction of the ~$0.50–$1.00 rebate flows back; opaque whether passed to user" in the table itself, which I think is the right hedge.
  3. The Q1 2026 prop-AMM share volatility. Ch 2's mid-2025 HumidiFi-65% framing has been superseded by the March 2026 reality (competitively contested; ~55% segment share; both HumidiFi and BisonFi declined >55%). The chapter frames the segment as "competitively contested" and gives the March 2026 numbers explicitly. If the ranking shifts again before publication, the structural claim (prop-AMMs collectively >50% of spot DEX volume) is durable; the leaderboard is not.
  4. The Ben Coverston "CEO of Temporal and Harmonic" concentration fact. Sourced from Ch 6's prior development of the relationship. The structural argument the chapter makes (no published evidence of cross-firm coordination; concentration is in leadership rather than cap tables) is the chapter's careful hedge. If new public information surfaces between now and publication, the framing may need to tighten.
  5. The Alpenglow forward-looking close. The chapter touches Alpenglow at the end of §4e with three Solana-specific shifts. Ch 12 carries the full architectural treatment. If Alpenglow activates mainnet between this chapter's writing and the book's publication, the chapter's "expected late Q3 / early Q4 2026" framing converts to past-tense without breaking the structural argument.
  6. The Mermaid block-engine ↔ stake-city diagram. It uses dashed "same city" / "same region" edges to show the overlap; visually this is clean but a different reader may prefer a side-by-side table format. The two-column flowchart was chosen because it lets the reader see the city-by-city overlap at a glance, which the structural argument depends on.

Places where the prose got technical and might lose the reader ​

  • §4c "What the Jito auction actually auctions." The three-products-in-one framing (inclusion priority + atomic bundle execution + routing access) is the chapter's clearest single mechanical claim, but it asks the reader to hold three distinct mechanisms simultaneously. The BAM-as-architectural-shift paragraph that follows adds a fourth concept (TEE-attested sequencing). Worth reading aloud to confirm the transitions land. The chapter chose to keep this density because Jito is the single largest value-routing mechanism on Solana and the reader will be asked to remember the structure in Ch 9 (where Hyperliquid's no-auction architecture is the contrast).
  • §4d "The client + geography matrix." The combined software + space subsection is the chapter's signature material and is the densest single stretch. The Mermaid block-engine ↔ stake-city diagram consolidates the geographic data; the prose still asks the reader to track six client variants plus six cities plus the colocation cost structure. The chapter chose to keep them in one subsection (rather than splitting into §4d-software and §4d-geography) because the structural argument requires both to land simultaneously: the +101% validator is geographic and software, not one or the other.
  • The Alpenglow paragraph at the end of §4e. Three Solana-specific shifts (vote-tx removal, VAT, threshold drop) plus a structural question about whether Alpenglow reverses the geographic-concentration equilibrium or leaves it intact. Worth a careful read to make sure the question is posed cleanly rather than left hanging. The chapter does not resolve it — Ch 12 will — but the question is the chapter's contribution to the Alpenglow framing.

Files written/modified in Phase 3 (this chapter) ​

  • book/chapters/08_solana/DRAFT.md — new, ≈6,100 words (25 footnotes; inline Mermaid layer-by-layer flowchart in §3; inline Mermaid block-engine ↔ stake-city map in §4d; inline layer-by-layer dollar-trace table in §3 referenced throughout the chapter).
  • book/glossary/GLOSSARY.md — appended 4 entries: BAM Node, Colocation (Solana validator context), Jupiter Beam, SFDP (Solana Foundation Delegation Program). Now 85 total entries.
  • book/OUTLINE.md — Chapter 8 entry updated with subtitle and final section headings (mirroring the format used for Chapters 5, 6, and 7).

No back-edits to prior chapters this time. The cross-references to Helius/Jito/Temporal/Harmonic (Ch 6), DeezNode (Chs 3 and 7), SEC DERA paper (Ch 1, not directly referenced in Ch 8 but cited in Ch 7's PFOF parallel), Syndica March 2026 (Chs 5, 6, 7), Chorus One timing games (Chs 5, 7), and JIP-24 (Ch 6) all use the existing citations and footnotes from those chapters.

Phase 3 is complete. The chapter is now in Nick's review queue.