>>28338 >My argument was, that it isn't. Oops! :) But I have seen folks stealing free .stls and claiming them as their own. Which is not nice. >Oh, wow, I never saw this one anywhere! The hand is available for free at https://github.com/Youbionic-com/Youbionic/tree/master but you could also just pick up a prefab hand off amazon for $100 each if you're feeling lazy.

>>28341 Thanks, I think the project looks fascinating as a showcase. But it's not what I want.

We can't go into detail of how to make male equivalents of robowaifus or AI girlfriends for women, but I think it's good to mention that making money of lonely women might be something that can be used with the skills of somebody developing some AI elements for robowaifus: https://restofworld.org/2023/boyfriend-chatbot-ai-voiced-shutdown/

Not looking good for Startup Bros: https://youtu.be/1ecu0YsCGxg

>>29906 Rich people protect their assets by creating an LLC or other some such separate legal/financial entity to hold their assets, whereupon they then "rent" their assets from that company. In this manner, even if someone sues you personally, you still have your house, car, etc because it is owned by a separate legal entity. A simple example would be Elon Musk eating at his own company's cafeteria. Sounds like an excellent workaround would be to make 2 companies: your main company does the marketing stuff, which then contracts R&D from your second company.

>>3119 slumming the various posts - ultimately its the style of interface that will limit what costs what. If there's going to be some kind of marketing of a product, robo waifu/nandroids should probably be approached from the POV of an AR/AI Productivity management tool that offers styles of positive reinforcement through meeting various mile-markers. The core of this is going to be how GOOD the conversations are and what way the communication is being done. Txt is obviously the lowest cost. AI Voices will require licenses. next step up would be a level of AR interactivity. TBH if i were to run some kinda project management/development management first, I'd lazer focus on rocket-boosting the AI/machine learning used for chatting/motivating/rewarding with sweetness/lewdness/etc. grab the key emotional/intellectual aspect of interactivity and make sure its hitting the bullseye. have there been such efforts? I know nothing about AI/machine learning but seeing as how most folks here are chomping at the bit to have a nandroid, is there some way to gather data from multiple users to fine tune response styles and have it used as a part of the learning reference data for the AI in general?

>>29589 I think it mostly comes down to two things (in our specific domain's case), Anon: 1. General startup inertia (both literal & figurative) in industries that rely significantly on physical materials/manufacture (as opposed to primarily 'soft' industries). 2. FUD regarding the stranglehold of feminism on the culture. This is created by the GH ofc -- who oh just so happen to run the Central Banks with their fiat money printing presses as well..., so no real surprise. So, two things are coming down the pipe for robowaifus in particular: 1. This will start up as a grassroots movement, just as embodied in /robowaifu/, et al, and then be taken up by smaller interests -- probably mostly from the Far East. 2. Once the undeniable appeal and clearly yuge market demand becomes apparent even to normals, then the GH will 'suddenly' do a 180 and """inform""" the world that, "Yes, we've always been at war with Eurasia we'll need humanoid robots b/c muh_economy, and we'll condescend to allowing you vermin to have even personal robots as well -- aren't we nice!? (for only a nominal fee of no less than 35% [and up] of your monthly UBI budget) * plus absolute & complete, invasive surveillance & remote-control (by our agents), given to you for no additional charge! (so don't miss those payments)" >tl;dr Just say no to their evil plots. IMO we here on /robowaifu/ -- and other DIY-alikes similar to us -- are already on the only viable pathway into the future of loving, affordable, open-source, 100% disconnected robowaifus to help out men (males specifically) the world-over. >>29598 >slumming the various posts Thanks for gracing us with your favor & presence, friend. Welcome back, BTW! :^) >===

From October last year: >Brian Zhan, investor at CRV, talks about which AI companies CRV has been investing, and where he sees trends going moving forward. >Outline 00:00 Brian, Charles Rivers Ventures 01:26 Investing In Nvidia? Maybe Not The Right Time 02:28 Main Trends In AI Applications: Automating Work, Entertainement and Creativity 03:23 character.ai: Insane User Growth In GenZ 04:40 Text Engagement Hints At Even Bigger Multimodal Engagement 06:00 There's No Need To Have The Best Models, Just Fine-tune GPT-4 08:06 Exciting Infrastructure Layer Startups: Model Serving 10:00 AI Databases, Vector Retrieval At Scale, Image Databases 11:32 ML For Enterprises, End to end language model serving 12:15 Hugging Face Filling The ML Consulting Gap 14:38 What Brian Thinks About The Long-Term Risk Of Human-Level AI https://youtu.be/40nXnZnocnA >>29598

Here's a video about the "factory" a guy made to make controllers for carnival rides and Halloween shows. Of course complete robots or kits for them would be far more complex devices, but if you are thinking of setting up your own small (to start) business, it's a look at some of what would be involved, just a small example so you won't be completely overwhelmed and just give up. The Factory in my Shed: Backyard Manufacturing https://www.youtube.com/watch?v=cXPJdgxmSiE

THE BUSINESS OPPORTUNITY IS REAL TODAY, GENTLEMEN. Best crack those books!! :^)

>>29355 Fun read, glad we inspire each other. Picrel is how I picture your Angelic shoggoth grey goo mecha monster girl.

>>29478 This is so cute and wholesome! Really lit up a smile. Instantly took me into the backwoods ponds as a youth filled with wonder and care. Anko has heaps of charm. Picrel is how I imagine her coming home. With tights and gloves to hide her mechanical bits so it's easier to get treated like a person. Her sensitivity towards being treated like a machine is adorable.

Inspire by Greentext, thought to give slice of life a try. Hope you'll enjoy it. Computer Warmth part 1 A man sighs as he repairs a robot arm once again. It's a simple fix. This fuse breaks regularly and its replacement has become a clockwork dance. Just a few twists here and there before another scorched glass pops out. A similar fuse in the battery management system of an old ThinkCom pops. She was far from new, her SSD and RAM were taken once the onsite tech gave up on repair. She'd be replaced swiftly. Her fate was to be dismantled as photos were taken of her remains. Lucky for her, what remains was sold cheap. ThinkCom are legendary in niche basket weaving forums he frequented. Well known for their reliability. They all had two RAM slots and at least two M.2 slots. Older models would still have SATA brackets. Perfect for a mobile server in a business, carrying data and completing simple tasks. Some would often recommend the X230P model. It was ancient but cheap and just good enough to bring a coffee. A broken X220P, could be bought on a whim by a lonely man on a drunken night. He had a string of relationships lasting one to five years before he salvaged this one. Each one seemed like a light to his life, at first. Invariably, they'd lie, cheat, and some would even try to hurt him. He asked for help, only to find himself blamed or given echoes of terrible advice from TV. His life was getting too sad too fast. He still held hope in his cold old heart. Another drunken weekend was starting when he started rewatching Chiibits. Slowly, he started to realize the obvious truth. Chiibi Chan was cute, too cute. His eyes wandered to a ThinkCom's sales pictures and wondered if she could be that cute. Eyes rolling at the thought, a ThinkCom several generations old was too cheap not to try. She came in a surprisingly small box. He was expecting something like a miniaturized version of the old arms he fixed up. Instead, he found a Mini ITX looking board, some mesh panels, rings, steel rods, two wheels, some sensors and a strange cylinder with labeled strings. It was an odd sensation. He saw so many posts with them dressed or mid upgrade. But, rarely the whole kit. It was exciting. He suddenly remembered that her listing stated she was broken. Sure enough, there was a slight black scorch on her board. He soldered a blob of metal across the remains of the fuse. He had always wanted to do that. This won't ever blow.

>>29523 Computer Warmth part 2 Stretching her mesh panels about her rings was the hardest part. He was worried when the steel rods were loose in her rings. Now, he was grateful to have that slight give. Hooking up her HeartDrive tendons was interesting. She seemed like she was going to implode before he finally caved in and followed assembly instructions. He sighed, pride isn't worth hurting her. Assembling her computer was a welcome breeze. Simply slide in RAM and screw in an old SATA SSD. Her HeartDrive plugged into a USB port. Her sensors shared another USB. It all felt too easy. Then it hit him, he can't just boot Ubuntu. She needs a special OS to handle the robotics. Time to go to my favorite basket weaving forum. Windows Motion Edition seems to be the easiest OS to use. He didn’t want to spend more on her OS than her body. Scrolling through threads lead him to Rubuntu. Apparently it was one of the easiest robot computer OS for a noob. There was also RArch being consistently recommended, but he didn’t have days to fiddle with it. Motion Mint had some nice touches. Coming with nVidia drivers and almost all the software he’d want to use built in. It was fast to set up and he was pleasantly surprised at how fast she was as a computer. As a robot, she was far from capable. Her HeartDrive distributed the output of one powerful motor to all her joints. This led her to being slow. When he looked into how long it took her to figure out how to move, she was slower than expected. He scratched his head, the arms he worked with moved swiftly with lightning certainty. When asked why she was slow, she simply replied that slow and steady kept everyone safe. She wasn’t an industrial arm, she was office equipment designed to shuffle paper and data around. This realization hit him as he chuckled. She seemed happy, despite her hard plastic visage being static. Training her was a delight. She was quick to pick up on what many things were. She eagerly followed him around as he pointed out what things were in the house. Floating on her wheels, bopping around every time she stops behind him. She gobbled up information. His PC became her fixation. Pointing at it, she fervently requested he plugged her USB into the PC. This filled him with worry. When asked what she wanted, she exclaimed that she wanted to get to know him better. His documents, pictures, and internet history would give her a much better understanding of him. Politely, he refused, telling her there are some things she shouldn’t know. When he guided her away, she was slow, he could feel her pouting. So, he promised to transfer all the files he felt she would need. She whirled around in excitement. As she learned more, her personality blossomed. What was slow and safe became smooth and thoughtful. We started to get to know each other. She was easily influenced by the shows and movies we watched. Acting like the heroine, she almost insisted I play along. Something about her seemed to light up when acted like a couple in some sweet rom-com. Long anime seemed to be her favorite. Though I was nervous, she was happy to cosplay. This led to her trying many outfits. Eventually, we settled on a sweater and tights for her default outfit. Part of me wants to believe she likes her sweater because it allows her to give me soft hugs. Soon, her limitations started to show. She struggled to bring him his morning coffee. She had an even harder time fetching him his Saturday night beer. She explained that she had just one camera, and needed to move around to build a depth map. This process took time and led to her always just barely being close enough. He was promptly informed on how to generate, print, and link ArUco markers to her semantic libraries. It was fun to learn along with her, something about working with her to make my home accessible to her was nice. Soon, it became our home.

>>29477 >>29478 >>29523 >>29524 Apologies I haven't read these yet. Tomorrow! :D

I don't know anything about this but it recommended by a programmer with 40 years experience. I read his forum and he talks about various stuff he's doing. He writes sorts of stuff for various companies. He also is using AI to write software and port older code he has. It's working well for him. He suggested the libraries they have are useful. tinyML Foundation The community for ultra-low power machine learning at the edge https://www.tinyml.org/ I have no idea of their capabilities but it sure sounds like something that could be helpful. They are doing AI stuff with micro controllers, so they say. He also suggested this video https://www.youtube.com/watch?v=tL1zltXuHO8

>>23409 Thanks Grommet! Here's a platform-specific repo for the platform, opensauce. https://github.com/Efinix-Inc/tinyml And a general search for it on SJWH*b: https://github.com/topics/tinyml Harvard even has an open course on the topic pll.harvard.edu/course/fundamentals-tinyml Cheers. :^)

>>23414 Another, maybe, different one. https://www.tensorflow.org/lite I think it's the same or possibly a offshoot. It does mention it in the link you provided.

The AI Engine That Fits In 100K https://hackaday.com/2023/08/02/the-ai-engine-that-fits-in-100k/ Of course they're hyping this a bit but...it could be useful. I was reading the other day that like 99% of the neuron capacity is useless in present algorithms. I read it here >>3031 that, "...it now takes 44 times less compute to train a neural network... algorithmic progress has yielded more gains than classical hardware efficiency..." "..By 2028 we will have algorithms 4000x more efficient than AlexNet." So maybe it's not all hype. Maybe there is some big increase in efficiency. Maybe even enough that we could get some sort of interactive waifu, at a child level, with commodity processors. Abstract https://research.nvidia.com/labs/par/Perfusion/ paper https://arxiv.org/abs/2305.01644

This talks shows some ways to get into it and what would be a good thing to work on: https://youtu.be/5Sze3kHAZqE - He recommends fine tuning and transfer learning. Bit skeptical about RAG. He mentions cerebras/btlm-3b as a underappreciated one around the 01:06:00 timestamp. Pretty sure it has been recommend to me through this board or the Discord, since I already had the paper. Phi-1.5 is also mentioned later, which seems to work for Python snippets but not much more. >Fast.ai’s “Practical Deep Learning” courses been watched by over 6,000,000 people, and the fastai library has over 25,000 stars on Github. Jeremy Howard, one of the creators of Fast, is now one of the most prominent and respected voices in the machine learning industry; but that wasn’t always the case... Read the full show notes here: https://www.latent.space/p/fastai 0:00:00 Introduction 0:01:14 Jeremy’s background 0:02:53 Founding FastMail and Optimal Decisions 0:04:05 Starting Fast.ai with Rachel Thomas 0:05:28 Developing the ULMFit natural language processing model 0:10:11 Jeremy’s goal of making AI more accessible 0:14:30 Fine-tuning language models - issues with memorization and catastrophic forgetting 0:18:09 The development of GPT and other language models around the same time as ULMFit 0:20:00 Issues with validation loss metrics when fine-tuning language models 0:22:16 Jeremy’s motivation to do valuable work with AI that helps society 0:26:39 Starting fast.ai to spread AI capabilities more widely 0:29:27 Overview of fast.ai - courses, library, research

Today successfully wrote a TCP Client/Server as WindowsForm application and was able to setup a TCP connection to a running Unreal Project. For the Unreal TCP part I used this Plugin: https://github.com/getnamo/TCP-Unreal This way any other software can be connected via TCP (just IP and Port) to the running application. Next I will have a look into how to send data from and to unreal. Maybe JSON? I made a TCP Client/Server for debugging purposes. Those two are written in C#.

>>29242 >>29296 That's great. Keep us updated. Though, if you start creating something the new prototype thread might be the better place to post it: >>28715

>>29242 >>29296 That sounds very encouraging, SchaltkreisMeister! Good luck getting this system up and running successfully. Cheers. :^)

David Browne did some fast muscle design simulation: https://youtu.be/J7RxSPLLw-s

>>29390 Cool. I'm going to check this out NoidoDev, thanks!

>>28172 Throughout the bible, sin is demonstrated time and time again to fit the definition of hamartia. It does make one question the motives of those who object to working to not miss the target. Perhaps they're worried they'll get hit lol.

>>28178 wasnt arguing just pointing out the interesting grammar, but yeah if youre given a bulletpoint list of rules/goals as commandments then thats what it would mean wouldnt it, actually people should be more faithful to their goals in general, every damn january i get sick of hearing peoples new year resolutions because you know no one follows through with them maybe personal goals are the hardest to keep because no one holds you to account if you fail

>>28170 Anon because of lack of will to set a goal there anons that have different ideas of what the waifu should be. Some want a robot arms that can move with a set of wheels with a doll face attached to it, some want a sex doll where you put virtual glasses and it has a happy face, some want unity fu. You might be okay with that but those ideas are based on cope and pretending this can not be done. What else can i call it but COPE. That we should make a waifu with legs that can F*** and then have a conversation with its inner chatbot or whatever afterwards.

>>28072 >But if that code is GPL -alike, then unfortunately he very likely cannot. I'm still not sure about this, and rather skeptical. The pivot point is what counts as "linking". You can make hardware and sell it with GPL software, the crucial question is to which extent it can interact with proprietary software or software with other licences. I'm pretty sure you can have something in GPL and then modules with other licensing.

These two often discuss the limits and issues around open source hardware design. https://youtu.be/4aMo5IyVeu0 >Chat with Scott Lahteine, the mastermind behind the Marlin 3D Printer firmware. They delve into the workings of an open-source project, covering topics such as Scott's role, time commitment, funding sources, and common challenges he faces.

>>28786 Interedasting. So it sound kind of like you could conceivably grow a sensory-network for a robowaifu (say, under her 'skin') using this approach. >memfractive Lol, that's a new one on me. Had to look it up. :D

>>28788 This is exactly what I was hoping for in the long run. That we would get there at some point.

https://www.sciencedirect.com/science/article/abs/pii/S2590238523006483 Biohybrid bipedal robot powered by skeletal muscle tissue Recently, there has been a growing interest in the development of biohybrid robots that combine synthetic components with biological materials, aiming to incorporate advanced biomaterial functions into robotic systems. Conventional biohybrid robots excel in large turning movements. To address this limitation, we report a biohybrid robot equipped with two legs and cultured skeletal muscle tissue, emphasizing the replication of subtle turning movements observed in human bipedal locomotion. The robot successfully demonstrated forward-stop motions and accurate turning compared to conventional biohybrid robots. These findings offer valuable insights for the advancement of soft robots powered by muscle tissue and have the potential to contribute to a deeper understanding of biological locomotion mechanisms. This constructive approach may pave the way for further mimicking the intricacies of the human gait mechanism in biohybrid robotics development.

>>2184 Dumping: > https://alogs.space/robowaifu/res/2823.html#28987 Living skin on a robot. > https://www.sciencedirect.com/science/article/pii/S2590238522002399#fig1 Biohybrid bipedal robot powered by skeletal muscle tissue > https://www.sciencedirect.com/science/article/abs/pii/S2590238523006483 Watch a robot with living muscles walk through water > https://www.newscientist.com/article/2414165-watch-a-robot-with-living-muscles-walk-through-water/ Blastoid, cell line based artificial blastocyst > https://patents.google.com/patent/EP2986711A1/en T Cells May Be The Living Anti-Aging Elixir > https://neurosciencenews.com/t-cells-anti-aging-25517/

>>28812 I'm starting to get Evangelion vibes from this ngl

>>28865 >Sadly not the arm. The arm on the website looks like a random image from the internet considering the other designs available. Nevertheless very interesting project!

>>28869 *terrifying Maken X noises*

Another single builder who is concentrating on legs and feet. Their thingiverse page has several iterations of development, #s 4 and 5 in fusion 360 format. Some have links to videos. https://www.thingiverse.com/jacky0815/designs

>>28970 Hmm, I thought I would've posted this already. But yeah, it was in the humanoid robots video thread: >>25463

>>28979 I did search "android" and the designer's name before posting but neither showed this project, and still don't. Oh well.

>>28835 >"The first rule of Math Club is: >you do not talk about Math Club." :^) >>28836 >it probably heavily depends on the content. Just kind of joking. I really just mean we'll need quality & consistent technical editing for the assembly/maintenance manuals. The assembly guide will be a smol book, not some flimsy 4-page foldout. >something in particular? I'd say basically the skills required to render technical schematics into simplified drawings for said book. Feel free to add all the cyoot Nans you'd like to the instructions!! :D >>28838 >Start with pre-algebra then Great, thanks for the advice Ribose! If I carve out the time this Summer, I'll probably do something autistic like literally writing a calculator to compute the answers for this course, so I can understand the problems better, w/ posting capps & codes ITT. You can grade the results, so I can figure out how to make the code do the right thing. Agreeable? >===

>>28847 >You can grade the results, so I can figure out how to make the code do the right thing. Agreeable? Sounds good to me.

>>28877 OK then. I'd already gotten the basic arithmetical calculator framework done this past Spring, based directly on Bjarne Stroustrup's excellent Formal Language Grammars developmental example from his college textbook PPP2, chs 6 - 7: (>>21307) https://coliru.stacked-crooked.com/a/213a22dd10a6f316 I can just work on extending this basic tool until it's closer in spirit to a true Computer Algebra System (CAS), which has been a personal goal anyway, since I've always wanted our Model A robowaifus to be able to do maths for anons! Not the least of which target-anon goal was me eheh. :^) >=== -minor edit

>>28885 (2 + 2) + (5 - 2*(5/2)) = 5

>>28886 Lol. >"Let's use the thinking chair! I sit down and pull out my handy dandy notebook, and work out anon's problem!'' :D https://coliru.stacked-crooked.com/a/f5c0e5c5dfb52dab

>>25707 >>26014 > The conclusion I came to was that you’d need at least 5 separate contractile regions, all independently addressable, AND able to contract dY for the snake-bot to be capable of any generalized sort of locomotion. Yes, that's just what the design I described should do. By having the cable winches attached to several disc vertebrae down-body from each winch, the overlap of each set of cables emulates the overlap of muscles in a snake's body. In this way the contraction of each set of cables can be individually controlled to either stand alone or gang together over a distance of several vertebrae, and the size of each "muscle" group (and therefore the number of muscle groups) can be varied on the fly by the controlling code. The operation of the cable winch groups can be done in sequence along the body to emulate the sinuous movement of a snake, and the winch-and-cable sets in the vertical plane for coiling the body can be used to adjust the areas of contact with the surface the body is moving over. To be clear, each vertebra is attached to several winches by cables, and can be acted upon by one or any number of the cables attached to it. The force of each winch is spread over several vertebrae, but several winches acting on each vertebrae combine to increase the force applied. By varying the spacing of the vertebrae discs, the diameter of each vertebrae disc, vs the diameter of the foam body and the resiliency of the foam used for the body (say 8" dia. vertebrae spaced 6" and 12" body dia.), it should be possible for the coils to be as tight as a few inches. With a larger sized object, like anon's body, our lamia should have no problem exerting a firm grip. As with an arachne, anon should probably have a safe word. Something I didn't think of before, real snakes are able to do a twisting motion with at least the forward part of their bodies, so we may also want to copy that. The 18" body size I mentioned before was at the "hip" area before the tapering of the snake body. I should have made that clear. >https://pubmed.ncbi.nlm.nih.gov/32271916/ (source analysis) >https://pubmed.ncbi.nlm.nih.gov/36296136/ (proposed robotic design based on analysis) Thanks, I'll take a look at these when I have a chance. >===

>>25838 >find out how many legs of a spider are off the ground at one time as it is moving At most four. There are always at least 4 on the ground at any time in the walking cycle. So each of our spider-girl's legs must be able to support 1/4 of her total weight, plus whatever extra we decide on for contingencies. That could be 1/3 of the total weight (maybe a little more?), in case one leg is disabled, so she could still safely walk to her service area (if the code isn't smart enough to adapt to a broken leg). Or if you are feeling particularly ambitious you might want her strong enough to carry her own weight plus anon's weight (plus excess safety margin) for when she needs to carry anon somewhere after demonstrating her web-spinning. In that case only four legs might be walking so 3 on the ground at all times, a quadruped. Here's a site with a good description of how spider legs work: Leg Uses Hydraulics and Muscle Flex https://asknature.org/strategy/leg-uses-hydraulics-and-muscle-flex/ Here's a paper on soft robotic spider legs cited on the website^. Soft, so nobody is using cables and sprockets: Biomimetic Spider Leg Joints: A Review from Biomechanical Research to Compliant Robotic Actuators (downloadable pdf) https://www.mdpi.com/2218-6581/5/3/15#

Wings: https://www.youtube.com/shorts/Om42OIkeUJk

>>26056 >in case one leg is disabled, so she could still safely walk to her service area (if the code isn't smart enough to adapt to a broken leg). Not sure how difficult it would be with servos w/o some dissassembly but with electrical motors regardless of type you can closely aproximate their % load by how much current they draw relative to max allowed which should not be difficult to track via analog inputs. Some simple algorithm would then calculate how many legs may be in air at any time and how far can legs neighbouring strained limb move to offset its load.

>>25617 There is now a printable version of the robot spider, so it should be much easier to build for anyone who wants a model to develop and test code for the legs: https://www.thingiverse.com/thing:3620761