The original note, as written at the time
We could kick start the humanlike AI programmes with this function. We could have carrier androids that carry a womb which can be removed after impregnation. Another womb can be implanted and the carrier android can be leased to another family for a separate session. It can also be that each family gets one of its own.
Ongoing development
We could start the humanlike programmes with this function, and it is the one I would argue about longest.
The proposal, stripped of its cruder framing, is this: gestation moved outside the body, into a carrier that can be given a womb, emptied after conception, and leased again to a different household. The machine is a vessel; the womb is the asset; the lease is the transaction.
The technology is not the interesting part and it is not far away. The interesting part is that this arrangement would turn the one capacity human beings have never been able to commodify cleanly into a service with a tariff, a term and a return policy.
And it is worth saying why that might be wanted. Pregnancy kills women. It removes them from work, from study, from the room where the decisions are made. For a great many people the choice to lease gestation would be a genuine liberation, and any argument against the technology has to answer that first.
But look at what the leasing structure does. A womb that can be moved between households makes the capacity separable from the person, and once it is separable it can be priced, and once it is priced there will be a market, and a market will find the cheapest supplier. The cheapest supplier of a leasable womb is a poor woman, and the moment her capacity is a line item the argument about her consent becomes an argument about her price.
That is the failure mode I would design against from the first drawing: not the machine, but the tenancy. If the womb is leased, who holds the lease, who may terminate it, what happens to the child if the lease ends mid-term, and what the household is allowed to do with a womb it does not own.
We keep building the capability and leaving the tenancy to be worked out by whoever profits from it. That is how we ended up with the rest of the internet.