ai gen
Editing the human · a 2026 field map

What's
Missing

The biohacking ecosystem isn't a catalog of what exists. It's a map of what's under-built, what's needed, and who should build it — read on a clock that AI may be winding fast.

Read the map ↓
20–45%
P(transformative AI by 2030) — the honest range from two independent models. High enough that every timeline below should condition on it. This page does: flip the switch in the nav between AGI ~2030 and a slow world to see how the whole landscape re-weights. You're viewing the AGI-this-decade lens.You're viewing the slow-world lens.
01

The Money Isn't
Where the Goal Is

Read as flows rather than a catalog, the ecosystem has one defining feature: capital, talent, and attention cluster in the scenes furthest from the stated goals, while the decisive bottlenecks sit under-owned and cheap. The loudest claims attach to the thinnest data; the things that would actually move the needle have almost no one on them.

Where the capital clusters

well-funded · mostly off the critical path
Longevity / "don't die" VC~$6–8B/yr
Cellular reprogramming (Altos, Retro, NewLimit)~$4B raised
Brain–computer interfaces~$2.3B/yr · Neuralink $9B val
Supplements / wearables / "optimization"huge consumer mkt

Where the leverage actually is

decisive · under-owned · cheap
Whole-brain-emulation field (total people)<500 people
The "which-details-are-you" sufficiency test~$0 funded
Brain-preservation R&D (Nectome, total raised)~$1M
Vascularisation / perfusion toolingtiny vs reprogramming

Numbers are directional (vendor estimates disagree by ~2×) but the gap is orders of magnitude, not percentages. That asymmetry — billions chasing the goals' periphery, ~$1M on a literal pillar of mind-uploading — is the single most actionable fact on this page. It means the highest-leverage moves are cheap and un-crowded. Detail: doc 14 →

02

The Gap Register

Nothing here is fixed. These are the things that are missing, under-built, unconnected, or unfunded relative to their leverage — each with what's needed and who could build it. Many are connect-what-exists, not invent-something-new. Sorted by tier: S = highest leverage × tractability, then A, B.

03

Two Worlds,
One Decision

The biggest variable isn't in any lab — it's whether transformative AI arrives this decade. The earlier analysis (and the "Biopunk 2037" worldbuilding) quietly assumed it wouldn't — one even assumed an AI freeze. That's a load-bearing omission. Condition on it instead and the map re-weights hard. Below: the four futures, and what each means for where to put effort. [ viewing: AGI ~2030 ][ viewing: slow world ]

AGI this decade · aligned

  • Uploading: the ASI cracks connectomics & the "which-details" science fast — bio-science is leapfrogged. P(FiO) ≈ P(aligned ASI) × P(you survived with brain intact).
  • Biopunk: compresses into a bridge; ASI-designed repair may make growing whole bodies moot.
  • Lever: survive to ASI (preservation) + raise P(alignment).

AGI this decade · misaligned

  • Uploading: never happens for humans, or happens coercively. FiO becomes a warning, not a goal.
  • Biopunk: mostly irrelevant if humans lose control.
  • Experience machine: the high-S-risk branch — the wrong machine, built well.
  • Lever: almost entirely AI control / governance.

No near AGI · aligned later

  • Uploading: human-led connectomics & preservation stay slow but meaningful; preservation is a bridge.
  • Biopunk: matters directly — organs, xeno, regenerative medicine are the central 2030s projects.
  • Lever: the bio-science bottlenecks below.

No near AGI · prolonged muddle

  • Uploading: speculative for decades; preservation a long-shot.
  • Biopunk: this is the world where "Biopunk 2037" matters most — gated by capital, regulation, hard biology.
  • Lever: advance bio-science directly; build institutions; reduce mortality.
The single highest-leverage move

If you put real weight on AGI by ~2032, it stops being "advance the bio-science" and becomes (1) survive to ASI with your brain's structure intact — preservation now, the one thing an ASI can't fix retroactively — and (2) raise P(aligned ASI). Almost everything else is a bridge, or gets redone better by the ASI. Preservation + alignment are the only moves robust across all four futures.

In a slow world, leverage is in the under-owned bottlenecks: the small-organism sufficiency test, vascularisation/perfusion tooling, and brain-preservation logistics. The bio-science is the work, and it's where a motivated individual or funder can still move the frontier. Preservation logistics + alignment hedge well into the other futures too.

Full reasoning, the probability scorecard, and the independent gpt-5.5 corroboration: doc 13 →

04

The Three Goals,
Re-weighted

Not static maturity scores — what's needed to get there, and how the odds move when you flip the world-switch.

I

Uploading

"Friendship is Optimal"
1–5%≈ P(aligned ASI)full FiO this centuryFiO requires an ASI by construction

Bottleneck: data acquisition + the unsolved "which details are you" science. Compute isn't the limit. Need: connectomics throughput + a sufficiency test + preservation.

Bottleneck flips to alignment + whether you survived with brain structure intact. The ASI handles the bio-science. Preservation now is the one rung it can't fix retroactively.

II

Biopunk 2037

grown organs & bodies
3–10%leapfroggedwhole-body cultivation by 2037ASI-designed repair may make it moot

Organ-supply spine (xeno, bioengineered, ectopic) is real & in human trials. Whole bodies are walled off by de-novo ectogenesis. Two models lowered the report's 15–25% to ~3–10%.

The organ spine compresses into a bridge; the whole-body moonshot is the element most likely obsoleted before it's finished — ASI-designed repair, or substrate change, instead of growing spares.

III

Experience Machine

Nozick, built for real
VR + dreams= Goal Itractable stack today; full machine distantcollapses into FiO

Build the external stack now — best XR + body-ownership + lucid-dream / Dormio-style incubation — designed around agency & reality, not pleasure (wanting ≠ liking).

An aligned ASI building a wish-fulfilling world for minds is the experience machine. An unaligned one builds the wrong one. There's no separate project left — only a pre-ASI stopgap.

05

The Scenes

The substrate the gaps sit on — eight overlapping scenes, from garage grinders to $3B reprogramming startups. Dot = maturity center of gravity (demonstrated · emerging · speculative · ideology). Each links to its full dossier.

06

Two Models,
One Verdict

Every conclusion here was researched twice — once by Claude, once from scratch by gpt-5.5, neither seeing the other. Four independent passes, and they converged on every load-bearing point — including lowering the author's own worldbuilding optimism. That's the honesty signal.

Convergent across both models (high confidence)
  • Compute isn't the uploading bottleneck — data + "which details matter" is
  • Brain preservation is the only personal lever; its sufficiency is unproven
  • Biopunk's organ-supply spine is real & on-trajectory
  • Whole-body cultivation ≈ 3–10% (report's 15–25% too high)
  • P(transformative AI by 2030) ≈ 20–45%, central ~30%
  • Under AGI: preservation + alignment dominate; bio-science is leapfrogged
  • The experience machine collapses into uploading
  • A pure pleasure/wirehead machine is the wrong target (wanting ≠ liking)

Full reconciliation: doc 12 →

07

The Full Dossier

Every section above distills these documents — now rendered as HTML, with source URLs and confidence tiers inside.