Real Fembots - What do we want from them?

[SyntheticBuilder]

I’d like to discuss real fembots. Not fictional images or videos, but actual physical machines And not from a hypothetical point of view, but from a very much realistic one. We all know with the current state of AI, having a somewhat functional fembot is not a fiction anymore. In fact, there are many examples of very real, existing machines. I’m sure you know them

So, to start the discussion, here’s my own very basic set of requirements for a fembot:
1. It should be attainable. Not $175,000 like what Realbotix is charging, but something any average person can afford.
2. It should be available. Basically, go to your nearby Best Buy and grab one Or at least order it on the jungle website with next day delivery.
3. It should be functional or at least entertaining. I’m not asking for an all-purpose maid, but even a simple robotic friend is good enough for starters.
4. It should be good looking. This is obvious However, different people might interpret “good looking” in different ways, so…
5. It should be smart enough to have engaging conversations, play games, and of course to have some romantic options.
6. It should have a functional body. Not necessarily walking or running, but all arms and legs should be moving, so she can change poses, use gestures, interact with objects, etc.
7. I’d really like to be able to mod it – make new parts, change existing ones, reprogram her, etc. So it means open architecture and preferably fully FOSS.

Now what do you think? I’d be really interested in discussing a perfectly-balanced (functionality vs price) real fembot.

[Kube²]

"It should be attainable. Not $175,000" then describes the features of a machine 10 times more expensive...

The problem is that what you describe is impossible to achieve at the moment, and even less so for the price you are aiming for. It's no use saying people should do this, they can't, no one can.
But it's an interesting topic of discussion. We'd already have to live in a world where autonomous domestic androids are mass-produced in order to see how low prices can go. We're not necessarily very far from that world; maybe in 5-10 years we'll have a good idea of ​​what's possible or not...

[jolshefsky]

Yes... ish.

My expectations of reality are already quite low (and by current examples existing today at any price, even that's too high.) I've mentioned several times in the past how I recognize my fantasy as fantasy: there cannot be a lifelike charming young woman that is actually a machine. My own preference is for a perfectly human body that is paradoxically idealized and plastic; a completely independent person that is paradoxically programmable and obedient—these paradoxes in my fantasy make the reality simply impossible.

But should such a robot be available someday, she most likely will be exciting for the first 5 minutes (okay, 5 days to be honest) but after that she would be too not-actually-intelligent and seem like a boring toy. And given that the manufacturer would absolutely have cloud-based surveillance, any private interactions would not be possible.

Also, her skin won't stay ... appealing ... for long: dirt and, um, organic matter would quickly make her body pretty gross, and only the most careful cleaning and lack of touch would make her stay attractive. The way living things slough off dead and damaged skin cells revealing fresh ones, ironically enough, is what makes us feel okay to touch.

[SyntheticBuilder]

The problem is that what you describe is impossible to achieve at the moment, and even less so for the price you are aiming for. It's no use saying people should do this, they can't, no one can.

First of all, let me tell you - I liked your "Fembot Instruction Manual" series!! A LOT!

I might slightly disagree with the statement quoted above, and I might have an actual proof of my point. You see, it all depends on your expectations. If you want a realistic-looking, fully skinned, indistiguishable from a real woman fembot, then yes - no one can produce such a machine for a low price. But! If we adjust our expectations a bit, and say that having a spandex-based skin cover instead of silicone is okay, that having a bit rough movements is also okay (some might even find it better ), and the battery life is not really a concern... Then what we end up with is a female android with arms and legs, capable of moving, wearing high heels and normal clothes, who has an obvious robotic look, but otherwise serves the purpose. Just like this one - https://youtu.be/xKOH8TngKsU
Would you like to have such a machine now, in year 2025, or would you instead just sit and hope that a realistic cheap fembot would arrive to the store nearby some decades in the future?

Same goes to functionality - it's all about expectations. We all understand that a realictically behaving fembot would probably require an AGI level of AI. Which is... well, a bit far away But if we accept the quirks of the current LLMs and VLMs, and imagine a fembot based on those AI models - we can have a more than real possibility to have it right now! Not in 20 years.

For me, it's all about having something "NOW" and not tomorrow - because after all I might not live until tomorrow. Simple as that But we all different, and I respect other opinions on the matter, of course. So I'd like to have a fembot which can talk and play with me, look good, move the way I want, and be cheap. But more importantly - I can buy that fembot today, and not in gajillion years Think about it as of times when you had your Commodore 64 or ZX Spectrum. You might have wanted a powerful IBM mainframe, but you had only a cheap home mini-computer. Time moved on, and now you have a PC (or whatever) on your desk that is millions times better than that IBM mainframe back in the day. So in my opinion, it's better to have a simple fembot now, and then I'll be more and better prepared for having a better fembot tomorrow

If you think a full-size robotic woman with current-gen AI can not be cheap - I have a proof of the opposite. You can take a look at https://syntheva.no - a Norwegian company making fembots today for just $2000
Full disclosure - it's my company, and yes, anyone can order The price is adjustable depending on what you want (which configuration). I won't say our fembots are perfect, but they are very good for the price! And since the whole architecture is mostly software-defined, they can be upgraded quite easily.

[SyntheticBuilder]

My expectations of reality are already quite low (and by current examples existing today at any price, even that's too high.) I've mentioned several times in the past how I recognize my fantasy as fantasy: there cannot be a lifelike charming young woman that is actually a machine. My own preference is for a perfectly human body that is paradoxically idealized and plastic; a completely independent person that is paradoxically programmable and obedient—these paradoxes in my fantasy make the reality simply impossible.

Believe or not - I totally understand what you mean Yes, it's paradoxical, but I think not as impossible as it seems. Being an AI researcher myself, I might tell you - "programmed independence" is quite possible, and in fact, it is one of my primary research topics in application to multi-agent AI structures. We can talk about this to a much greater length if you want - just let me know

But should such a robot be available someday, she most likely will be exciting for the first 5 minutes (okay, 5 days to be honest) but after that she would be too not-actually-intelligent and seem like a boring toy. And given that the manufacturer would absolutely have cloud-based surveillance, any private interactions would not be possible.

Well, my hopes here are that the robot would not become "boring" in 5 days. It no doubt will loose some novelty value. But I would always like to continue communicating and playing with it, making new stuff as we go.

Now, my favorite topic - THE CLOUD. In short - I hate it! All my AI systems are based on local inference (training is slightly different matter, of course). For example, my fembot "Synthia" is running 100% locally on an SBC inside her head, she doesn't even has an ability to connect to Internet. So all your private stuff will be left private forever
"But what about updates then?!" I hear you asking Well, it's as simple as plugging a USB pen drive, and letting Synthia update herself

What me and my company is doing now is exactly an attempt to show and prove to everyone - "Look! It's possible!" Open-source software? Check! Local inference? Check! Total security by removing all connectivity? Check! Affordable design ready for mass-production?? Check!

Also, her skin won't stay ... appealing ... for long: dirt and, um, organic matter would quickly make her body pretty gross, and only the most careful cleaning and lack of touch would make her stay attractive. The way living things slough off dead and damaged skin cells revealing fresh ones, ironically enough, is what makes us feel okay to touch.

True! Generally, the robotic skin is a hardcore matter. Extremely expensive to produce and extremely hard to maintain. As I said in the post above, if you don't mind having a non-silicone skin on your fembot for now, then it's not really a problem
For example, in case of Synthia, users might choose spandex-based skin covering (the only skin covering available atm, though) and then simply wash it in the washing machine every week or so

[ProchazkaJBG]

Sorry but to me this all reads like a big ad and sales pitch for your company


IDK if those are your intentions, I'm just tossing that out there

[SyntheticBuilder]

I see your point. And in a way - it is a bit of a pitch. But... Only partially. And for a good reason. Or so I thought.

Put yourself in my shoes for a second. Let's say you made a thing which works as you want and has a set of definitive features. Doesn't matter what it is. It could be a lemon peeler for all intents and purposes. Now if someone would tell you "you can't peel lemons for less than $50/lemon", wouldn't you defend your argument by saying "hold on, please, but I have a peeler right here which costs $10, and I can peel 9000 lemons with it, so your argument is not exactly 100% correct"? I think anyone who has a point of proof would use said proof to strengthen their argument.

You see, my intention was (and still kinda is) to see what people who're (allegedly) into fembot scene would like to get from a real-world machine. Not from an image, or a short video. But from a real-world thing.
And I myself was (as I stated before) a big fan of FembotWiki, and it gave me a lot of inspiration for what I'm doing now. So my reasoning was simple - if those guys gave me these ideas, they'd probably be glad to see a beginning of a transformation of their ideas into reality. So I thought...

But now, I see there's seemingly no interest. Looks like some people (for clarity - I'm NOT addressing participants of this thread) prefer to live in fantasy world, potentially with ideals so high (thanks to years of evolving the fantasy) that NOTHING in then next 70 years would satisfy them. And by that time... I'm sorry to bring bad news, but not many of them would still be kicking by that point.
Funny thing, right? To understand that ordinary people are more interested in humanoids for home use than people who're supposedly should be defining how the whole entertainment robotics industry should be shaped.

Well, this is the truth. My intent was to make something very special together with the community which I thought would be welcoming the idea (not a specific product) of such development given the fact that the OP can demonstrate a proof of their abilities to develop the idea into reality.

[ProchazkaJBG]

I see your point. And in a way - it is a bit of a pitch. But... Only partially. And for a good reason. Or so I thought.

Put yourself in my shoes for a second. Let's say you made a thing which works as you want and has a set of definitive features. Doesn't matter what it is. It could be a lemon peeler for all intents and purposes. Now if someone would tell you "you can't peel lemons for less than $50/lemon", wouldn't you defend your argument by saying "hold on, please, but I have a peeler right here which costs $10, and I can peel 9000 lemons with it, so your argument is not exactly 100% correct"? I think anyone who has a point of proof would use said proof to strengthen their argument.

You see, my intention was (and still kinda is) to see what people who're (allegedly) into fembot scene would like to get from a real-world machine. Not from an image, or a short video. But from a real-world thing.
And I myself was (as I stated before) a big fan of FembotWiki, and it gave me a lot of inspiration for what I'm doing now. So my reasoning was simple - if those guys gave me these ideas, they'd probably be glad to see a beginning of a transformation of their ideas into reality. So I thought...

But now, I see there's seemingly no interest. Looks like some people (for clarity - I'm NOT addressing participants of this thread) prefer to live in fantasy world, potentially with ideals so high (thanks to years of evolving the fantasy) that NOTHING in then next 70 years would satisfy them. And by that time... I'm sorry to bring bad news, but not many of them would still be kicking by that point.
Funny thing, right? To understand that ordinary people are more interested in humanoids for home use than people who're supposedly should be defining how the whole entertainment robotics industry should be shaped.

Well, this is the truth. My intent was to make something very special together with the community which I thought would be welcoming the idea (not a specific product) of such development given the fact that the OP can demonstrate a proof of their abilities to develop the idea into reality.

Again, to me it sounds like you're trying to convince people to lower their standards just for the sake of owning a 'fembot'. I want one as much as the next person here... but I'm not going to just purchase ANY fembot... it has to be one I actually want. Unfortunately that's not going to be possible for a while... probably.

Maybe I am being too harsh or misreading things. If that is the case just ignore me lol.

I do hope you have success with your project, and find people that share your enthusiasm for it.