/sci/ - Science & Math

AI is making huge strides lately, as are humanoid robots for use in the home and workplace. But research into self-replicating machines seems to have stagnated since the mid 1990s. What are the principle roadblocks inhibiting advances in this area?

>>9136381
Abso-fucking-lutely
>>9136437
>> wouldn't be made of nuts and bolts
Bullshit, they damn well could. We could have robots capable of assembling copies of themselves from parts and then the machines to make those parts.

>>8987322
We need to build self replicating machines

>>8938857
Self-replicating machines with no replication limits.

>>7677276

Good job, you reinvented Von Neumann machines 60 years late

What career tracks are relatively automation-proof? Which jobs will get easier to do because of developments of stronger AI, instead of disappearing completely and rendering one's education useless?

Thread inspired by that one about incoming automation of anesthesiology.

>>6488929
>implying

Will there ever be useful Von Neumann machines?

>>3314282
Of course it cannot, though it does pass the requirements for a feasibility demonstration called for in the NASA study.
See fertility test
http://www.islandone.org/MMSG/aasm/AASM56.html

pic related

Gentlemen.

By now it should be clear to everyone that the very raison d'être of our species is to flood the galaxy with self replicating machines and fuck over any other intelligent life out there with the sheer brunt of exponential growth.

However, how do we even begin? Self replication is a difficult thing to achieve. Yes, the components are right there, in the asteroids and planetoids across the galaxy, but it's a long way from a floating chunk of iron rock to the intricate circuitry necessary to form a new probe.

Discuss.

A robot that builds other robots.