AI & Robotics
AI cybrid (AI persona in a grown human body)
A human body grown from a person's DNA and driven remotely by an AI that has reconstructed that person's personality.
In shortAI cybrid (AI persona in a grown human body), from Hyperion (Dan Simmons, 1989), is at technology readiness level 2 of 9 (technology concept formulated) at its weakest link. Held back by Full-term ectogenesis. Science status: consistent with known physics; key effects unproven at scale. Estimated new funding to unlock: $625B–3.6T.
In Brawne Lamia's tale, her client and lover Johnny is a cybrid, a reconstruction of the poet John Keats running in the TechnoCore and inhabiting a body cloned from Keats's DNA, linked by the datasphere.
Closest todayLanguage models that imitate historical writers from their texts, and teleoperated humanoid robots; growing an adult body or controlling one remotely is not possible.
Dependency stack
Research TRL 1–3Development TRL 4–6Deployment TRL 7–9
Sort by
- criticalArtificial general intelligenceThe persona must think and act as a full person, not a chatbot
✕3 blockers›
- Reliability collapses on tasks longer than a few hours at high success thresholds
- No persistent learning from experience after deployment
- Weak grounding in physical and social context outside text and code
TRL 3- Must keep learning from experience like a human employee
✕3 blockers›
- Weight updates on new data degrade prior tasks and safety alignment (spurious forgetting)
- No method to validate a model whose parameters keep changing after release
- Per-user weight updates multiply serving cost and memory
TRL 3- Online updates and replay at frontier scale need large training compute in the loop
✕2 blockers›
- Grid power and interconnection queues
- Advanced-node chip and HBM memory supply
TRL 8
- Base reasoning, language and tool-use competence
✕2 blockers›
- Reliability and calibration over long horizons
- Continual learning after deployment
TRL 7- Training and inference computeTRL 8
- Training and running agents at human-equivalent throughputTRL 8
- Running a body remotely needs full motor output and sensory input over a neural link
✕4 blockers›
- Simultaneous stimulation and recording on the same tissue; stimulation artifact swamps spikes
- Coverage; immersive vision alone implies millions of controllable sites across V1
- Suppressing the real senses so the artificial stream is not in conflict with them
- Total implant power and heat once millions of channels are driven
plausibleTRL 3- Translating a synthetic scene into the activity patterns that evoke it
✕3 blockers›
- Models are subject-specific; decoders do not transfer across brains without recalibration
- Non-invasive signals are slow and spatially coarse; invasive coverage is a few square centimetres
- Encoding (write) models are far behind decoding; we cannot specify the activity for an arbitrary percept
TRL 3- Fast, fine-grained signals needed for anything beyond coarse category decoding
✕2 blockers›
- Tissue response degrades signal over years
- Channel count and wireless bandwidth
TRL 6- Channel count is limited by wireless data and power budget
✕3 blockers›
- Tissue heating caps implant power at a few hundred milliwatts, which bounds radio and compute
- Raw broadband neural data exceeds the link, forcing on-implant spike detection and compression
- Battery or wireless-power lifetime versus the decades a permanent interface implies
TRL 5- Implant power density sets what telemetry and on-board compute are affordable
✕2 blockers›
- Cycle life of lithium-metal anodes
- Thermal runaway safety
TRL 6
- Deep generative and language models supply the latent spaces that brain activity is mapped ontoTRL 7
- The write path for each sensory modality
✕4 blockers›
- Channel count and spatial resolution; phosphenes are coarse and do not tile a visual field
- Charge-density safety limits cap how often and how strongly tissue can be driven
- Stimulation artifacts blind the recording channels needed for closed-loop control
- No model specifies the pattern for an arbitrary complex percept
TRL 5- Stimulation needs the same dense chronic arrays as recordingTRL 6
- A full sensorium is a multi-gigabit bidirectional streamTRL 5
- Dense chronic coverage of multiple cortical areasTRL 6
- The body has to be gestated and grown without a surrogate mother
- The gap between day-14 embryo culture and artificial-placenta viability covers organogenesis, the hardest and least substitutable phase
- No artificial implantation or placental organ exists to supply hormones and nutrient exchange during that gap
- Scaling implies thousands of parallel units with sterility and uninterrupted power for nine months each
TRL 2- The early half cannot be built without reconstructing implantation and placental exchange
✕3 blockers›
- The placenta's endocrine and nutrient-exchange functions have not been reproduced by any device or culture
- Maternal immune tolerance at the interface is not understood well enough to substitute for
- Human implantation cannot be studied past the 14-day limit, so the critical window is inaccessible
TRL 3 - An engineered placenta is a vascularized tissue that must be grown, not just machined
✕4 blockers›
- Capillary-scale perfusion cannot be printed and must self-assemble, which is uncontrolled today
- Printed tissue must mature into function (filtration, metabolism, contraction), and maturation protocols stall at fetal-like performance
- Cell demand of 10^10-10^11 per organ exceeds current manufacturing
- Mechanical integration - ureters, bile ducts, valves - has no demonstrated fabrication route
TRL 4- Without a capillary network, tissue thicker than a few hundred microns dies
✕3 blockers›
- Printing resolution stops at tens of microns while capillaries are 5-10 microns, so the smallest and most numerous vessels must self-assemble
- Engineered endothelium is thrombogenic, so perfused constructs clot
- No method guides anastomosis of a construct's vasculature to host vessels at scale
TRL 3 - Banked-cell organs are not patient-matched, so acceptance must be engineered
✕3 blockers›
- Chronic antibody-mediated rejection proceeds even under current immunosuppression
- Tolerance protocols trade rejection risk for conditioning toxicity and graft-versus-host disease
- No assay predicts which recipients can be safely weaned off drugs
TRL 5 - One organ needs 10^10-10^11 clinical-grade cells of several types
✕3 blockers›
- Genomic instability and drift accumulate over the expansion needed for 10^11 cells
- Each organ needs many cell types differentiated on a synchronized schedule, not one lineage
- Cost per cell at current media and reagent prices makes a printed organ unaffordable
TRL 6
- Supplies the late-gestation half of the problem
✕4 blockers›
- Miniaturization - the lamb model weighs 1-3 kg while a 23-week human fetus is about 500 g, with shorter umbilical cannulation distances
- Umbilical vessel vasospasm and thrombosis during cannulation and over weeks of support
- Infection control in a continuously exchanged sterile fluid environment
- No agreed proof-of-concept endpoint or animal model that maps onto human neonatal outcomes
TRL 5- Sterile fluid exchange, nutrition and waste handling must run continuously without interruption
✕2 blockers›
- Food production closure
- Microbial and ecological stability over years
TRL 5 - Gas exchange through the umbilical circuit for weeks is the core function
✕3 blockers›
- Blood-surface contact activates coagulation, so runs depend on anticoagulation that causes bleeding in fragile patients
- Oxygenator plasma leakage and fiber fouling limit single-circuit life
- Hemolysis and platelet consumption scale with pump shear and circuit length
TRL 7
- Cloning a body from recovered DNA
✕3 blockers›
- Incomplete epigenetic reprogramming of the donor nucleus produces placental defects and large-offspring syndrome
- Efficiency is a few percent of transferred embryos, so each live birth consumes many donor eggs and surrogates
- Donor egg supply is the limiting input for any primate-scale program
TRL 5 - Reconstructing a voice and style from a person's writings is the part current models already doTRL 7
Who could unlock it
Full-term ectogenesisTRL 2bottleneck
- Jacob HannaResearcher, Weizmann Institute of ScienceGrew natural mouse embryos ex utero from E5.5 to E11 through organogenesis (Nature, 2021) and generated transgene-free post-gastrulation whole embryo models from naive stem cells (Cell Stem Cell, 2025); holds an ERC grant titled ExUteroEmbryogenesis
- Alejandro Aguilera-CastrejonResearcher, Janelia Research Campus, HHMIBuilt the roller and perfusion ex utero culture platform in the Hanna lab and now works explicitly toward birthing a mouse from an incubator
- Renewal BioStartupIsraeli company co-founded by Hanna in 2022 that licensed the Weizmann ex utero and synthetic-embryo patents (about $9M raised via NFX) to grow early tissue as a cell source
- Vitara BiomedicalStartupWorking the opposite end of the same problem - fluid-filled support from roughly 22 weeks of human gestation onward, the only part of full-term ectogenesis anywhere near clinical use
Artificial general intelligenceTRL 3
- OpenAIStartupBuilding Stargate with Oracle, SoftBank and MGX, with ~7 GW planned by Sept 2025 toward 10 GW by 2028; states AGI as its mission
- Google DeepMindCompanyReleased Gemini 3 (Dec 2025); Demis Hassabis frames AGI as the lab's goal
- AnthropicStartupClaude models led METR long-task time-horizon measurements in 2025-2026
- Demis HassabisResearcher, Google DeepMindGoogle DeepMind CEO and 2024 Nobel laureate (AlphaFold); directs Gemini and the lab's AGI research
Full-sensorium bidirectional brain-computer interfaceTRL 3
- Robert GauntResearcher, University of Pittsburgh Rehab Neural Engineering LabsLeads the Pittsburgh ICMS program that showed biomimetic stimulation produces more natural touch in humans (J. Neural Eng. 2025)
- Eduardo FernándezResearcher, Miguel Hernández University of ElcheImplanted 96-channel arrays in visual cortex of blind volunteers who recognised letters and shapes (JCI 2021)
- NeuralinkStartupRaised $650M Series E in 2025; has human motor-BCI participants and FDA Breakthrough designations for vision and speech restoration
- ParadromicsStartupFirst long-term Connexus implant in June 2026; participant used it for real-time speech, with sensory feedback on the roadmap
Somatic cell nuclear transfer cloningTRL 5
- Qiang SunResearcher, Institute of Neuroscience, Chinese Academy of SciencesCloned the macaques Zhong Zhong and Hua Hua (2018) and ReTro, the first cloned rhesus to reach adulthood, using trophoblast replacement (2024)
- Institute of Neuroscience, Chinese Academy of SciencesGovernment labThe only group to have produced primate clones; its 2024 result consumed 113 transferred embryos for one surviving animal
- ViaGen Pets and EquineCompanyThe commercial SCNT operation, cloning cats, dogs and horses and the Przewalski's horse Kurt; acquired by Colossal Biosciences in November 2025
- Colossal BiosciencesStartupUses cloning from edited cell lines as its route to live animals and bought ViaGen in 2025 to bring that throughput in-house
Beyond engineering
- Personhood and rights of a cloned body with no native mind
- Consent and estate rights over reconstructing a dead person