>>3005837
Show prototype

>>3005837
i often wait till i have something show to post but how you just make some place holder thread is gross. it tells me what kind of person you are

>>3005529
>He's a bit meh at it sure, but I wouldn't say bad. He has helped me immensely in this area, bringing awareness to me of things I'd have never known even existed. This is invaluable. Game changing.
lmao actual living people gave you tons of advice and valuable critique months ahead of your 'invaluable discoveries' with chatgpt. And all these discoveries are is just changing trajectory and plans due to complications that arise from implementing any significant advice by the AI. Chatgpt is a data skimming machine, btw. It will ask you a lot of questions about your work and ask for a lot of photos but in return won't even be able to draw correct pinout for an IC half of the time.
The most insane part of all of this is that you pathologically defend chatgpt that's not even specialised in robotics or electronics but disregard real critique and advice from people with experience. You think you will become some kind of robot master while being led on a wild goose chase by a mediocre chatbot

>>3005837
>code just enough AI that the AI can begin coding itself
why do you think you can do this when every major tech company is attempting to make their already capable ai systems better through self training and ai assisted engineers with literal trillions of dollars worth of incentive
not to mention this is near the top of the list for most major governments

obviously if the best newest version of claude fable with unlimited vram and unlimited processing power couldn't make a better ai, why can you?

>>3005940
>Game changing.
What's even more astonishing is that people used to design and assemble absolutely amazing circuitry before AI was even a thing. The whole consumer/business chip ecosystem was set up in such a way that you should be able to find what you need because that's a sale for the manufacturer - so you had whole yellow pages sized catalogues of ICs and advent of internet brought in databases with every possible spec to filter your searches on wholeseller's website. All you have to do really is to go to a website and select what you're looking for. chatgpt has nothing to do with this.

Your problem though is that you don't really know what you're looking for because you have no meaningful knowledge in the field and refused to gather it for past 12 years. Think about it; CTE takes 4 years. You're not a child anymore so you'd rather go to community college, which takes 2 years. Both of these institutions would teach you how to build basic electronics, giving you a good deal of momentum to build up on. Instead, you absolutely wasted those 12 years of project time so you're condemned to buying bogus parts and playing with kapton tape for 6 months by your own ignorance.

File: 1778737862237202.jpg (33 KB, 323x342)

33 KB JPG

>>3005529
>>if you can't be bothered to improve at soldering/handling electronics and try again?
>the first try I probably got 98% there but it had to be 100% perfect to work.
No shit, why don't you try again and improve to 100%?
>And finding that last 2% was extremely difficult to do because you can't inspect connections that are visually obstructed on the bottom of a chip blah blah my scenario is so dramatic
If you can't inspect them then don't - just solder again. Do you think SMD soldering is some kind of fucking epic feat? Why can't you be bothered to improve at soldering?
>Now I would gladly do this,
Yeah you would if you were good at soldering, but you're not. You can't even hypothesize because you simply can't, you can't choose to do this because you never chose to improve.
>but at that point, seeing how the frustration ramping was climbing rapidly, I looked for an easier alternative. To potentially act as a release valve of mounting frustrations.
Yeah you gave up after one failed attempt

also
>frustration ramping was climbing rapidly
Jesus Christ we're talking about one smd here

>>3005837
So, still 0 progress, no prototype no nothing?

for whatever abuse I get on 4chan I know one thing: if it weren't for you guys I would not have found the ESC I'm going to be using for my robot's motors so I'm thankful to you guys for that. That's a game changer for me. I'm working on its firmware now.

>>3006043
CAD for all systems of entire robot done in details fully - check
Physical skeletal structure of arm fully assembled and built and tested - check
First motor mounted and custom pulley based downgearing system designed and implemented (albeit roughly) - check
electronics for that first motor made ready - check
custom firmware for that first motor coding - in progress.

>>3006054
I respect the fact that you want to be fully vertically integrated but don't you think testing one (1) motor and one joint before doing the full design would've been smart ?

>>3006049
You know, for the ESC, there's a lot of 4in1 ESC that you can buy if you are concerned by size it might be interesting for you to look at, since one ESC can cover 4 motors. Actually I work in a very big company in Taiwan that makes ESC and QDD control boards so I might have a lot to say about your idea, I'm not going to read all your threads tho so just TLDR me if you are interested. For the firmware part you can consider Odrive, they have published the code already

>>3006063
No I don't. To know which motor would fit and would be strong enough you first need the full design of the entire robot to have the ability to determine how much space you have to work with, how strong and fast it has to be etc. Then you can pick an appropriate actuator and select appropriate downgearing and if appropriate downgearing does not exist, make your own - which I did. Only then are you in a position to make the robot frame, mount electronics, mount the customer downgearing, and eventually then be capable of testing the first joint. And that test will expose all issues with the downgearing, the motor, the electronics, etc. Which is what we are moving toward ASAP but has eluded me for a long time due to life and due to rabbit holes upon rabbit holes I had to traverse to get this far.

>>3006065
no, I don't have space for a 4 in one ESC. Having each ESC separated enables me to distribute it one per motor next to the motor whereas a 4 wide would be too big to fit anywhere in the arm. The ESCs I have do fit already and will work I think. As far as Odrive, I thought that's FOC firmware made for Odrive boards. I am not aware of it working across just any ESC like you imply. I don't think it would. But in any case I don't think it would have the specific behavior we want and it probably is closed loop which I plan to not have in my firmware. In any case, the firmware of the motor has a HUGE role in overall control of the robot and latency as well and how the communication is as far as robustness and latency and so rolling my own firmware will ensure success in all of these critical areas and so I'm not open at this point to any firmware advice I'm already in DEEP to making my own and very far along and its going very well and I'm very positive about it, have great momentum and have total confidence in what I'm making. What I will say is I recall Odrive had like $100+ motor controllers and so I consider that company to be predatory ripoffs.

>>3006054
Sticking primitives inside a 3d skeleton still isn't CAD, my dude.

>>3006090
>designing a robot on computer 3d software is not computer aided design (CAD)
ok I used my computer to make my design but just because the entire design was aided by the help of a computer doesn't mean it's comuter aided design - says the retard?

>>3006099
>make my design
but 'your design' is the equivalent to drawing a human outline and then drawing a series of circles and lines and labeling them things far more complex

>>3005837
This is Terry Davis level deranged. You don't even know what you're doing.

>>3006503
Respect the schizo dedication. Try to guide him tho. Maybe you can say stuff like : "oh maybe you should test one motor before building everything" or "maybe you should improve your soldering and order a PCB directly" or even "Maybe you should consider that PhDs with years of experiences can't make the AI you describe so maybe just stop being delusional" or something like that to help him slowly recovering from his pipe dream

File: 1768896020865051.jpg (1.75 MB, 2890x2530)

1.75 MB JPG

>>3006503
Terry was formally and practically educated in the field of computer science and he actually designed and produced his own variation of C language, compiler for it, bootloader and an operating system - which makes him one of maybe a dozen of people on Earth that could make this claim. His only weakness really was his mental illness and maybe lack of connections.
Arty on the other hand has absolutely no idea about any of disciplines he engages in and lives in a fantasy world despite having no life ruining psychotic conditions - evident for instance by the fact that picrel is his idea of an electornic circuit schematic, and picture of original post of this thread is his idea of CAD of a humanoid robot.

>>3006565
He needs to be broken down in a shock therapy like in some kind of bootcamp, amigo. He clearly works on this project for an ego boost and attention, considering how he consistently puts his convenience and uneducated ideas above any form of efficiency. He's obviously protective of his egotistical way of doing things and for any kind of reasonable advice he will find dozens of excuses, in writing which he puts more effort than into the actual project. By not giving him any pushback you just give him an opportunity to make up a random excuse and stay assured he's right, effectively enabling his behaviour. Once you call out his bs you wear him down by negative association and give him an incentive to question himself before he makes a statement and/or a decision, so he doesn't turn up to be a dumbass failure again and hurt his ego. You can't appeal to any possible currency other than his ego.

uh oh.... he might be scared off again

File: 1786913923601496.gif (547 KB, 136x119)

547 KB GIF

>>3006503
But this is a good thing.

>come to /diy/ for kiddie level lego mindstorm shit (btw, what is the modern equiv? Rasp-pi?)
>discover autist trying to build androids

Also fuck lego mindstorm. That shit was ass 20 years ago and I'm glad junior comps gave up on it. Looking for shit to make with a niece.

>>3006917
>>discover autist trying to build androids
It's actually frustrating because this project has been going on for 12 years and produced nothing. It's stuck on surface level promise that will never get delivered, plus arty despite being a moron is a calculating dark triad person so it's really not the same as watching some confused innocent sperg

>>3007150
did you actually think the guy who says he's building a robot with a mech lungs and a heart to fulfill cooling and something like a few hundred motors with fishing line 'tendons' and somehow has the ai ability to manage everything autonomously to the point that one of his self designed arms will simply build the rest of the robot...
was going to actually do anything?

the fact that he thinks that last part is true, yet can't bring himself to just purchase literally any existing robot arm to build his robot for him might be one of the more baffling things
i think he last explained that one away offhandedly by saying an existing arm wouldn't look enough like a human and/or his would somehow be superior

>>3007167
>did you actually think the guy who (...) was going to actually do anything?
Obviously not, but in such scenario normally a person would pump out at least some half assed prototype and/or eventually give up and move on. Here we have a specimen that keeps on failing with no progress since Obama administration and he's continously unresponsive to the failure. He's perpetually stuck at the beginning of dunning kruger curve as if he was new to the project and that absolute lack of character develooment comparable to genuine mental retardation is frustrating to watch.

>>3007407
This claim of no progress and keeps on failing is not true IMO

Robot frame:
Huge successes
- fiberglass version of skeleton went very well and is a WIP with one arm and hand done and half of ribs done and 1/5th of spine done. Will return to that in the future to finish it and build a robot off it (Adam)
-metal skeleton - looks very viable plan to use it to build a robot and have one hand skeleton mounted to it successfully. Plan to use it to build Dinah - so its a successful progress
-PVC medical skeleton - discovering this alternative skeleton method in itself is a great affordability and labor savings success. These seem very robust and up to the challenge. Have successfully been using the arm of one of these to build the first arm that will build the rest of its own body and then more robots etc. The "seed" that grows into a tree. The arm has all been attached with artificial ligaments taped and sewn into position which has been a successful method for attaching bones together - huge successes compounding.

So the only real dunning kreuger is the people who think they understand and can assess my project's success for failure and deem it a seriies of failures based on their own ignorance!

Robot electronics
motor mounting - success - a great method found for that
motor and electronics fitment: success - realized for all the degrees of freedom I want that fitting this much stuff into humanoid body volumetric constraints was no joke so invested the necessary hundreds of hours to build it all in CAD and it went swimmingly well and very successful. my designs all fit! huge successes and miniaturization was a success for pulleys and electronics plans. also cooling systems are very intense, elaborate, and brilliant and very innovative and were a major challenge to bust through. Solving all of that was a success.
-mounting the first motor was a success. is very solidly mounted in a way that won't budge and is robust and still allows for my cooling systems to work with it and is easy to remove it if needed later which I've done a few times already.

>>3007519
-pulley based downgearing - massive multiple successes. Turn in place downgearing was a huge success so far solving every issue with it from miniaturization solved, smooth low friction operation solved, finding appropriately sized thumb tacks for their axis to cut down labor solved, finding appropriately sized fishing crimp sleeves to act as plain bearings solved, building first ones and verifying their smooth and awesome operation solved, and figuring out a way to always keep them tight and pull out all runout and looseness always with the elastic bracelet cord method solved. So that is one awesome totally silent totally easy-ish to make downgearing method confirmed to a small degree in testing solved. Now just need to test it WAY more to iron out any kinks as needed. Next, the Archimedes silent pulley based downgearing system has also been solved, is easy to make, is very cheap to make like the winch in place pulleys method, and is miniaturized enough to work as well. It appears a great method and that it will work amazingly. I have tested it and it was a huge success so far in early testing. Still need to test way more to iron out any kinks.

>>3007520
motor controller: huge successes all around. designed multiple great motor controller designs including huge success after success at miniaturizing it all while maintaining peak performance needs all the while. Figuring out and solving all the ways this can be done was a great learning experience and confidence booster, proving I can successfully do electrical engineering at a high level. Early prototyping of these designs were full of successes from diy pcb etching successful methods found to deadbug methods and various techniques surrounding that perfected and successful as well. Then the icing on the cake was finding ESCs small enough, cheap enough, and strong enough to buy off the shelf and use as is - a HUGE labor savings pivot and major success for my project.
Software:
Big successes with working with arduino IDE to upload bootloaders to arduinos who had none and making my own circuit for achieving that, then big success working with arduino IDE to making a flashing new firmware/sketches to arduinos and making my own circuit for that. Then big successes with creating a brushed dc motor program in Arduino IDE and flashing that to a arduino and using that to drive a brushed dc motor. All great firmware experience and all successful. But now no longer planning to use brushed motors, however, those are all successes and will be reused for SOME brushed dc motors exceptions to the rule that we do plan to use and 10 of these are for the hands actually.

>>3007521
So far big success at developing my own custom firmware for my off the shelf ESCs to make the work for my robot actuation needs. Still a WIP with great promise and a great momentum and headway.
3d graphics - great success so far building my own 3d graphics rendiering engine which will be used in my internal physical world building realtime simulator taht the robot will use to motion plan and simulate his actions before taking them for planning purposes and will also be a physics engine that all runs in real-time and very lightweight and optimized all from scratch in c++. Is a WIP.
Web browser - from scratch web browser for robot to use to scrape websites for reading them and interfacing with them to do shopping or be a personal assistant as well as looking up things for his own research and learning. I can also use it for my own personal browsing needs in the future which will be nice. It will have a windowless mode so the robot can interact with unlimited websites in a lightweight way without using mouse and keyboard to research many things concurrently and interfacing with their various forms and input boxes without keypress events or mouse events freeing up those for other things. So far hugely successful and a WIP.

>>3007522
DIY operating system - so far a big success. from scratch operating system can make my robot even more lightweight and efficient and remove spyware-ish windows problems entirely in the future perhaps. a fun side project that is not mandatory but is a excellent leveling up platform for learning and mastering programming to a even deeper level
AI chatbot - so far a big success. so advanced it can learn and think and philosophize in theory according to its designs which are next level. Very innovative AI method. Is superior to LLMs in many ways. Is what robotics needs. Is the higher level brain function taht the robot will need to be logical and "understand" the world on a high level similar to human's higher level thinking. Will enable it to really be a great companion and assistant and "smart" and really autonomous, even taking initiative in researching things and watching youtube to learn on its own to fill its own knowledge gaps and better enable itself to achieve its goals. All coded from scratch - industry leader IMO
Vision system - so far very promising developments here. Huge progress and very promising. all coded from scratch. Industry leader IMO.
speech recognition and synthesis - so far very promising, all coded from scratch, excellent headway so far, very exciting. Should be lightning fast, lightweight, and highly accurate unlike anything available off the shelf.

>>3007518
>Huge successes
>WIP
>half done
>1/5th done
>will finish it in the future
>plan to use it
>these seem very robust
>>3007519
>great method found
>realized I want
>mounting the motor was a success
>>3007520
>finding finding finding
>method confirmed to a small degree in testing
>Now just need to test it WAY more
>is easy to make
>It appears a great method
>huge success so far in early testing. Still need to test way more
> great learning experience and confidence booster
>>3007521
>designed multiple designs
>buy off the shelf
>big success working with arduino IDE
>>3007522
>So far big success
>Still a WIP
>great success so far
>Is a WIP
>It will have
>So far hugely successful and a WIP.
>>3007523
>so far a big success
>so far very promising developments here. Huge progress and very promising.
>so far very promising, very exciting
You have completed nothing, there is no measurable success, just endlessly chasing the rabbit.

>>3007518
>Have successfully been using the arm of one of these to build the first arm that will build the rest of its own body and then more robots etc. The "seed" that grows into a tree.
if this is the true then an existing alibaba collaborative robotic arm would solve the entire thing

clearly, if the ai is capable of building the rest of its own body, it would be similarly capable of using an arm to build all of a robot
and also, as you seem to be the singular conduit of physically manipulating things on behalf of whatever ai system you are using short circuiting that by giving it direct access to an arm is the obvious path

>>3007662
not true. The amount of time it would take to program the AI to use a alibaba arm to build the robot arm would be prohibitive and a alibaba arm would just have a pincher. I am not guaranteed a simple pincher can build a robot. It's not as sophisticated at grasping and holding human tools and stuff like a high end human arm and human hand would be with full human DOF. So I am not convinced this is even viable. The other issue with this proposal is that extra work to teach a industrial robot arm to build the humanoid robot arm is then causing me and my AI to master industrial robotics which is completely off the path of my robot project it ends up a HUGE side tangent that may take even more time than simply building the humanoid arm myself then the AI for that arm all of which I had to do no matter what. So that is the straightest and most efficient path then. You are treating this as though the same AI could use any arm but it can't. You'd have to master industrial robot arm AI that does NOT cross over into humanoid arm AI so adding years of coding work for a one off operation that does not them have any use after that one time. You don't code AI and automation for single use. You code AI and automation that will be used countless times which over time pays you back for that initial effort. I'd never get paid back for that initial effort if it does it just once.

>>3007559
a step completed or a entire subsystem completed IS completing things for hte robot. Even if I plan to revise or test and approve with no revisions, the thing completed still counts as being completed. So that said I completed a TON of stuff.

>firmware for highly complex body part kinesthetic systems running on arduinos
>MS Paint schematics
>"DIY operating system" success!
I'm not convinced you have a single working proof of concept for any part of any robot.
If you post any proof of what you're claiming besides a vague screenshot I will eat my shoe.

>>3007686
You're being misleading to anyone who doesn't already know what to expect of you.
Take
>Big successes with working with arduino IDE to upload bootloaders to arduinos who had none and making my own circuit for achieving that, then big success working with arduino IDE to making a flashing new firmware/sketches to arduinos and making my own circuit for that. Then big successes with creating a brushed dc motor program in Arduino IDE and flashing that to a arduino and using that to drive a brushed dc motor.
for example.
All of this is child's play. Literally, actually - there's children playing with arduinos and doing all of that. In order to "create a brushed dc motor program and flashing it and making a circuit for it" you just downloaded a premade library with example sketch provided, connected 3 wires over a breadboard or whatever and pressed "upload" with adruino connected via USB to your PC. Is this success of great magnitude requiring it's own paragraph in the context of a project about humanoid sentient robot that makes itself? Do you realize how you sound when you're presenting this as a big success and a noteworthy step in process of making a robot? You could as well add "I brushed my teeth today and tied my laces".

Then;
>Early prototyping of these designs were full of successes from diy pcb etching successful methods found to deadbug methods and various techniques surrounding that perfected and successful as well. Then the icing on the cake was finding ESCs small enough, cheap enough, and strong enough to buy off the shelf and use as is - a HUGE labor savings pivot and major success for my project.
This story played out in previous thread and it was essentially that people kept telling you to buy the controller, you kept buying bullshit like kapton tape and only after a failure to solder you gave up and bought it - how is that a success, you never even used methods you "found out" because you opted to buy instead.

I could go on.

>>3007672
bro it has been used 0 times by 0 robots to do 0 things. for 12 years. your "countless uses" is just a lie to keep your attentionseeking going. industrial arm ai doesn't cross over into humanoid arm aiexcept the entire brain of the operation, planning, kinematics, force control, the actual thinking transfers the second you swap the wrist. the part that doesn't is like a few weeks of gripper shit. you just made up "years of non‑transferable work" to justify doing nothing for another 4 years.

>>3007672
>industrial robot arm AI that does NOT cross over into humanoid arm AI
How so? How can you even tell? You're claiming you'll write truly sentient AI but it won't be able to operate two kinds of the same tool without some hardcoded libraries that will take 12 years to write?

my discord for my project btw: discord.gg/AdbKtvurh

Hey OP, can you post a recap of your actuator strategy ? I'm interested in what you think is the best way for moving the limbs.

From what I understand, you use an electric motor to drive a pulley reducer which then moves the limb using a cable.

What kind of motor, BLDC ? Do you use something like an Odrive ? Paired with the pulley, that would give you a backdriveable compliant actuator, which for your project is really important I think. A stiff non-backdriveable actuator won't be able to give you the life-like movements you want (at least for walking).

What kind of joint are you using ? Hinge or rolling joint ? If you use cable-based actuators with hinge joints, you might run into problems where moving one joint (ex: the elbow) will change the tension of a cable used for another joint (ex: the wrist) if this cable doesn't go throught the axis of rotation of the elbow hinge perfectly. May or may not be a problem depending on the alignment error and the degree of positioning precision you require. Maybe you could also compensate this in software.

Do you have a protoype limb yet (in sim or irl) ?

they don't realize i NEED to build that arm. its PART of my NECESSARY personal fulfilment journey. I WANT to do that. but arm #2, leg, rest, I can pass because IMO its more rinse repeat of all the challenges and methods the arm already proved and solved and is actually easier than the hand/arm was. so I feel FINE to hand that task off to the robot to do at that point and having the robot build the rest of its own body is probably far more impressive than me doing it to boot. so I want to be a decorated roboticst with my own hands and then AI can be my focus after that and forevermore from there is fine even. but I don't want to exit the physical field of robots as far as MY hands BUILDING until I have at least achieve that seed bootstrap stage. then I'm ok

>>3007818
Great questions and best post I've seen for ages. I'll answer thoroughly:

>electric motor > pulley reducer > move limb with cable
yes exactly. well several pulleys and whatnot but yes

>bldc?
yes

>something like odrive?
more or less. its going to be a hacked off the shelf ESC given software that makes it behave like a Odrive would but simpler than odrive and $3 instead of $200 per ESC and about 1/50th the physical size envelope for saving precious space.

>for your project is really important I think
yes agreed

>Hinge or rolling joint ?
neither - well sort of rolling joint but ball and socket/rolling joint mix. I'm using a PVC medical skeleton so the exact joints our bodies use i'm using. Minus cartilage for now so more grindy. Might need to use Vaseline (or Delrin plastic artificial cartilage?) in there some day to slow wear? Not sure. We'll see how it holds up.

>moving one joint (ex: the elbow) will change the tension of a cable used for another joint
this can be a challenge but bicycle brakes solved that the same way mine will. the slack on the cable sheathing is the key.

>Do you have a protoype limb yet (in sim or irl) ?
that's kind of hard to answer. I have a assembled skeleton with artificial ligaments for the arm and first motor mounted along with its silent pulley based downgearing and routing to a finger joint. All I need to do to test first moving joint is mount the ESC, finish the ESC firmware, then send the ESC commands by way of a arduino connected to my pc via usb and boom, moving finger joint that we can test to failure. Is it a finished prototype of the entire arm fully working? No. Its a WIP. People in here say I have NOTHING so I have to say no, I have no prototype because that would entail I made something and that would be a lie. So I made nothing and have done nothing apparently.

the other huge problem is if I had even a off the shelf full humanoid, say the neo robot could build my robot, that doesn't really save much time honestly... now i'm just programming a neo rather than focusing on the iterative designs. i'm doing programming a bit OUT of order IMO. if I nailed the whole arm I'd be sufficiently satisfied with hardware to move on from hardware and let the arm do the rest. I've proven enough. but if I have not done that much, moving onto AI only dev for decades is not ok for me. i'd feel i'm leaving behind potential there in hardware land that needs to finish first. I'm a hardware first AI second builder/roboticist. Some are AI first hardware last. It's tough to say which is superior. Some could say build a robot and off the shelf AI can animate it years later let it come to you. Some could say build the AI and wait for the hardware to come to you off the shelf years later. Let others do that part. Some could say build nothing and let off the shelf AI + hardware completed robots come to you and just watch TV. It's complicated and personal preferences sway which path you choose. I'm set on hardware first then AI BUT I am a bit AI ALONGSIDE the hardware in between because I just can't wait. So I'm a bit hybrid actually.

actually i think people generally DO NOT consider total envelope constraints for their humanoid builds and fiddle with actuator options without considering that, and of the ones who did consider it, they OFTEN over simplify to save room and sacrifice as many DOF as they can to make it easier with goal to minimize DOF for MVP and they also don't care ENOUGH about final aesthetics other than final client at a factory will buy or w/e. they don't go for the aesthetics holy grail of perfect human beings ai generated objects of utter perfection tier etc

i can honestly say i think i've nailed the AI foundations, NAILED the frame, NAILED the motors selection and placements, NAILED the downgearing, NAILED the motor controllers, am NAILING the firmware, and NAILED the skin approach, i have them ALL everything NAILED. cooling systems too, chef's kiss. its ALL there. I have it ALL. people are SLEEPING on this

in a way, honestly, my slower approach is more efficient because it allows the collective to chime in for longer and as much as I may resist suggestions at first or not find HOW they can be implemented, i find so often the suggestions ARE steering my path and the collective consciousness of my audience IS enabling me to prioress albeit slower in a straighter line avoiding years of redos. i don't work on it as much as i'd like but that plays in my favor. For example, few years back, if i had built all the arm and got it moving, the suggestion to just switch to bldc instead of dc motors would be a non starter but since i only had to swap the FIRST dc motor I had installed so far at that time, it was doable albeit hard to suck up and tear out the first motor i already felt good about

you say "I'm deliberately allowing the design to mature before I lock it in. " --- that's not true. its a natural outcome of work life balance and having no funding to go full time on this. which forces it into hobby territory competing with netflix and chill and YouTube etc. so its harder to get done than clocking in 9-5. you do it as much as you can. but the side benefit of this is the information flow is heavier between each step so its a accidental benefit and DOES improve outcomes i think oddly as that may be. it isn't the goal to get done fast and ugly but to get done at near peak tier excellence and that requires the meat to marinate

in 2003 on some fembot fan forum I found, one guy wrote: "Like people do like to own a nice car, i would like to own a fembot.

To me, fembots have no mind of their own, they are just machines running complex programs and executing routines and subroutines. But machines that i would like to have pleasure with, sexual or playfull, like a kind of walking Playstation for adults... :P"

its amazing that forum had tens of thousands of posts for decades running and yet despite them ALL wanting desperately to own one we have NONE we can buy 23 years later. really proves how hard it is yet 4chan acts like I should whip one out in a year tops while maintaining work life balance, a job, real estate investing, a family, etc

even ones like moya and xpeng iron which come close are not perfect and look a bit rough and move slightly rough and stuff like that. haven't seen dextrous manupulation from them either - doubt they can well. elon musk said the hands are like 40% of the entire engineering challenge so i'm sure those companies more or less skipped that part to a great degree frankly. its my main focus. but like if its THIS hard, then me kind of going slow and carefully really thinking hard and taking every step very very cautiously and thoroughly is NOT dumb. it saves me countless hours of failed prototypes and potentially MIGHT lead on a more direct path. but they act like spam out first idea and it all just will work one shot style

OMG I LOVE the visual picture of crossing a huge wilderness with this project as a visual analogy/metaphor! it DOES feel that way! and that's also why its so dumb that I am MILES AND MILES deep in this trek, have crossed streams and mountain ranges, entire countrysides on this journey and some idiot on 4chan says I've completed nothing as though I was at home trying to find my hiking boots in my closet still like WTH!?

i mean its literally been a good 10k hours no joke if you include ALL chatgpt talks, ALL my youtube video watching, ALL forum reading, ALL prototyping, ALL coding, ALL actual physical work sessions, ALL discussion with family, ALL imagining and problem solving laying in bed etc. No joke!

if you spent 10k hours on a wilderness trek, 4chan saying you've not treked at all yet is what is so absurd - they emphasize NOTHING has been done.

I FEEL EXTREMELY accomplished regardless what they say. i feel i have a VERY decent grasp on bldc, the terminology, the tech stack, the methods, what does and does not work for a humanoid, the overall plan and options and pitfalls and approaches and overall solutions. in the deepest rabbit holes of minutiae perhaps i get a bit rusty or vague on details when zoomed on on something 10x level but at birds eye view and even 5 foot view I'm VERY VERY sound and knowledgeable on it ALL

this IS 100% knowledge work if you think about it. vastly so and seriously so and predominantly so. if someone had ALL knowledge the path would be super straight and somewhat quick and to the extent you lack the knowledge the path is largely a learning one honestly. even when you TRY to build you end up just finding why the thing you built won't work 99% of the time which can be discouraging. so simply learning almost feels like the most efficient path forward but yet 4chan pretends you can just one shot this goal: >>3005837 ---. Like the first actuator you pick with ZERO knowledge in robotics or electronics on day one of your journey will be PERFECT, the FIRST frame you pick will be ideal, the FIRST AI approach flawless. its STUIPD to assume all of that from them. they treat the build like its that simple

When I said ALL knowledge I meant as in omniscience and omnipotence. then it would be trivial. simply a knowledge of EXISTING frontier humanoid robotics ALSO would not get you far enough! not if when you start the best robots were in darpa challenge. they are miles from the goal I stated. so my knowledge had to far exceed even what was being DONE by others and dive into NOVEL and INNOVATIVE solutions that did not exist yet when I STARTED even though over the past 12 years more and more answers are coming out from other people doing it too

>>3007827
lil nigga things his lil drawings with no real moving hardware mean he's solved it irl.
Put your retarded little creature in fucking mujoco and your NAILED AI, and let us see it walk.

>>3007832
anyways those are some highlights from a long talk with chatgpt i had tonight. I love talking to chatgpt for ages. Its like a personal diary and I write my thoughts and hear some commentary about them and that leads to more thoughts. Its so fun and satisfying. I was thinking I should publish our full conversations on my website at some point if they were really good. I personally want to read other's conversations with chatgpt if its about some subject i'm interested in. I want to see what peteblank and chatgpt talk about with the miku project of his for example.

File: screenshot-18_09_58-11_09_2026.png (5 KB, 511x385)

5 KB PNG

>>3007822
>this can be a challenge but bicycle brakes solved that the same way mine will. the slack on the cable sheathing is the key.

How so ? Do you pre-bend the cable sheathing to compensate for the required cable slack when the joint is rotated ?
See the diagram I drew: pink is the wrist actuator, green the wrist cable, red the wrist cable sheathing, the black sphere is the elbow joint, the yellow arrow points to the intentional pre-bend.
Is this what you are talking about ?

If yes, I wonder how to prevent the cable tension from messing up the sheathing's bend.

Larry don't you think it's insane that you can barely solder a chip with help of chatGPT but you're designing a whole pulley system in your head with full confidence?

>>3007839
yes that's the idea. However, the bend has to be large (in terms of length of cable sheathing involved in the bend) and gradual instead of a sharp kinked bend like you drew. A larger (in terms of length of sheathing involved in bend), gradual bend distributes the cable tension and friction along a larger length of cable so it doesn't cause issues so much. Also there may be cases where you want to encase the low friction PTFE tubing sheathing in a metal spring outside of it as a type of sheathing for the sheathing which will add the additional property of making it uncollapsible if the tension is too great and its causing the PTFE to collapse in on itself in terms of crushing. The type of crushing I mean is like if you place a can vertical on the ground and stomp on it. Its walls literally collapse. That CAN happen to this type of sheathing if you don't have the spring and the tensions are greater than the sheathing's walls can bear. This is why I go with thicker walled sheathing for the finger joints to give greater wall strength in those areas in hopes of preventing that kind of failure mode. But I won't know if it will work for sure until its tested more.

>>3007842
chip soldering I did fine, but some unknown short or even a static electric discharge made the chip fail. That doesn't mean I cannot solder a chip. Plus, QFP chips are designed more for hand soldering but QFN chips are not really supposed to be hand soldered by design. They are designed moreso to be baked in a reflow oven or installed with a hot air station. Neither of which I was able to do because my PCB design was made in such a way that the solder tension when molten would not be equally pulling in all directions for self centering of the chip which meant the PCB was inadvertently designed to utterly fail if you heat it in a reflow oven or hot air station method as it would violently throw the chip out of position due to the physics. None of which I had considered when doing the PCB design. So all of those factors really have no bearing on whether I am able to accurately design a pulley system with full confidence in my head or otherwise. Plus even if designing the pulley system in my head were unlikely to do perfectly, I was able to prototype the pulley design and test it iteratively perfecting it further on each pass. So it wasn't merely a one shot perfect pulley design in my head but rather a iteratively developed through trial and error pulley system design that was based on pre-existing successfully made and demonstrated pulley designs already used for thousands of years. Hence the certainty that in principle the designs will work if executed well. It's not 100% novel by any means.

>>3007910
>but some unknown short or even a static electric discharge made the chip fail.
no, it was your dogshit process from top to bottom that caused it
literally every step you took contributed and added up to a guaranteed failure that was unavoidable the second you started down the absolutely retarded path that you chose.
the fact that you cannot even acknowledge your own failure without trying to make excuses highlights your own retardation and inability to properly assess your own progress/possibility of success

>>3007910
>chip soldering I did fine, but some unknown short or even a static electric discharge made the chip fail.
How can you tell? You said you're unable to diagnose it.

> That doesn't mean I cannot solder a chip.
Okay, let's assume you did a fantastic soldering job, but you either chose wrong tool or wrong chip or bricked it by handling it badly. You can choose any possible type of incompetence to be the reason of failure in case of that chip - it doesn't change the result. A simple SMD solder went wrong. You set it all up and you executed it, so it's on you.

Knowing that you stumble in such ways on basic tasks, isn't it insane that you're confidently discussing a whole pulley system? You used to disregard advice and estimate chip size based on your assumptions of the pulley system. There's a clear pattern of unexpected failures and problems mounting in your work, but you keep having surreal ammount of confidence in that pulley system.

File: 1781647965820772.png (125 KB, 344x342)

125 KB PNG

It's very amusing how arty's success bullet point list was minor stuff hiding behind broad, big words like
>OPERATING SYSTEM
>COMPUTER ASSISTED DESIGN DOCUMENT
>INTERNET BROWSER FROM SCRATCH
>MOTOR CONTROLLER SOWFTWARE (IN ARDUINO)
but when it comes to analyzing his failures on borader topics, it's always some very specific small detail that caught him off guard.

When he's talking about his unfinished work, he's focusing on desired results, but when he's talking about his faulty results, he delves into technicalities. And each time this technicality is usually some newbie mistake. You're supposed to believe he's SUCCESSFULLY developing an internet browser in c++ from scratch (he doesn't even know what 'headless' is, he used word 'windowless') but at the same he's so retarded he didn't know he's supposed to space out pins and pads properly on a pcb. He doesn't even know how to find a chip for his project.
You will never see him talking about the technicalities of his operating system or whatever SO FAR SUCCESSFUL project he's working on, because it's just so unimpressive you wouldn't call it an operating system. Only once he fails and ragequits the project, he's finding a scapegoat in form of a technicality - and that's when you learn he's retarded.

>>3005837
Start small retard. Make a child slave first then move up to a woman

>>3008256
You can't start smaller than perfecting moving a SINGLE finger joint which I'm doing first. So all this start smaller, start easier, etc NONSENSE in no way applies. A single finger joint cannot be dumbed down any further than that.

>>3008285
>A single finger joint cannot be dumbed down any further
sure it can, remove the joint
a three clawed 'hand' with a 2 axis 'wrist' should be plenty of control to grip, manipulate, and otherwise interact with anything worthwhile
do you NEED a pinky? a ring finger? would losing those two digits really impact YOUR life?

>>3008294
no, 3 claws and 2 axis wrist is NOT simply than actuating a SINGLE finger joint. And copy pasting that single finger joint is trivial. So later on steps to make a claw is literally retarded. And ugly, and would be extremely less performant and not look human at all. This has to pass for human. Really its like you guy have no clue what the goal even is. Also, no, a single finger missing VASTLY reduces a human hand's performance. It is shockingly so. Every finger is absolutely necessary and amazingly so. Its amazing you guys live with hands your whole life but fail to observe them enough to where you think a simple claw can do as much as them. You must have zero skill in your hands to think a claw can match them.

Have you made the working finger yet?

>>3008301
>later on steps to make a claw is literally retarded.
Now you're throwin stones from a glass house, buster

>>3008303
everything is setup mostly except the firmware to drive the motor which is about half way done now. I had to develop a custom communications protocol and commutation methodology to do this in the style that was easiest and best path forward IMO with no need for back emf nor encoders. Updates comign soon on that end btw. But yeah things will be really epic the day that finger starts moving i feel at lesat for me. but I guess the whole journey is as epic in its own way. Just having a moving robot is a new level of cool factor mainly for me.

>>3008307
how old are you calling someone a buster? 60s? Not knowing the age of people makes it harder to respond appropriately. I respect my elders. anyways... yes, I am very serious about a human passing robot. so a claw cannot cut it. and no, i don't see a claw even remotely close in what it is capable of doing as a human hand. can you imagine cutting off your hand and using a freaking CLAW to try to sew, pull out an eyebrow hair, dig into your pocket to find your keys, hold a drill and depress its trigger, turn a wrench, catch a football, shoot a basketball, grab and swing a golf club with perfect form, play pool, give a massage, wipe a table, or 10k other things? I mean you are probably like a claw can pick up a pop can so its just as good as a hand at grasping. But that is just ONE stereotypical example. When you get 20k more examples the claw falls short VERY HARD VERY QUICKLY.

>>3008314
>i don't see a claw even remotely close in what it is capable of doing as a human hand.
Salisbury hand much? This shit was solved 40 some odd years ago. You could build better in your garage nowadays.

https://youtu.be/Q2z8pJMN1dM?si=RAk65PuFShgBjEqB

The Utah/MIT hand adds a thumb (same basic technology). You're reinventing the wheel (badly).

Anyways, I did not rage quite the DIY ESC. I met resistance and that felt significant and I paused to investigate some other options as one should I feel in a big project, continually re-evaluating, testing resistances of various branches, and choosing paths of least resistance as they present themselves so as to not do needless extra steps or extra work which slows the project down. It's not a lazy thing, it's efficiency maximizing. But at times SOME paths thought easier do NOT lead to teh desires outcomes and you rule them out based on that. A claw does not lead to a human shaped hand. it serves no purpose other than a pointless derailment and side project that progresses you zero toward the goal. Every path chosen must always necessarily advanced you toward the goal and the goal of this project is plainly stated in the OP post.

>>3008342
When will you discover a path of least resistance called buying a robot arm to test out the concept

>>3008375
this thread and project announcement from its inception was not ok i'm going to get a job and save up to buy a robot. This is a /diy/ forum dedicated to people building things themselves. I did not set out on a grand journey to buy a robot as you foolishly suppose. I am describing path of least resistance to ME BUILDING a robot, not path of least resistance in OVERAL LIFE APPROACH. Because if you go that far, the path of least resistance would be to just watch TV and not build anything at all. Saving time and money that way. So please you point is retarded..

>>3008415
Why didn't you DIY the ESC then

>>3008439
you aren't supposed to make the PC, every wire from copper you yourself mined, the tools that you have to blacksmit, etc. You are supposed to buy components off the shelf if they are possible to buy and assemble them into a robot. When you build a diy electric car you don't build the motor yourself, the breaks yourself, the metal you mine yourself etc. That's not how this works. You build parts that you can't buy or need to customize but you don't just avoid buying everything for no reason. It's still DIY building a house if you buy lumber and screws you don't have to cut the wood down from a forest and mine ore to make screws/nails.

>>3008463
Why aren't you buying the arm then. It's not a whole robot.

what don't you understand about an intermediary step of a non-human hand/arm? why does it HAVE to be the final product? you've already declared the intent of building an arm to then have it be controlled by ai and have it 'complete itself'. so what's the big deal with making that hand/arm not the absolute perfection of a replica human hand. Who cares about extra work, once the arm is hooked up to ai it can work 24/7 and refine its own ability beyond anything you could ever hope to do.
delaying the connection of any physical manipulator to your ai is the obvious roadblock and your ego compelling your middle manager insertion of substandard work serves only to ultimately hinder your own fucking project

>>3006565
I was like you once. I tried to help OP. The sooner you realise he is simply mentally unwell, doesn't actually want to be helped and understand the difficulties involved in what he is doing (because it will break his narcissistic fantasy) the sooner you will realise the kind thing to do is bully him in the hopes he starts putting all this wasted time into getting a job so he can provide for his wife, kids, and parents who he mooches off of.

>>3006899
He literally can't leave. He needs his narcissistic supply, but everywhere else including real life has banned him.
Look at this (>>3007827). He's in a narcissistic high right now, NAILING everything (yet having nothing to show for it...) Have a look in the archives and you can find him confessing to cheating a narcissism test to obscure his narcissism, if it wasn't already obvious from literally everything about him.
>>3008502
An intermediary step would involve a resolution to the question "is larry retarded?" (yes.) Since he does not want this question to be resolved, even for himself, he must never test anything at a small actionable scale. Hiding behind grand scale designs allows him to put this off indefinitely, and his belief that he will be finishing this robot in heaven means you will never be able to check him for this, because he literally believes he will die before he does anything that can answer the question "is larry retarded?" (yes.)

>>3008463
>you just can't!!!! OK!?!?!

>>3008487
there are no customizable off the shelf realistic, passing for human robot arms so stop making stuff up.

>>3008502
Any step that directly goes outside of a direct path to final goal including going with a non-human hand/arm is a step I reject on principle. Every step by the principle must go direct path toward the end goal with no detours. Not only that but we humanize inanimate objects so that non human arm would become like a friend/family member like Thing on Adams family and it would be nearly impossible to destroy it or throw it away after so it would BE my first and only robot for decades which I do not want to be associated with. I only want to be associated with the highest quality human passing robots and nothing less or nothing other than. Also, IF I did get a non human arm to a point its building whole robots, everyone would just say no need to build anything else, patent or otherwise commercialize and mass produce that arm only and it can do all manner of amazing work and make big money. To make a humanoid then would be called retarded if I already have a non human arm able to do anything. Which would remove the focus off the original goal and shift the whole narrative onto this other thing which is now seen as the real prize. That would further derail, confuse, and distract from the original mission massively. And would infuriate everyone to just ignore and move on from something that amazing like it's nothing just because I want a realistic humanoid instead. So better to NOT EVEN GO THERE at all. And just build the human arm to begin with so that whole branch of realities is never experienced to begin with.

>>3008533
>Also, IF I did get a non human arm to a point its building whole robots, everyone would just say no need to build anything else, patent or otherwise commercialize and mass produce that arm only and it can do all manner of amazing work and make big money.
what?
a non human arm already exists, you getting one wouldn't unlock some magic realization that would lead to your imagined scenario
maybe the issue is you are somehow viewing this industrial arm as a replacement or substitute
it's just a tool, just like cad software or chatgpt is a tool
>That would further derail, confuse, and distract from the original mission massively
the only people paying any attention to your dumb bullshit are in this thread reading this right now and actively laughing at your fucking asinine 'updates' that result in absolutely nothing

>>3008521
I have decided to resolve to NOT die before completing the robots. I now have resolved to complete them all before I die and declare it with faith. The begrudging acceptance that there is a statistical strong possibility I don't finish before I die was a concession I made from a sober minded statistical probability perspective and is in part something I did from giving too much weight to naysayers and trying to hear them out too much. Better to just put my foot down and declare NO, I WILL WIN.

>>3008537
>your project is only seen or discussed in this one thread on this one place on the internet
not true at all

>the industrial arms already exist
no they don't, not with high level AI in them able to build anything. Actually it just hit me this is what iron man did in the first couple movies. He had that smart industrial arm helping him buld all his suits. So at least if I did go that route I'd have Jarvis and be even more like iron man which is pretty sick but still. Jarvis was a beloved character you could never throw away after it exists. It even had a personality. I want the ONLY beloved character with a personality to be my robot ITSELF, not a Jarvis intermediary step robot. Not to mention the space taken by a Jarvis arm would cramp the space of my bedroom too much. One humanoid in my bedroom is enough its NOT a huge room to work in. Plus I often have my wife and kids in here so its even MORE cramped then.

>because industrial arms exist, selling them with AGI-lite AI built in would not be possible
are you retarded? you saying a dumb and gcoded simpleton industrial arm is the same thing as freaking Jarvis on Iron Man movies? Hell no. If you could take the industrial arm and inject Javis, then sell that, it would sell like hotcakes.

>maybe you see the industrial arm as a replacement or substitute
it would be. IF the end goal of the humanoid is largely to do the same work the industrial arm would now be doing, then it would be hard to justify making the humanoid at that point. Which then would replace/substitute the humanoid. Or if I did still build the humanoid I'd be criticized for doing that since I already have a industrial arm doing the same work so it would be seen as pointless to make the humanoid.

Update: I now plan to use a USB hub to breakout USB from the main brains PC and have USB then run out as the main communications lines to the various Arduinos distributed across the robot's body. USB has the benefit of low latency, high bandwidth, and the cabling itself is already shielded and has ferrite rings built in so already deals with EMI concerns over these longer travel distances. Then to go from these distributed arduino hubs littered across the robot over to the individual ESC MCUs it will use my custom protocol single wire communication. Because these final communications paths are much shorter (think 5" or less-ish) then EMI is far less of a concern and our standard twisted pair wiring for the signal should be enough to make EMI a fairly non-issue for this.

As to my firmware, I now have a interrupt section written that records communications PWM and stores it in memory, paying attention to changes in the signal line and recording duration of low pulses and the pulse duration will correleate to encoded data being sent. The main while loop of the firmware then goes into that stored information on a scheduled routine and checks for valid communications stored there. It checks for known commands listed in there and if it finds one it executes on it. This too I have written. I just need to write the actual execution code next. However, I need to stop where I'm at and test and debut what I've written so far since the file says it has 1700 lines of code already and I have not tested nor debugged any of it so I'm due for that next. Once it is fully tested and bug free I'll move into the execution of the commutation stuff. I plan to test the existing code in windows first and then test it on the MCU of the ESC second. Well I might not even test it on the MCU at all for now maybe windows only and then test everything at once later on the MCU not sure. In any case, I'd say I'm about 60-70% through with the firmware making process now.

>>3008542
meanwhile in the real world :
https://youtu.be/GHySQMMrIa4
VLA are already able to do basically everything

>>3007764
>a project about humanoid sentient robot that makes itself
also robots/ai can never be sentient and to suggest perhaps one day they can exposes you as a blasphemer in danger of God's wrath. Say it again and get banned.

>>3008531
Nobody is saying there's off the shelf stuff like that, but why don't you just buy a robot arm off the shelf and feed it AI to make a realistic hand you want? Human fetus doesn't start with pair of complete hands as well, since you love analogies to much. You would skip the whole mechanical/electornic process and could focus on making the AI that would take care of the rest of the project.

>>3008569
already answered/explained this. but just one bonus answer: the hands on skillbuilding I personally learn with MY two hands is helpful in teaching my robot who shares the same hands structure and musculoskeletal designs on the proper form to execute the various tasks to build the robot the rest of the way. A pro golfer makes a good golf instructor. A industrial robot arm has no past experience with hitting a golf ball with human arms so would make a poor instructor for teaching a human to golf with their human bodies. I have to be able to code in or say hold the tool like this and demonstrate and know the way to hold it by experience etc. So the skills I personally build as I make the human arm directly assist my abiity to code the AI to mirror what I did. Just as the skills I have in thinking and talking directly help me code the conversational AI. If I skip doing ANY building and just a buy a arm, I can't code the skills I never acquired into the AI. Not only that, but coding an industrial robot arm to do anything is alien to someone who never lived their life doing tasks as an industrial robot arm. I have never BEEN a industrial robot arm so can't say how I'd approach a given problem since I have no experience with doing tasks AS a industrial robot arm. So coding one would NOT be intuitive and would be a slow slog and not wise.

>>3008533
>IF I did get a non human arm to a point its building whole robots, everyone would just say no need to build anything else, patent or otherwise commercialize and mass produce that arm only and it can do all manner of amazing work and make big money
why do you care? You don't listen to feedback anyway. Why do you even assume world at large gives a shit about your project

>>3008571
>the hands on skillbuilding I personally learn with MY two hands is helpful in teaching my robot
Okay bro, have you not acquired a decent skill of using your arms througout your entire life? Do you have to engage in some kind of extracurricular activities in order to learn how to use your hands because you have been just jerking off your entire life up until now?
>I have to be able to code in or say hold the tool like this and demonstrate and know the way to hold it by experience etc.
Do you seriously have to waste 3 months of your life on bricking an ESC chip in order to lean how to hold a soldering iron?
>Just as the skills I have in thinking and talking directly help me code the conversational AI. If I skip doing ANY building and just a buy a arm, I can't code the skills I never acquired into the AI.
You either have 'thinking skills' and can come up with way to code an activity using similiar life experiences or you're a complete retard with no imagination that has to go through the grinder. Inability to use analogous experience is a sign of low IQ - it's literally inability to apply your knowledge in a broader sense and reliance on repetition. If you want to tell us all you're a retard then it's okay but why do you assume you're going to finish the project then?
> I have never BEEN a industrial robot arm
You never will be btw, what an insane argument to hold onto.

>>3008542
>Jarvis
I just love how arty is endulging into these pipe dreams, daydreaming about the worldwide reception of him and his robot; how the arm is going to be his beloved sidekick with a personality, but only if it looks realistic. And the damn hurdle of cramming it in the bedroom along with the rest of the family - straight from 101 dalmatians, or ALF even, with some hilarious interactions between Jarvis and Arty household ensured!

>>3008567
>Arty thinks he can drop the banhammer
Fuck around and find out, asshat.

>>3008571
This is already how VLA models are trained you stupid retard, watch the video in this post
>>3008564, you can 100% teach a robot to do whatever you want even if it's not exactly the same structure as your body. To be fair this is kinda new and it didn't exist when you started your project

Will add "flesh" to them eventually?
Would they make good target practice as a stand-in for enemy combatants in motion?

>>3008567
a robot could be sentient one day, just not yours
and not because of some moral or religious thing, but purely because of your ineptitude

>>3008572
Because if large amounts of people unanimously and confidently and repeatedly tell me a thing that is logically and objectively and obviously true, it becomes extremely difficult to ignore when part of myself agrees wholeheartedly to boot. I can't ignore such effects and this would instill permanent doubt and frustration and what ifs that would be difficult to deal with for the rest of my life. Better to avoid that entire.y. Now to have a vocal minority naysaying my apporach or shitposting with no objectively good points being made and vague general insults that I tend to not take very seriously, particularly when most of it is literally made up out of boredom to boot, which is the case genearlly when it comes to my robot, of course I don't take that seriously or give it much heed. It's NOT the same as would be the case with a functioning elite tier coded AI built into and perfected onto a industrial robot arm that then can work wonders and make all manner of things. That WOULD objectively be a money printing machine and to try to make anything else ever at that point without absolutely focusing almost all resources to maximize the money printing machine's economic impact for me would be unwise at that point and downright frustrating to anybody observing the scenario. It's not the same thing at all as you foolishly suggest.

>>3008573
>Okay bro, have you not acquired a decent skill of using your arms througout your entire life?
While I may have a library of semi-related skills from working with my hands that do play a role in enabling me to learn new specific skills and techniques when working on the robot arm, taht does not mean it isn't extremely helpful and useful to develop specific skills and techniques specifically for the relevant tasks in building the actual robot arm by hand. Not only that, but the techniques themselves are developed to best match how a thing is to be made when being made by human hands. The way humans hands can best make a thing is how I end up optimizing how I make a thing intuitively and naturally. So if I never make anything to begin with, then it is less obvious how that thing is to be made when I'm just trying to hypothetically make it in my imagination while sitting at a keyboard with the thing not in my hands. Which would be the case if trying to build the robot arm just by way of coding a robot to do it for me without first doing it myself. Not only that, but building it myself has a iterative aspect to it where I can naturally fine tune the methods, streamlining them and improving their efficiency which would all greatly improve the AI coding instructions when teaching the human robot arm to make similar parts and do similar steps when constructing the rest of its own body. These factors cannot be ignored as is being foolishly suggested here.

>>3008573
>3 months of your life on bricking an ESC chip
so you blatantly ignore the iterative improvements I made in how the etchant solution is mixed, how to heat it, how to agitate it, how to block off copper from etching process that need not be etched to speed up etching of areas taht do need etching but must avoid over etching so as to minimize undercutting effects, and about 200 other very important revelations I had during that timeframe. Instead saying the entire time I was doing nothing but moving a soldering wand around exclusively. You are clearly arguing in bad faith and outright lying for all to see. What purpose does that serve other than to expose the blackened condition of your lying heart?

>>3008573
>code an activity using similar life experiences
Unlike you, I have built inventions where constructing the first prototype took an enormous amount of time comparatively speaking and where I fine tuned and optimized the process of constructing said product iteratively through experience so that by the time the product was being made by me commercially repetitively, due to so much hands on experience and little improvements on each pass, I was able to cut down the production of one product from 20+ hours for the first prototype to say 3-4 hours tops for the 50th build of that same object. That's a 17ish hours reduction out of 20+ hours on that first pass. And with taht increased speed also was increased accuracy and skill and the many mini optimizations and techniques that made it easier and faster in every way down to the tiniest of details. You can't get that just by pulling on relevant general life skills. You only got that through hands on experience. You are down playing and ignoring the importance of large amounts of hands on experience in doing physical tasks and the huge impact that makes on your skill and speed and technique which any person working in trades knows and any person who is a craftsman of any sort knows. And then while speaking utter nonsense ignoring these facts have the nerve to call ME the fool when you are hte one speaking utter foolish nonsense. Classic 4chan buffoonery.

>>3008577
I don't use off the shelf AI for my projects because its garbage compared to my AI IMO. So watching what some turd AI can do in a video is pointless. Won't click that link.

>>3008577
oh and if this VLA off the shelf turd AI were so good, Elon woudl have already used it and teleoperation demos of Optimus would not be a thing stinking up the internet and causing people to distrust robot companies. We also would not have robots barely able to do any useful tasks if it were working so amazingly like you are claiming. Instead we have no humanoids able to do much in the way of useful tasks beyond even the most rudimentary stuff that hardly justifies their existence.

>>3008586
If you want to spend $500,000-$1,000,000 to buy a target dummy you destroy with a single bullet then sure?

>>3008654
lmao this is next level delusional. Your "AI" doesn't even exist and yet you delude yourself into thinking it's better than the current state of the art.
>>3008655
>Elon is so smart that if he doesn't use something the something isn't good
first your statement is absolutely retarded and second optimus does indeed use vla now, the teleoperation videos are rather old now
>We also would not have robots barely able to do any useful tasks if it were working so amazingly like you are claiming
Nigger they are coming into factories now just watch the fucking video they can do a lot of things now you are just unaware of what's happening in robotics because you decided you were smarter than everyone else

>>3008653
>You are down playing and ignoring the importance of large amounts of hands on experience in doing physical tasks and the huge impact that makes on your skill and speed and technique which any person working in trades knows and any person who is a craftsman of any sort knows
NTA but this is exactly how vla works actually you show them some demonstration of a task and they can generalize and get better and performing it, the only problem is rn there's no way to upload human knowledge directly on the VLA as you imply, you should demonstrate the task multiple time to the robot for it to learn how to complete this type of tasks. Of course for most households tasks this is already solved and you can download a database already to train your model

>>3008649
>I can't ignore such effects and this would instill permanent doubt and frustration and what ifs that would be difficult to deal with for the rest of my life
So it's just about your ego? You're going to grapple with the whole planet? In your fantasies, at that

>>3008654
>I don't use off the shelf AI for my projects
You're constantly using chatgpt for help

>>3008542
>>the industrial arms already exist
>no they don't, not with high level AI in them able to build anything.
cmon man
really?
do you really think you will be the first person to give an ai access to a robotic arm?
do you really think this isn't something others might be doing already?

>>3008652
>so you blatantly ignore the iterative improvements I made in how the etchant solution is mixed, how to heat it, how to agitate it, how to block off copper from etching process that need not be etched to speed up etching of areas taht do need etching but must avoid over etching so as to minimize undercutting effects
Everything you just mentioned is steps to one simple etching process which is usually done right on the first try by a kid in high school doing a science fair project. If I'm wrong then please enlighten us all on your arcane ways of mixing the etchant solution or how you developed an improved way to shake the container during the etching
>>3008651
>So if I never make anything to begin with, then it is less obvious how that thing is to be made when I'm just trying to hypothetically make it in my imagination while sitting at a keyboard with the thing not in my hands
I think you're pretty good at imagining things, actually

>>3008661
>there's no way to upload human knowledge directly
Just as I suspected. Turd off the shelf AI. If you want something done right you have to do it yourself. My AI you can explain why each thing is done and whatnot. Really give it a full understanding of every technique in plain english with analogies and examples and it can grasp it in a deep way. Not just monkey see monkey do copying BS. When will these kids (VLA turd AI devs) learn? They should be using real AI - rules based AI

>>3008664
I didnt' say ego. I'm saying if people all say you made a mistake, and you feel you did, ignoring yourself and all of them to do something stubbornly that is not ideal is hard. The doubt and frustration and wavering it makes is hard. What does that have to do with EGO!? IT's NOT related to ego at all... It's moreso related to confidence you have been going the right path. You see, if you live a life of solid decision after solid decision, you can shape your life like a skilled sculptor. That's how I roll. So if I doubt a major sculpting decision, it can cause me to have less confidence in my sculpting (decision making) abilities which is going to make me less confident in future endeavors. Confidence and ego are NOT the same thing IMO

>>3008665
using chatgpt as google does NOT mean I'm using off the shelf AI APIs as are nested part of my software implementations. That's apples and oranges. I am assuming the training data of REAL PEOPLE is reaching my eyes THROUGH chatgpt as a stupid medium content aggregator. SO I am assuming I'm PULLING FROM real people's takes and the consensus view and the correct view often times when I ask chatgpt something. I say I'm not asking chatgpt, I'm asking the countless articles chatgpt read that were written by pros and it is going to sum up for me.

>>3008683
others are giving arms retarded ai written with laziness, vibe coded, turd ai. I'm talking about truly elite ai, the likes of which would never be open sourced. That's what I bring to the table here.

>>3008712
>My AI you can explain why each thing is done and whatnot. Really give it a full understanding of every technique in plain english with analogies and examples and it can grasp it in a deep way. Not just monkey see monkey do copying BS.
Few hours ago you wrote you have to go through every step of your manufacturing process in order to be able to code the AI properly with your knowledge in mind and now you're saying your AI will require only a bunch of theory (avaliable on the internet already) and analogies

>>3008710
no, they aren't etching with this fine of boards in a one shot way I don't believe that prove it. I followed all directions and it didn't work. undercutting WOULD NOT EXIST as a term and as a issue in etching if LITERALLY NOBODY EVER EXPERIENCED IT UNTIL ARTBYROBOT DID IT. How dumb.

>>3008717
its not just give the AI theory or just give it analogies or just give it instructions. its all of the above. the more instructiosn paired with understanding the better. Just copying is not enough. copying while also understanding the step and being then able to improvise or improve on the step ITSELF due to understanding the step in a deeper way is key.

>>3008719
ok i forgot i had noted that i wanted to thank people more for constructive feedback so guy who suggested vla video thank you. I do appreciate it and I might eventually watch it. I think I get a bit over defensive on here due to a bit of excessive hostility and abuse I face here but like VLA at least sounds interesting and I might be able to pick some peanuts out of the turd AI to eat.

>>3008718
>I followed all directions and it didn't work.
Okay, let's have a little thought experiment here; you followed all directions in soldering your ESC. The final result was you getting it bricked. Now, was that your intention? OR did you do something wrong along the way? If you did, then does this mean you're always capable of following the directions correctly? Like, are you capable of recognizing or at least accepting possibility of your own fault?

By the way, how does that change the fact that you just made a big R&D heavy deal out of simple etching? I can make a fucking precise etch with sodium persulfate no problem

>>3008720
you would face a lot less aggressivity if you actually listened to people I had to tell you 3 times about it and call you a nigger before you admitted it was actually interesting. Anyway you are welcome.
>>3008719
>Just copying is not enough. copying while also understanding the step and being then able to improvise or improve on the step ITSELF due to understanding the step in a deeper way is key
You should really watch this video then :
https://youtu.be/cqwKceUSZ5Q?si=h4C5GmmLrPQEK6VZ
What you are trying to accomplish is actually being done rn, it's actual research. I'm not saying it's perfect I'm saying it's getting better and better. VLA and world models are exactly trying to do what you describe, building a deep understanding of the world and its causal relationships to actually predict it and generalise using rules it already understands.

>>3008721
nope i bet you can't. use my design and make a board lets see. bet you get undercutting and bad traces

>>3008723
and the fact this new video isn't how they did it from the outset proves their AI is turd AI and they are finally trying to address that

>>3008733
So um when are you gonna release your super cool AI(TM) that is totally going to beat 20 years of research founded by billions of dollars and finally prove to the world the heuristic based computing (you know the AI branch that died when neural networks got bigger) is actually the way to go ? I'm legit curious on how you'll react when you'll inevitably get nothing out of it, I legit hoape you don't harm yourself I don't think you are a bad person you just need some help

>>3005837
Dude, don't you think if your idea could be done, the mega corporations would have already done it? How the fuck do you think your dumb ass can make a self-replicating arm when nobody else on the planet can do it? You have ZERO artistic ability, your circuit traces show that. You're sloppy. You're short-sighted. You're BULLSHITTING YOURSELF, of all people. Just admit defeat for once, stop wasting your life, and do something productive for yourself and your family. It's OKAY to fail at this. Failure builds strength. Making these threads just pisses people off, because they see how fucking retarded you are and nothing short of you getting a lobotomy will stop this god-forsaken nonsense. Just STOP. You're learning NOTHING. You are INCAPABLE of learning ANYTHING of value, and you're wasting precious time of everyone involved.
At first I thought this was an elaborate trolling attempt. My mind wouldn't accept anything else for an answer. But after seeing your insanity over and over again, I understand that you are deeply mentally handicapped. For the love of GOD, just fucking stop, man. At least stay off the internet, you're only bringing negative value to everyone who sees this. Your posting and behavior makes normal people want to enact violence upon you.