Hacker News Reader

Comment by buboard | Hacker News Reader

Self driving terraforming robots would be an excellent testbed for AI technologies. What is the reason for pushing humans again?

View on HN

The discussion around humans vs robots in space in historically deep conversation with scientific, technical, and philosophical aspects. Do not simply brush it off without understanding the context.

that doesnt answer my question

AI doesn't work nearly as well as you seem to think it does. For something as excruciatingly difficult as a space mission, humans have proven to be well up to the task. The headline of this article here is one illustrative data point.

Furthermore, many methods that are currently classified as "AI" act because of very complex and often opaque emergent behavior, and we often have a hard time (or don't know at all) why a CNN for example behaves as it does, or even how exactly it behaves. Do you want something that you neither understand nor can predict to perform a crucial job in your space mission?

If space exploration were a robot-first, robot-only affair, we would probably have cheaper, better, far more capable robots today... And also more humans in space then the 3 that currently occupy the ISS.

actually, since the end of the cold war, space exploration is a robot-first endeavor (and apart from ISS robot-only -- think of the countless exploratory missions, hubble, the rovers etc). the ISS barely counts as space exploration

Define "robot"?

This is one of those classic situations where the "why not both" meme is really appropriate. AIs and Robots and Humans working together.

the cost of sending humans is disproportionately larger

The cost of failure is even higher.

the mars rover missions cost $2.5 billion. sending humans has an estimated cost 500 billion

you do realize the mars rover was supposed to last 90 days? and we re still receiving data from the voyagers. any robot missions to mars will still be remotely-orchestrated, however AI can be used to manage the time lag between earth and mars

I'm not sure how the supposed duration of the mission relates to AI? Also, how would AI help with the time lag?

> I'm not sure how the supposed duration of the mission relates to AI?

i was responding to the assertion that humans are more 'up to task' for space exploration

> Also, how would AI help with the time lag?

giving the robots some autonomy to navigate / build between commands would speed up their mission and making up for the 15 minutes delay

We absolutely should make extensive use of robotics on their own and as preparation for manned missions. Stuff like sending ships and fuel factories in advance of people, to place fully-fueled ships for return missions, habitations, etc. for them on arrival.

Manned missions still make sense for two reasons: 1) humans are (presently) more versatile than computers/robots, and this will probably continue to be the case for 20+ years (although one human can do the work of many through technological augmentation), and 2) the emotional/sociological/etc. value of humans actually being there (plus in the longer term, actual settlement).

I'd like to be approximately the 1000th person to move to Mars, sometime in the next 15-30 years.

Reply on news.ycombinator.com