Claims register

This is the detailed evidence trail behind index.md §6. The index table is a quick reference; this file is authoritative when the two differ. Every entry records what was claimed, when, by whom, what changed, and what would settle it.

The register exists because the corpus spans 2024-08-04 to 2026-10-22 and Tyler repeatedly revised his own positions. Quoting a two-year-old video as if it were current is the most common way to misuse this database.

How to read an entry

Each claim has:

• Status — one of the categories below.
• Established — what the record actually supports, separated from inference.
• Position history — dated changes, with the confidence label used in the source summary.
• Assessment — why the current position is where it is.
• What would settle it — the artifact or statement that would end the debate.
• Sources — the summaries to read, plus event counts where useful.

Status vocabulary

Evidence tiers used in the position histories

Fact in a summary means a named artifact supports the claim; it does not mean this project independently verified the artifact. Verified in events.tsv means the artifact is directly checkable. The two labels overlap but are not identical.

Summary

C01. Project White Sands is HLX

Status: Corrected. “Project White Sands” is a Deadlock voice-casting codename. It is not HLX and should not be cited as HLX evidence.

Established. On 2024-08-04 a voice actor’s portfolio/resumé listing for “Project White Sands” surfaced. Tyler discussed it in HLX Files #0 and treated it as a Valve casting cover name connected to HLX. In Every Last Detail Ep1 (2026-01-29) he corrected the record: White Sands was a code name used for Deadlock auditions, and the voice actors themselves confirmed it. Miro #0849 and #0850 record the same conclusion.

Position history

Assessment. The correction is well supported by multiple actor statements and is one of the few places where Miro and the latest video agree outright. The original portfolio page was not archived in a privacy-safe form, so the initial discovery still lacks a primary snapshot. That is a documentation gap, not a reason to doubt the correction.

What would settle it. A privacy-safe archived copy of the 2024 casting listing, or an on-record actor statement naming Deadlock and White Sands together. Neither is required to stop using White Sands as HLX evidence.

Sources. HLX Files #0 · Every Last Detail Ep1

C02. HLX is only an engine feature or save format

Status: Superseded. The “HLX is just a save format / engine flag” argument was a real early hypothesis. Tyler rejected it in 2024 and the subsequent asset record made it untenable. The stronger claim — that HLX is a full Half-Life game — remains his inference, not an official statement.

Established. The codename appears in public Source 2 builds from 2020-10-09 onward. By 2024 the associated material included HEV suit strings, off-world references, creature names, vehicle simulation, and Hammer tooling. A save-format or single engine flag would not accumulate that surface area. Tyler’s own framing in #0 was “a fully-fledged non-VR Half-Life game,” labeled Speculation.

Position history

Assessment. The negative claim (“not merely a save format”) is well supported by the breadth of HLX-labeled content. The positive claim (“a full game”) is still an inference from that content. The distinction matters: a project can have extensive engine work without being the game people expect.

What would settle it. A shipping build, store page, or Valve statement. Until then, describe HLX as “a Half-Life-shaped project,” not “confirmed Half-Life 3.”

Sources. HLX Files #0 · Every Last Detail Ep1

C03. HLX is open world

Status: Debunked. HLX is better described as “open linear”: wide, open-ended levels with procedural activities, bounded by Source 2’s technical limits. It is not a seamless open world.

Established. The open-world claim did not originate with Tyler. In HLX Files #2 (2024-09-12) he explicitly corrects a public claim attributed to Gabe Follower, saying Half-Life 3 is not open world or pseudo-open world, but a linear game with much wider levels. In Ep1 (2026-01-29) he supplies the technical argument: Hammer’s walk limit remains 16,384 units, and Source 2’s baked-lightmap architecture makes a seamless open world impractical. He compares the intended scale to Doom 2016’s Foundry or open-ended Uncharted 4 levels.

Position history

Assessment. This entry corrects a mistake in the earlier index table, which implied the open-world reading persisted into 2025. Tyler rejected it within five weeks of the first leaks. The technical argument is stronger than the terminology: the 16,384-unit limit and lightmap dependency are checkable engine facts, while “open linear” is a design comparison.

What would settle it. Map dimensions and streaming data from an HLX build, or gameplay footage showing contiguous traversal. Neither exists publicly.

Sources. HLX Files #2 · HLX Files #5 · Every Last Detail Ep1

C04. HLX uses voxel-based destruction

Status: Debunked — and this was Tyler’s own claim. This is the clearest case of a position that reversed twice and is now rejected.

Established. Tyler argued in 2024 that HLX environments were voxel-based and destructible, tying the work to the cancelled Arty project. By February 2025 he said the voxel-destruction systems appeared to belong to a different project. In Ep1 (2026-01-29) he rejected voxel geometry and Hammer support outright and noted that create_iso_surface, present in 2021 builds, is absent from current ones. Miro contains the underlying string work but does not overturn the debunk.

Position history

Assessment. The 2024 claim was never a single artifact; it was an interpretation of Arty-adjacent work that Tyler later separated from HLX. The 2026 debunk is stronger because it points to an absent command and absent Hammer support. Treat “voxel destruction” as a discarded hypothesis, not a leak that was later confirmed or denied by Valve.

What would settle it. An HLX build containing voxel/ISO-surface tooling or destruction code, or a shipped game demonstrating it. Current evidence points the other way.

Sources. HLX Files #2 · HLX Files #5 · HLX Files #9 · Every Last Detail Ep1

C05. HLX is a VR game

Status: Debunked by an official statement. Valve’s Steam Frame team said on 2025-11-12 that it built no VR game or demos for the headset, and Tyler reports that this covers HLX. The design evidence also points away from VR.

Established. Tyler said HLX was non-VR as early as HLX Files #0 (2024-08-04), and repeated it in #1 (2024-08-25), while treating Valve’s Deckard headset as a separate hardware project. Ep1 adds that VR support was stripped from Source 2 and Dota 2, and that HLX’s design — vehicles, gravity shifting, conventional FPS movement — is not VR-first. Miro #0809 preserves the “VR sequel to Alyx” hypothesis apparently as a note to debunk, not as evidence.

Position history

Assessment. This is the most firmly closed design question in the corpus. The official statement does not name HLX, but it removes the product Valve would have been building it for. The residual possibility — a optional VR mode — is not supported by any artifact and should not be presented as likely.

What would settle it. A Valve statement naming HLX and VR together, or a build with VR binaries. Nothing in the corpus suggests either.

Sources. HLX Files #1 · HLX Files #16 · Every Last Detail Ep1

C06. HLX has a day/night cycle

Status: Debunked. The Source 2 day/night code that circulated as HLX evidence belongs to Dota 2. No HLX artifact shows a cycle.

Established. Ep1 (2026-01-29) is the only video in the corpus that addresses this directly. Tyler identifies the day/night system as Dota 2’s, and pairs the debunk with the lighting argument: HLX is a baked-lightmap project, which is hostile to a dynamic time-of-day system. The claim had circulated in community datamine discussion; Miro does not provide an HLX-specific counterexample.

Position history

Assessment. The negative finding rests on one video, but it is a specific attribution rather than a general denial: the code is traceable to another Valve game. Baked lighting is consistent with the conclusion. Treat as debunked, with the caveat that a scripted lighting change in a specific level is not the same thing as a global day/night cycle.

What would settle it. An HLX map or entity graph with time-of-day logic.

Sources. Every Last Detail Ep1

C07. HLX ships or is announced in 2025

Status: Superseded. 2025 passed with no announcement. Tyler’s latest prediction (#20, 2026-10-22) is that there will be no HLX announcement in 2026 either. No official date or window exists.

Established. The timeline of expectations is well documented:

Assessment. Every date in this chain is a prediction. The reporting about other people’s expectations (#19) is secondhand and unnamed; the prediction itself is Tyler’s. The only safe statement is the negative one: no official date exists, and the 2025 window has closed.

What would settle it. A Valve announcement, press release, store page, or rated build. Nothing else.

Sources. HLX Files #1 · HLX Files #7 · HLX Files #16 · HLX Files #19 · HLX Files #20

C08. HLX’s vehicle entities were deleted

Status: Nuanced. The shorthand “vehicles were cut” is wrong. An early vehicle schema exposed in Dota 2 in June 2022 was removed after dataminers found it; the vehicle system itself continued to evolve.

Established. Ep2 (2026-03-01) describes a vehicle schema and related entities added on 2022-06-09 and deleted in subsequent updates. Miro #1486 confirms the removal and calls it “likely a mistake.” But Miro #1487 records further schema changes on 2022-11-04 and 2023-03-13, including CWheeledPhysicsVehicleVData, and #0378 describes a full vehicle system in base FGDs, with #0410–#0544 adding passengers, roles, animations, destructible parts, and suspension parameters.

Position history

Assessment. Two different layers are being conflated: the accidental public exposure in Dota 2, and the ongoing development of the system. The first was removed. The second did not stop. Public summaries that say “vehicles were cut” are repeating the first layer as if it were the second.

What would settle it. A build-to-build diff of CBaseVehicleVData / CWheeledPhysicsVehicleVData and the base FGD entries across 2022–2026.

Sources. Every Last Detail Ep2 · index §7.2

C09. Animgraph 2 shipped in CS2 viewmodels only

Status: Updated. Ep6 (2026-04-16) reported AG2 in CS2 viewmodels from 2025-07-29 and said player models were “planned.” Miro #0857 records that CS2 moved all systems to AG2 on 2026-04-01 and removed AG1 entirely.

Established. AG2 is Valve’s next-generation animation graph system, already used across Deadlock’s heroes and progressively adopted in CS2. Ep6 documents the architecture and the viewmodel rollout. Miro, which is later on this topic, closes the open question about player models.

Position history

Assessment. This is not a contradiction; it is a snapshot taken before a later update. The important correction is that AG2 adoption in CS2 is now complete, which strengthens the case that AG2 is production-ready across Valve’s projects. HLX-specific AG2 use is still inferred from nav-link variants and graph tooling, not from an HLX-tagged graph.

What would settle it. An HLX-tagged AG2 graph, skeleton, or clip in a public build.

Sources. Every Last Detail Ep6 · index §7.2

C10. The Final Hours’ 2018 small team was HLX

Status: Corrected. The small-team passage in The Final Hours of Half-Life: Alyx describes the lineage that became Citadel → Neon Prime → Shadow Line → Deadlock, not HLX.

Established. The book describes a small team exploring a new project around 2018. A common reading assigned that team to HLX. Ep1 (2026-01-29) reassigns it to the Citadel lineage that eventually became Deadlock. The codename chain itself is Tyler’s reconstruction and needs independent follow-up.

Position history

Assessment. The correction matters because it changes the project’s origin story: HLX is not a 2018 skunkworks that survived, but a later project built on shared Source 2 work. The chain needs a follow-up because the summaries do not provide a single artifact that maps all four codenames in order.

What would settle it. Valve casting/production records, or a dated internal document mapping the codenames.

Sources. Every Last Detail Ep1

C11. A chapter was rebuilt during the semi-reboot

Status: Retracted. Tyler no longer believes his earlier explanation that a specific chapter was rebuilt because it was weaker than the rest of the game. He declines to state the replacement claim.

Established. The original chapter-rebuild explanation appears in earlier coverage that is not identified by title in the corpus, which is itself a gap. In #19 (2026-07-06) Tyler says he is “not under that impression anymore,” based on people he has spoken to, and does not repeat what he was told. He still describes HLX as having survived a semi-reboot and as occupying an entire floor of Valve’s offices — both unnamed-source claims.

Position history

Assessment. The retraction is more informative than the original claim: it tells us the semi-reboot story is unstable even in Tyler’s own sourcing. Do not pair “semi-reboot” with “chapter rebuilt” as if they were one finding. The floor claim is separate and equally unsourced.

What would settle it. Named reporting on the reboot, or a Valve statement. The corpus cannot supply either.

Sources. HLX Files #19

C12. HLX started in late 2019, or was greenlit around Alyx’s ending

Status: Unresolved, probably compatible. The two statements describe different events: work “taking shape” versus a formal greenlight.

Established. Miro #1182 says HLX started taking shape in late 2019, while Alyx’s ending was being finished. Ep1 says the project was greenlit around completion of Alyx’s ending and was in production before 2020-09-22 (the Sandbox access date). Alyx shipped in March 2020, so “around the ending” can mean late 2019 to early 2020.

Position history

Assessment. No artifact dates the decision. “Late 2019” and “around the Alyx ending” overlap enough that the disagreement may be semantic. The defensible statement is: HLX was underway by late 2019/early 2020 and demonstrably in production by September 2020.

What would settle it. A dated greenlight document or a Valve production timeline. Neither is public.

Sources. Every Last Detail Ep1 · Miro report

C13. VData work began in late 2020

Status: Corrected (Miro note error). VData existed by 2020-08-12; the first public reference is 2020-09-04 18:42:33 UTC.

Established. Miro #0891 says work on VData began in late 2020 and cites September 5, 2020. That note is internally inconsistent with its own dates. Miro #1178 dates the first public reference to 2020-09-04 18:42:33 UTC in a Dota 2 update; #0010/#1189 date the VData assettype image to 2020-08-12. Ep1 uses the same chronology.

Position history

Assessment. Use the precise timestamp and the earlier existence date. The September ⁴⁄₅ difference is most likely UTC versus local date. The “late 2020” phrase should be treated as a drafting error, not a competing timeline.

What would settle it. The original assettype metadata and the 2020-09-04 Dota 2 build. Both are checkable in principle.

Sources. Every Last Detail Ep1 · index §7.2

C14. HLX is Half-Life 3

Status: Unconfirmed. This is Tyler’s identification, stated with increasing confidence, but Valve has never confirmed it. It is the strongest open claim in the database and the one most likely to be misquoted as fact.

Established. The codename HLX appears in public Source 2 builds from 2020-10-09. The associated content is Half-Life-shaped: HEV suit references, Xen fauna, Combine off-world material, gravity shifting, and Half-Life-specific weapon names. On 2026-10-21 a Valve audio panelist said “HLX” out loud at a conference, which Tyler presents as the first direct public utterance of the codename. That confirms the codename is real and known inside Valve; it does not confirm the title.

Position history

Assessment. The evidence supports “a Valve project codenamed HLX with Half-Life content.” The step from that to “Half-Life 3” is an inference about title and scope. It may be correct — Tyler has tracked the project for years — but it is not verified. Every page in this database should keep that distinction visible.

What would settle it. A Valve announcement, store page, rating, or official reference naming Half-Life 3 and HLX together.

Sources. HLX Files #0 · Every Last Detail Ep1 · HLX Files #20

C15. The 4chan art, the Shapiro story, and ARG 197 are HLX clues

Status: Debunked. Three separate episodes, all rejected:

• 4chan concept art (#15, 2025-10-29; #16, 2025-11-18): judged AI-generated and fake.
• Mike Shapiro “Prepare for the Valve opening” Reddit story (#15): Tyler says flatly, “Mike didn’t say that.”
• “ARG 197” (#8, 2025-01-31): Tyler rejects the numerology around Shapiro’s New Year post and says Valve is not running an ARG. Valve announces projects by social post, following the Alyx precedent.

Established. The debunks are direct. The 4chan images were assessed as AI-generated; the Reddit anecdote was denied; the 197 pattern was catalogued and then rejected. None of these should appear in a leak timeline except as debunked items.

Position history

Assessment. The pattern is worth naming because these items keep recirculating in summaries. The correct framing is “debunked during the corpus,” not “early speculation that was later confirmed.”

What would settle it. Nothing — these are negative findings. Reopening them would require a new primary artifact.

Sources. HLX Files #8 · HLX Files #15 · HLX Files #16

C16. The Raising the Bar reprint is a timing signal

Status: Superseded for Raising the Bar. A different book’s reprint — The Final Hours of Half-Life: Alyx — is still read by Tyler as a signal, and that variant remains speculation.

Established. In #15 (2025-10-29) Tyler speculated that a Raising the Bar reprint was timed to Half-Life hype. In #17 (2026-02-02) he noted that the 2025 deadlines, the reprint, and the Shapiro tease did not materialize. In #19 (2026-07-06) he describes Valve failing to get the Final Hours reprint out in time and calls that, in his opinion, a signal that something was planned for 2025. These are separate publishing artifacts.

Position history

Assessment. Do not merge the two books. The Raising the Bar reprint signal failed. The Final Hours reprint observation is a separate, still-unresolved speculation. Both are weak compared with a store listing or announcement.

What would settle it. A publisher’s dated announcement or a credible editorial account of the reprint schedule.

Sources. HLX Files #15 · HLX Files #17 · HLX Files #19

C17. HLX is the Steam Machine’s flagship

Status: Unresolved. Ep1 describes the pairing as a pattern-based expectation rather than evidence. #19 later theorizes that Valve decoupled HLX from the Steam Machine minimum spec because the machine is weaker than a base PS5.

Established. Valve announced the Steam Machine, Steam Frame, and Steam Controller on 2025-11-12 for 2026. Prices were announced and reservations opened in September 2026. No HLX bundle, launch title, or platform statement exists. Ep1 explicitly labels the flagship idea as inference.

Position history

Assessment. The two positions are not contradictory: a flagship launch was one hypothesis; decoupling is a later attempt to explain the absence of one. Neither has artifact support. Treat the Steam Machine as context, not evidence about HLX timing.

What would settle it. A Valve bundle, launch lineup, or platform announcement naming HLX.

Sources. HLX Files #16 · Every Last Detail Ep1 · HLX Files #19

C18. HEV suit strings are stable evidence

Status: Nuanced. The shorthand hides two different renames.

Established. Two renames are being conflated:

1. HLX strings were renamed to “Burbank suits” in the weeks after HLX Files #0 (#1, 2024-08-25), along with the Xen gorilla removal and Houndeye rename.
2. Deadlock HEV references were also removed/renamed to “Burbank suit” (#10, 2025-04-19).

Meanwhile Ep2 (2026-03-01) still finds legacy HEV sound references in the engine and expects the suit to appear. So the string is scrubbbed in some places and still present as audio in others.

Position history

Assessment. “Burbank suits” is a real string, but it is not proof the suit was cut. Valve scrubs identifiable strings after leaks; audio assets lag behind. Missing names are weak negative evidence. The suit remains likely; the string evidence is unstable.

What would settle it. A current build showing the player model, suit logic, or HUD references, or a shipped game.

Sources. HLX Files #0 · HLX Files #1 · HLX Files #10 · Every Last Detail Ep2

C19. Barney Calhoun is a main character

Status: Unconfirmed. Tyler’s only recurring story claim is that Barney is “likely a part of it” (#9, 2025-02-26). A stronger version — Barney as sidekick while Alyx is captured — appears once as a rhetorical aside (#8, 2025-01-31) and is never repeated.

Established. Nothing in the HLX material summarized here names Barney in a model, voice file, script, or casting record. The #9 claim is offered while recapping what HLX “appears” to be, with no artifact. The #8 aside is explicitly rhetorical (“did you know that you’ll probably be walking around…”). Tyler’s other Barney references are about Mike Shapiro’s voice work, not HLX content.

Position history

Assessment. The downgrade from #8 to #9 is itself the finding: the specific scenario was not maintained. Missing names are weak evidence because Valve scrubbed HEV strings in the same period. Treat Barney as plausible but unestablished; do not build story summaries around him.

What would settle it. A Barney model, voice line, script reference, or casting record tied to HLX.

Sources. HLX Files #8 · HLX Files #9

Claims deliberately not promoted

These recur in community discussion but are not in the register because the corpus does not support even a speculative entry:

• A specific release date or pre-order window.
• Alyx’s rescue, Gordon’s role, or any chapter structure.
• A confirmed enemy roster or map list.
• A confirmed console version.
• A confirmed co-op mode.
• A Valve ARG or alternate-reality campaign.
• Any claim that “White Sands” is HLX.

If a future video supplies a named artifact for one of these, add it as a new C-entry rather than reviving an old one silently.