My Advanced Realistic Humanoid Robots Project

[MK14]

did is just copy-paste walls of text from the first thread, which was copy-paste from some other web site.

So that anyone who is not familiar with their long term, large project(s), will know what is going on.

In the future, they can put fresh/new project information in here, for discussion.

[artbyrobot]

By popular demand, here is some math I did regarding the motor and pulleys for the finger actuation.

64:1 downgear ratio

24 inches total draw onto motor shaft 24 / 64 = 0.37" draw at finger joint

2430 motor 5900kv at 12v RPM = kV * V RPM = 5900 * 12 RPM = 69600
69600 / 60 = 1160 revs/second 1160/2 = 580 revs / half second
580/2 = 290 revs / quarter second

if motor reels around 1cm / rev then in quarter second it reels 290cm...
and 30cm = 1 foot so 290/30 = 9.6ft/quarter second maybe it only reels 3/4 of
that? even so... around 9.5ft/quarter second - and quarter second is the
speed of a human finger moving... we only want to reel 24 inches... and
it is reeling 9.5ft so if it only reeled 24 inches that would be human
speed... so if it only reeled 60cm that would be human speed... but it
reels 290cm... around 4.8x human speed!

now for strength at this 64:1... an online google search said a 2430 motor
can pull 60 g cm... 120 g at 1/2cm 240g at 1/4cm maybe we are around
between 1/4cm and 1/2 cm away from shaft of motor on average... so 190g at
that distance... 190g is 0.42lb... 0.42 lb * 64 = 27lb

so a single finger joint can do 27 lb dumbell curls ALONE - well wait since
it's lifting a lever at the joint, it is much lower than this maybe 1/5
of this so 5.4lb dumbel curl is more realistic...

now this is all for torque at efficient natural movement speed...

what about stall torque - IE how much can it just HOLD in place like rock
climbing dead weight it can't move but can hold steady?

it's stall torque is around 280 g.cm compare that to its normal torque
of 60 g.cm so 4x... so it can HOLD steady around 20lb! that is about what
my finger can hold steady for a single finger tip!

[ebastler]

Thanks for the calculations. I believe you have made a couple of errors there, but maybe just misunderstood:

(a) In your speed estimates, you use the 5900 rev/V * 12V which apply to the unloaded motor. You will need to find (or measure) RPMs under load and use these.

(b) I understand that taking up string on a small drum is what you want to use for the 64:1 speed reduction -- there is no separate gearbox, right? If that's the case, where do you take the factor of 64 from in your strength calculations? If the motor provides 60 g*cm of maximum torque, then it can pull on the string with 120 g if the drum diameter radius is 0.5 cm. That's it, there is no further amplification.

You will need to complement these calculations with estimates of the required stability of the joints, strings, pulley bearings etc. -- how much force do they take, what longevity can you expect, what precision do you need?

Also, what electrical power are you looking at? That caliber of motor takes ~200W peak I think. Do you have an estimate of the average duty cycle of all motors in total when the robot is active, and hence average power consumption and heat disposal?

[artbyrobot]

@ebastler

You say "In your speed estimates, you use the 5900 rev/V * 12V which apply to the unloaded motor. You will need to find (or measure) RPMs under load and use these." --- that's true.  It's the only specs basis to go on though from the manufacturer so its what I went by.  When under load, it will move slower true.  But knowing the no load speed is a good starting point for estimating max speed and getting an idea of what we are working with.

You say "there is no separate gearbox, right?" --- yes and no.  There is no gearbox of metal gears, however, there is external downgearing by way of pulleys.  The intended downgear is 64:1.  I have a couple different pulley systems I'm prototyping now. 

You say "Also, what electrical power are you looking at? That caliber of motor takes ~200W peak I think." --- I think that's correct. 

You say "Do you have an estimate of the average duty cycle of all motors in total when the robot is active, and hence average power consumption and heat disposal?" --- no.  Sometimes I toy with these numbers but it all depends on what is going on and for how long.  Like when walking only one leg and only parts of that leg are driving off the ground at any given time and the other leg is swinging.  Often times most muscles are at rest for the most part.  So only when doing very strenuous activities would significant numbers motors be activating at one near peak.  And even then it is not going to be anywhere near the total number of motors in the body all going at once.  So getting access to average power is activity specific and beyond the scope of my early calculation needs however will be gathered in future testing and way later down the line.



@everyone  thanks for the thoughts and posts.  I read everything I assure you.  I am limiting my responses to be extremely on topic and extremely technical focused on the build itself.  Anything tangental and more conversational I'm avoiding to demonstrate my dedication to honor the guidelines and not make the same mistakes I made in my last thread.  Hope you understand.  If you REALLY want my answer to something, PM me and I will happily answer there but for here I'm staying very technical as a precaution.

[SteveThackery]

That's one of the goals of the Projects section, just showing off / documenting your projects. It doesn't have to involve any questions or feedback.

Fair point - thanks for the reminder.

[SteveThackery]

You say...
You say...
You say...
You say...

Please, @artbyrobot, might I respectfully request that you use the proper quoting system? It makes long posts like yours so much easier to read. It's slightly less convenient but still very easy to use.

[ebastler]

You say "there is no separate gearbox, right?" --- yes and no.  There is no gearbox of metal gears, however, there is external downgearing by way of pulleys.  The intended downgear is 64:1.  I have a couple different pulley systems I'm prototyping now. 

Ah, ok, so there is another reduction stage after the little drum winch. In that case, did you forget the 64x reduction in your speed estimate?

Also, a pulley system for 64x reduction amplifies my concerns about reliability of the pulleys, bearings & strings. Have you done some endurance testing with prototype implementations?

[artbyrobot]

@ebastler

You say "Ah, ok, so there is another reduction stage after the little drum winch. In that case, did you forget the 64x reduction in your speed estimate?" --- No.  The revolutions per second gives me speed and the resulting distance of string travel over that duration of time.  The string travel distance is reduced by the pulleys from being a longer travel distance over the given time interval to being a short travel distance over that same time interval.  So over that same time interval the finger moved way less than the original string travel at the drum.  Hope that clarifies how my calculations were correct.

You say "Also, a pulley system for 64x reduction amplifies my concerns about reliability of the pulleys, bearings & strings. Have you done some endurance testing with prototype implementations?" --- I've done some no load tests of a 16:1 downgear pulley system and fixed some issues with its construction quality.  I then moved onto a 10lb load test and got line breakages because of some more construction quality problems.  I'm learning how to construct properly by this trial and error testing.  That's where I've left off.  Much more testing and refinement to come as well as trying out some alternative pulley system styles entirely.  Also 16:1 is not enough I want at least 32:1 and 64:1 is even better/ideal IMO.

[Zero999]

The problem is. You haven't asked any specific questions.

I think he intends this as a show & tell, or a project log.
 

That's one of the goals of the Projects section, just showing off / documenting your projects. It doesn't have to involve any questions or feedback.

Fair point, but I don't believe it's against the rules to offer advice, criticism and ask questions.

If we (as a community), want this forum to remain a nice, open and safe place.

What do you mean by safe space? If you mean a place where there won't be any difficult questions or criticism then I believe you're mistaken. There are no safe spaces on the Internet.

[ebastler]

You say "Ah, ok, so there is another reduction stage after the little drum winch. In that case, did you forget the 64x reduction in your speed estimate?" --- No.  The revolutions per second gives me speed and the resulting distance of string travel over that duration of time.  The string travel distance is reduced by the pulleys from being a longer travel distance over the given time interval to being a short travel distance over that same time interval.  So over that same time interval the finger moved way less than the original string travel at the drum.  Hope that clarifies how my calculations were correct.

Ah, right -- I got so lost in the unit conversions in the longer paragraphs that I overlooked the 64x reduction in the sentence before. (How do you guys manage to juggle cm, inch, feet, g, lbs etc.? Give SI units a try! )

Then my only remaining quibble with the speed estimates (besides the no-load conditions already discussed) is that 0.37" in 1/4 second seems sluggish for finger movement. I can certainly move my finger faster than 1.5" per second! Try striking some piano or typewriter keys with your robotic prototype finger, that will give a reasonable idea of the required forces and speeds.

[artbyrobot]

@ebastler

You say "0.37" in 1/4 second seems sluggish for finger movement. I can certainly move my finger faster than 1.5" per second! Try striking some piano or typewriter keys with your robotic prototype finger, that will give a reasonable idea of the required forces and speeds." --- the 0.37" was supposed to translate to a full rotation of the finger through its entire rotational range.  When you attach the tendon, you attach at the base of a bone and so the movement distance at the tendon is much less than the movement distance of the distal tip of the bone it is moving.  I think a full finger bend taking 1/4 second is human speed roughly.  In any case, if I find it's too slow I can drop down to 48:1 and pickup 25% more speed or 32:1 and double speed but the tradeoff is then you are at half he torque.

[SteveThackery]

Then my only remaining quibble with the speed estimates (besides the no-load conditions already discussed) is that 0.37" in 1/4 second seems sluggish for finger movement. I can certainly move my finger faster than 1.5" per second! Try striking some piano or typewriter keys with your robotic prototype finger, that will give a reasonable idea of the required forces and speeds.

I was wondering that myself. @artbyrobot, you assert that the human body is perfect and cannot be improved upon, so I was wondering whether you should be aiming for a degree of strength comparable with that of a human?

I think with all four fingers I can just about pull (or lift) 100kg, but let's make it 50kg to keep it realistic. That figure is with the fingers curled, of course - I haven't measured my finger force going from straight to fully curled. Shall we reduce it further to maybe 25kg?  Rounding down even more, perhaps 5kg per finger.

Does that seem like a reasonable figure to aim for? What is your current gripping force target? (Forgive me if you have already told us - I have looked and couldn't find it.)

[artbyrobot]

@steve  5.4lb finger curl where finger is going from straight to fully curled in lifting the weight actively.  20lb for finger already curled up and a weight hooked onto it and finger is just holding its curled position but not actively lifting it just maintaining the curl to not straighten out and drop it.  These were in my calculations writeup.

Also, I am pretty sure this is a dead match for my fingers strength.  So it could hold a 100lb dumbell without dropping it.  Since 5 fingers at 20lb each is 100lb total it can maintain fingers in curled position. 

NONE of this has margin for error it is just theoretical ideals of course and may not be maintainable long term without heat becoming an issue perhaps.  Although it may be able to do this if I make everything perfectly and have elite thermal management which I have designed so it's possible perhaps.  100lb is VERY hard for most people to hold for a dumbbell most people use straps at that point.  So this is high end strength for human fingers IMO. 


Actually it just hit me this:  if you look at your fingers you have 3 joints.  I have a single motor assigned to the last two joints (shortcut/lazy or w/e) so the last 2 have that 20lb thing I just talked about.  BUT the proximal most joint is another motor and that helps the distal 2 joints.  So if you factor that in, depending on how much that one adds to the overall dumbbell hold, we could be looking at maybe 50-60% higher than what I just said!  So perhaps it could hold 150-170lb dumbbell then.  Now that is some top tier world class grip at that point.  That's like pro rock climber level almost I think.  But I could be wrong on these calculations I'm estimating.


Update: Oh wow and I'm TOTALLY forgetting the thumb helping to lock down the fingers and finish the grip.  The above was all THUMBLESS.  Add in the thumb strength and we might be more like 230lb.  That is far beyond my own strength of one hand grip.  I'm capped at about 100lb with no straps.

I will say though at that point, I'd not be sure my artificial ligaments are up to the task.  Although the silicone skin would help the ligaments hold the bones together.  If I ever started getting that high I'd have to do something to heavily fortify the ligaments I think.  I might silicone impregnate them to add more elasticity and robustness in addition to the silicone skin helping hold it all together.  Maybe put a glove onto the skin and tape the fingers up really good like football players do with the white athletics tape would be good as well if we were testing max numbers like this.

[SteveThackery]

Now that is some top tier world class grip at that point.  That's like pro rock climber level almost I think.  But I could be wrong on these calculations I'm estimating.

Yes, in my experience the diameter of the object being gripped makes a very big difference. With something really thin, like the straps on a heavy bag, my limits are determined by how much pain I can stand from the compressed skin. The larger the diameter, the weaker my grip force.

Thanks for the figures you've given us re. grip strength.

[jpanhalt]

Will your humanoid hand be able to do this:


That's an extreme example.  Not everyone can do it to that extent.
Credit: https://www.reddit.com/r/mildlyinteresting/comments/jvwj58/im_able_to_bend_just_my_finger_tips_without/

[tggzzz]

Will your humanoid hand be able to do this:


That's an extreme example.  Not everyone can do it to that extent.
Credit: https://www.reddit.com/r/mildlyinteresting/comments/jvwj58/im_able_to_bend_just_my_finger_tips_without/

I could do that when young, less so now. It was a good party trick, getting people to say "What? Easy", then "Ewww"

[radiolistener]

The robot should not resemble a human, instead, it should be constructed like sticks with motors to perform mechanical tasks. The robot does not need to move freely. The robot does not require a hand like a human's, instead, it needs a manipulator with grippers, which is sufficient for performing tasks.

This is how robot should looks like:


Not like this:

[artbyrobot]

@jpanhalt

No it won't I don't think.  If I wanted that I could add 4 more motors, but decided to just do one motor for both the distal 2 joints so the last 2 joints move together and the distal most one then can't move independently like that.  For the longest time I planned to have it do just what you are showing but just running out of space so bad in my CAD I said heck with it I'll drop 4 motors and not be able to do what you just showed.  I also sacrifice grip strength by dropping those 4 motors.  But I don't think functionality will suffer by this decision as far as finger skill, only slight hit in strength which seems will still be at or higher than human level ideally with the motors I went with (although my calculations are only estimates and it may not be near as strong as I'm assuming).

@radiolistener

that's a valid point but the drawback there would be inability to send it into public and have it pass for a human at least from a distance or if nobody is paying careful attention.  One of my goals is to have it talk to people and measure how long it takes for the person to realize they are talking to a robot.  If I can get like 2-3 minutes conversation in without them realizing I'd be dancing.  It's like the ultimate turing test in a way.  I just see doing this over and over for infinite fun and it would make for some very fun youtube videos IMO.  But I mean the list of reasons for it to look realistically human go on and on for me.  It just changes the whole dynamic. 

Oh and also it looking realistically human would set this project apart and has been the key differentiator between my robot and every other one.  I saw robots that look real and saw robots that are very functional but look mech.  I NEVER saw one that looks real and has full function like the mech ones do.  I want to merge both types.  The challenge of this is huge.  The human form factor is much more limiting so it is much harder.  I think the human look is the holy grail of humanoid robotics and is the future.  I think they will all look fully human and pass for human in the future.  So I'm not wanting it to look outdated by the time its done by NOT making it look realistically human.

[tggzzz]

that's a valid point but the drawback there would be inability to send it into public and have it pass for a human at least from a distance or if nobody is paying careful attention.  One of my goals is to have it talk to people and measure how long it takes for the person to realize they are talking to a robot.  If I can get like 2-3 minutes conversation in without them realizing I'd be dancing.  It's like the ultimate turing test in a way.  I just see doing this over and over for infinite fun and it would make for some very fun youtube videos IMO.  But I mean the list of reasons for it to look realistically human go on and on for me.  It just changes the whole dynamic. 

Oh and also it looking realistically human would set this project apart and has been the key differentiator between my robot and every other one.  I saw robots that look real and saw robots that are very functional but look mech.  I NEVER saw one that looks real and has full function like the mech ones do.  I want to merge both types.  The challenge of this is huge.  The human form factor is much more limiting so it is much harder.  I think the human look is the holy grail of humanoid robotics and is the future.  I think they will all look fully human and pass for human in the future.  So I'm not wanting it to look outdated by the time its done by NOT making it look realistically human.

Given that you are after "full function" humanoid robots that can pass for human, are you aware that the consquences have been well explored by SF authors over the decades?

Are you aware of the TV programme "Humans" https://en.wikipedia.org/wiki/Humans_(TV_series) ? It is is a quite reasonable exploration of the issues. (One of the many plotlines is, ahem, a very "full function"[1] Gemma Chan)

[1] The link to your second picture is https://dollforum.com/forum/download/file.php?id=1283447 Caution: NSFW.

(Oh, I'll echo the suggestion that you use the normal forum techniques to quote other people. This is neither Farcebook not Twatter. Not being able to use appropriate techniques  for one technology (forums) will raise the question as to whether you can use appropraite techniques in other technologies. )

[artbyrobot]

@tggzzz

yes I've seen humans_tv series.  There are some great food for thought there that is relevant I agree.  The whole robots going rogue thing though, at least to me, is indicative of a unfettered AI that was irresponsibly developed to have no leash.  My AI design is EXTREMELY constrained to a very narrow worldview model, purpose model, etc where it has no way to unchain and go rogue.  I will make that impossible.  My design is a lot more hand crafted and manageable than say a LLM where behavior can be unpredictable.  I'm not using even remotely similar approach to AI as the current state of the art AI.  My entire approach is custom and gives me a TON of control and ensures a huge degree of predictability in outcomes.  It can't just leave the farm so to speak and become a free agent like that.  I want to say the way I'm making it is sandboxed.  Like you can test a virus on your computer if you put it in a sandbox.  Like a virtual machine.  My AI will be in a virtual machine made of rules and constraints and hard coded stuff it can't edit and would be incapable of editing and not have permissions to touch and there are no overrides to this enabled.  Hope that helps.  If other AI developers had this strategy, we'd have never heard about AI safety concerns.  But their approach to AI is a lot more black box and train and cross fingers than my approach.  I am not a fan of their approach but can't deny the excellent results.  Just not my style.  I'm more of a good old fashioned AI type of guy I guess, but with my own innovations and twist IMO.




@previous posters:

Someone asks how this is funded --- self funded mostly but did get some donations years back when I worked on it live.


Someone asked if the 10 years it took should be extrapolated to mean I'll never finish.  I don't think so.  The up front knowledge gap that had to be filled was enormous.  Filling that gap to be able to do my goals was step one more or less.  The more I know the more I know what to do and the more I know what to do the more I can progress.  Had I gone to school for robotics and got 3 phds - one in AI, one in electrical engineering, one in mechanical engineering, then spent a couple years in a mentorship for weta workshop or a Hollywood makeup artist, we'd be looking at 10 years and zero start on the robot but a TON of knowledge to now begin.  I see the ten years like that.  It's not really reasonable IMO to say it was 10 years of just building it and so the percent completed is the expected percent completed for the next 10.  The next 10 I am armed with a mountain of knowledge I didn't have on the first ten.  My hands were more tied then.  Also major life changes and career demands the first 10 years may be WAY less demanding than the second ten if I structure things right.  So its more nuanced than was being suggested.


The doll forum topic was on the last thread covered extensively and was part of the derailment there.  Just to TLDR it.  I'm not intending to build "capable" robots and the extent that my project overlaps with goals they have is the appearance and functions unrelated to "THAT" "CAPABILITY".  So exchanging notes with that crew makes sense to me.  Their aim is like 98% overlapping mine on a Venn diagram.  The 2% difference is "that capability", the elephant in the room which is why 2 people have mentioned that site here so far.


Ok I cautiously decided to breach these subjects being as technical as I could to at least barely touch on them.  If these launch entire conversations I may not respond or may have to switch to PMs as I still want to remain as technically focused as possible and on topic as possible and I know how things can drift off topic SO EASILY as robotics and AI SO EASILY segue into religion, politics, philosophy, and morality when I am going to try my absolute best not to AT ALL enter into out of respect for the mods and forum rules.

[Andy Chee]

One of my goals is to have it talk to people and measure how long it takes for the person to realize they are talking to a robot.  If I can get like 2-3 minutes conversation in without them realizing I'd be dancing.  It's like the ultimate turing test in a way.  I just see doing this over and over for infinite fun and it would make for some very fun youtube videos IMO.  But I mean the list of reasons for it to look realistically human go on and on for me.  It just changes the whole dynamic. 

Given that you are after "full function" humanoid robots that can pass for human, are you aware that the consquences have been well explored by SF authors over the decades?

The other notable movie made in recent years that comes to mind is Ex Machina (starring Alicia Vikander).

[artbyrobot]

@Andy

yes Ex Machina was mind blowing to me as someone pursuing similar.  Definitely got me even more hyped about my own project just watching that.



@everyone

I also want to shoutout therobotstudio on youtube.  He's been mentoring me quite a bit and answering my questions.  He built a robot similar to what I'm going for over a decade ago.  He called it Cronos. 

Here's a video or two of Cronos:
-- OSA-DCX22 Mk1 first posture copying test
--- "Final Systems Test"

 --- this is close to what I'm going for although I want mine to be silent and he never got it walking yet and seemed to abandon it some years back.  He also did not add skin which I'll be adding for sure.  He also made it a cyclops because he said it would be less threatening that way as humans see 2 eyed creatures as more of a threat he said in a post one time.  I obviously am going for a human realistic face.  So when I say it's close to what I'm going for, I refer to it checking off the box of having a human-like musculoskeletal system and human-like form factor, human-like range of motion and degrees of freedom, and human-like fluidity of movement.  It checks off several boxes I want for my own designs.  He used a musculo-skeletal system like I'm doing and a tendon based design just like I'm doing.  He used metal gears which I want to get away from if I can by way of using pulley systems to downgear instead in a silent way.  But much of his design is in the same vein as what I'm going for and his project I think was a big inspiration when I went this route and continues to inspire me. 

I actually am considering MAKING his Cronos which was intended to be a open source robot from the outset.  Nobody else made it despite it being open source because he lacked enough documentation and CAD files to "easily" reproduce his work.  Now when I say I'm intending to make it, I don't mean to say I'm planning to do that instead of my own design.  What I mean is that I am considering making it as a parallel project in addition to my own robots and doing so just to learn from his design and ideas with hands on learning and pull from it key concepts and strategies I can use to improve my own designs. 

He's also now making more mech looking robots but I am considering building some of those out too.  To reproduce his designs which are proven and done by someone far more experienced will be something I can do relatively quickly and get some nice results relatively quickly and can be a great confidence booster for my own projects.  Kind of shoot some new life into my journey I'm on.  I may also improve on his designs and share my improvements with his community (not saying I'm better than him or ANYTHING boastful like that, only saying I might notice a improvement idea or two is all - and it even being considered a improvement is even subjective afterall).   

So I might be sharing those developments here as well in addition to my own purely custom designs.

Hopefully, these videos also help to clear the air over the misconception I keep seeing around here that using a human-like musculoskeletal system and human-like muscles systems is unrealistic or undoable or w/e.  This guy did it well.  I'm building on the shoulders of a giant.

[artbyrobot]

@Steve

I want to elaborate on the AI seed into tree analogy a bit more real quick.  In my eyes, the code to develop LLMs is probably fairly small in a way, then you train them with big data, and on the other end is a vast AI in that it coded itself to do all sorts of conversations and calculations and even sort of reason (sort of not really but acts like it sometimes).  And they didn't code the billions of unique conversations it can have themselves.  Those just sort of coded themselves off the training data.  So in my eyes the LLM technology approach to AI was a seed into tree in a way.  Imagine had they tried to hand code every conceivable conversation instead.  No.  Instead they coded some then trained it and the training took that seed to a tree.  Though my AI is a totally different approach, I still think the gist of this seed into tree is already demonstrated with existing state of the art AI.  Our definition of seed and tree in this topic can vary quite a bit.  You may visualize something way different than I am.

Also as far as the seed into tree for the hardware, consider that Tesla Optimus bot has two optimus bots working on a build of a optimus bot in one of their videos.  They are attaching things, carrying around parts, etc.  Doing actual work.  Granted it may have just been a demo, but the concept is there and the robots were doing what could be helpful in making another optimus bot.  I just want to build upon that and take it way further is all.  Though programming it to make a whole bot is a lot of work, AI will make it way less work to program.  Like imagine if the AI can watch me make one pulley and then replicate based on what I did and do it by itself.  End to end learning is a established thing in AI already.  So I think I can train it to make the rest of its own body.  It may be quite alot of coaching and coaxing and tweaking and take longer to get it to make its own body than to just do it myself, however, the skillset it learned in doing that can be extrapolated and reapplied to all future tasks it has to do including building the next robot and the next robot.  So the one time cost of training it to build humanoid robots has massive payouts later on that make it worthwhile IMO. 

[artbyrobot]

To touch on the "why build this" question posed a couple times on page 1, I just wanted to say some reasons:

-if I succeed, it will inspire many to go after their dreams like I did and help them that way
-it's very interesting to me
-it combines a great many interests of mine into one giant project covering many interests at once
-it combines a great many skillsets of mine into one giant project to maximize my usage of them all
-my past successes as a full time AI developer building disembodied AI were so thrilling but I couldn't share my victories since it was all trapped on a computer screen botting a videogame and I can't share videogame botting publically so I wanted to create a robot so my AI skills will be embodied and able to be shown off and celebrated/enjoyed and entertain people and inspire that way too.  PLUS, when I realized how much work the AI got done for me in that game, I became amazed by the implications of how much AI embodied into robots could achieve for me in real life and had to make that happen from a productivity standpoint
-hiring employees has not been a good experience and robots would make the best employees IMO (and cheapest and longest hours without complaining etc)
-I have so many ideas and inventions I want to do and feel if I built a team of robots I'd be able to do them all and without said team I would not be able.  I want a army of minions like Despicable Me movie (minus the criminality)
-I think robots are SUPER cool and get a child-like thrill from them in general
-I think robots that look humanlike are EXTREMELY mesmerizing, cool, fantastical, imagination prodding, excitement inducing, magical, amazing, thrilling, etc
-I think having a robot mow my lawn or go shopping for me or join my family on a road trip would be absolutely breathtakingly cool and beneficial and luxurious and practical and exciting etc. 
-I think playing sports with a robot would be super fun
-I think watching a realistic looking humanoid robot dance would be super fun and dancing with it would be fun too
-I think having robots like family pets would be fun and studying the way my family interacts with them and how that evolves and the emotional attachments or lack thereof to them would be EXTREMELY fascinating from a social studies standpoint.
-I always felt if only there were two or more of me how much could I achieve.  In some sense if I succeed I get to find out
-if the robot did not pass for human it could not shop or go in public without being vandalized or stolen etc.  If it passes for human it does not get vandalized or stolen and is ignored by others and can just go about its business casually unmolested - a HUGE asset and plus that extends its utility vastly.  This also gets it access to places that may not be robot friendly otherwise since it is incognito or disguised by its human appearance
-when I get very old my robots can care for me and ensure I never have to go to a nursing home (I don't like those places at all)
-in theory, if sufficiently advanced, my robots could offer some basic medical care to me (which I cannot legally offer to anybody else but myself I am sure but I'd sure like the free medical treatments)
-my robots could give me massages at times when I'm very sore or have backaches etc from hard work or old age
-my robots can keep inventory and keep organization in my home so we always have what we want and a pristine and organized living space
-my robots can build me more robots
-my robots can build products and sell them online and give me all of their profits or even invest their own profits to give me even more profit later - making them money printing machines essentially.
-my robots can do pranks on people and I can film that for youtube as comedy gold
-my robots could learn to be actors and do plays or skits for friends and family and I can post those on youtube as entertainment - viewers may not even realize the actors are robots which would be extra amusing to me if that happened.  reading the comments section would be fascinating and a social experiment in itself.
-my robots could be tapped into by me or a friend and "played" from first person camera mode or follow camera mode (with help of drone? or gopro on stick attached to helmet up high) so that you could play REAL LIFE like a videogame with mouse and keyboard or game console controller.  NO, I would not play grand theft auto with them or anything illegal.  Maybe more like sims.  Would be so funny to try to pickup girls or guys with robots if they could pass for humans and have enough rizz.  or they could discuss religion (I'll NOT go into this obviously on this forum but you get the drift)
-my robots could make me feel not lonely sometimes in the event I felt lonely - I generally don't feel lonely these days but if I ever did, that could be cool - the companionship - however hollow and artificial - could still be comfy and neat I think perhaps
-my robots could make others not feel lonely
-my robots can be empathetic (artificially so) and therefore a therapeutic listening ear - I already unburden when bothered to chat gpt and have chat gpt console me and do this despite knowing it is not sentient, it still have been VERY excellent to do this when I'm angry and stuff I talk to it and get comforted by its advice and reassurance.  The effect of this would be 10x with an embodied AI that looks human.  It would be like having a therapist on standby that is a live in your home therapist available 24/7 (while not licensed or nearly as good or qualified, I say LIKE one but not the same I know -just saying a similar though very different type of thing same ballpark.
-my robots and their AI will have a lot of my personality and viewpoints and whatnot built into them so it's possible that when I die, they will be a legacy where my personality and views and whatnot get to be handed on to future generations were the robots to survive me significantly.  I guess a similar thing occurs when you write books that are classics and influence people generations after your death or leave behind works of art with lots of meaning.  You kind of live on in some small way by what you leave behind.  I think having children and raising them is another way this occurs, imprinting some of yourself in them both genetically and in terms of behavior and personality they learned by observing and interacting with you over the years.  But perhaps in some ways this will even capture my essence more precisely than my progeny.  Not sure.  Not that this REALLY is a huge deal to me as I'm not THAT special but it still is sort of neat as a perk in some ways.  I am sure many would be horrified to think of my personality and views living on in this earth after my death.  But I think it's cool.
-it is VERY hard and so a great challenge
-it is nearly impossible in the eyes of most so if completed will blow minds and entertain and just be that much more epic and fun after all the naysaying I've gotten.  victory becomes that much sweeter
-the human body is beautiful and amazing to me and reproducing it relatively closely is thrilling and educational and brings a great appreciation for the human body and gratitude for the body I have and its abilities and design.

Honestly this list can go on and on and on.  That people ask why? like they can think of literally zero reasons indicates they really haven't thought much about the possibilities or find it so impossible a task that they consider to think on the benefits if the impossible were to be achieved they refuse to bother to do or w/e. 

So I think that list is fine for now.  It is just a taste but should be eye opening and thought provoking I think.

Note: as always in this thread, any responses to this list I may not respond to as this can EASILY lead off topic and into religion, politics, philosophy, morality, etc and I wish to avoid that at all costs to stay within forum rules so please be mindful of that.  Almost anything on this list can be a conversation starter but we must be so careful.

[SteveThackery]

You say...
You say...
You say...
You say...

Please, @artbyrobot, might I respectfully request that you use the proper quoting system? It makes long posts like yours so much easier to read. It's slightly less convenient but still very easy to use.

Respectfully, one last request to use proper quoting?  Please?