Re: [HBRobotics] tablet design

[Chris Albertson]

> On May 7, 2026, at 1:29 PM, A J <aj48...@gmail.com> wrote:
>
> Hey Folks,
>
> I was thinking about global robot design and the modern tablet came to mind.
>
> If we build tablets like Apple, Samsung & Google with many options for configuration
> it will be well received globally.
>
> Maybe the Bot is the same, we build the best mass produced product that we can.
>
> But we leave the OS open to the customers preferences.


The problem is that literally anyone with a decent budget can build a mechanical humanoid robot. University labs and, of late, literally any company that wants to can.

But not one person on Earth has built the “OS” that enables useful behavior that creates any economic benefit whatsoever. No one is even close or even on a path that will get them close. We will need another fundamental breakthrough, not just bigger LLMs.

So “leave the OS to customers”? What customer? There is no even one customer who could do this.

[Dave Everett]

We are in agreement on this Chris.

There is no generally useful humanoid robot and this is what people would want. It must be practical out of the box.

Nothing has changed since the 1980s, companies are still producing robots that might appear exciting, but could they do without extensive user programming? 

The new Robotis humanoid is cheaper than any, but it still has solid hands that if you have seen it could be useful for <ahem> one task, but I'm not going to trust a robot with that.

Dave

--
You received this message because you are subscribed to the Google Groups "HomeBrew Robotics Club" group.
To unsubscribe from this group and stop receiving emails from it, send an email to hbrobotics+...@googlegroups.com.
To view this discussion visit https://groups.google.com/d/msgid/hbrobotics/F560EBA6-05A0-4302-AD5A-B5EA15CB2ADB%40gmail.com.

[Thomas Messerschmidt]

If you have funding, I’ll build that robot with you.


Thomas Messerschmidt

-

Need something prototyped, built or coded? I’ve been building prototypes for companies for 15 years. I am now incorporating generative AI into products.

Contact me directly or through LinkedIn:

https://www.linkedin.com/in/ai-robotics/



On May 7, 2026, at 9:10 PM, Chris Albertson <alberts...@gmail.com> wrote:



[Dave Everett]

On Fri, 8 May 2026 at 15:30, Thomas Messerschmidt <thomas...@gmail.com> wrote:

If you have funding, I’ll build that robot with you.

I can't imagine funding being less than $200,000,000 and a lead time of  10 years, so $20,000,000 a year with no product until a decade later.

Dave

To view this discussion visit https://groups.google.com/d/msgid/hbrobotics/D1C5AEDA-6A18-4CD6-9F9B-FB78EA8D422D%40gmail.com.

[Chris Albertson]

Real question: How do you handle footfall placement? I notice that we humans have two modes. When we know the floor is flat and level we do not think we seem to let the feet hit where they will based on the stride length and to maintain balance. But with uneven ground, we look at each place the foot needs to go. I don’t think this is a binary thing. There are degrees smh we move smoothly up and down this scale of attentiveness. I think the answer is “MPC” (model predictive control theory. ) But I am always baffled by transitions.

My goal is a single algorithm that was a parameter like “attion to footfall location andcan move from 0 ro 1.

I bring this up to show that there are many things to solve before it makes sense to own an actual robot. Buying one now will not solve this problem. All I really need is a pencil and paper and to be a bit smarter. And then of course, I don’t need a humanoid to test my ideas, a MUCH lower cost quadruped has the same problem

Moving to real hardware should be the last step, not the first. Otherwise, the hardware sits in the corner while you wish you were smarter.

> To view this discussion visit https://groups.google.com/d/msgid/hbrobotics/D1C5AEDA-6A18-4CD6-9F9B-FB78EA8D422D%40gmail.com.

[Dave Everett]

On Fri, 8 May 2026 at 16:16, Chris Albertson <alberts...@gmail.com> wrote:

Real question:  How do you handle footfall placement?      I notice that we humans have two modes. When we know the floor is flat and level we do not think we seem to let the feet hit where they will based on the stride length and to maintain balance.    

I can't recall seeing a humanoid robot do anything other than flat hard floors. Perhaps some have, but I have not seen them. Even a standard footpath is likely to cause issues with the divisions between the segments or tree roots that have lifted and broken a segment. Grass or weeds growing through, steaming dog waste, delivery robots blocking the path and requiring the humanoid to walk on the grass or gravel.

Now I recall one humanoid doing a relatively flat, slight downhill grassy slope before pointlessly doing a back flip. It's also possible I am not aware of a lucrative back flip market.

Dave

[Chris Albertson]

On May 7, 2026, at 11:09 PM, Dave Everett <daveev...@gmail.com> wrote:

On Fri, 8 May 2026 at 15:30, Thomas Messerschmidt <thomas...@gmail.com> wrote:

If you have funding, I’ll build that robot with you.

I can't imagine funding being less than $200,000,000 and a lead time of  10 years, so $20,000,000 a year with no product until a decade later.

A physical humanoid robot isn’t necessary to solve the problem. Subsystem design and simulation come first.

The robot will remain a novelty toy until we have a theory of self-actualized purposeful behavior. Until then, they’re no smarter than a 5-axis CNC milling machine. They run programs without knowing what they do..

You say “ten years.” I ask, “How long until we find a verifiable theory of quantum gravity?” It’s impossible to answer. Timing depends on pure luck. Someone must have a brilliant, revolutionary idea. We can’t predict that. It could be 60 years or tomorrow. We can’t know.

So it is even worse, you ask for $200 million and then say you have no way to know if or when you might have a product.    At this point it is basic science research , not engineering.

[Robert Hasslen]

I'm new to this group, but looking forward to becoming more involved, even though I moved out of the immediate area.

This humanoid robotics craziness feels like what happened with General Magic.  Brilliant and visionary,  but lots of lost investment.  Hard to fathom 100's of millions of dollars being invested in this.

Bob

--
You received this message because you are subscribed to the Google Groups "HomeBrew Robotics Club" group.
To unsubscribe from this group and stop receiving emails from it, send an email to hbrobotics+...@googlegroups.com.

To view this discussion visit https://groups.google.com/d/msgid/hbrobotics/9DB1B0E1-7C06-4B4A-BBF5-ED14D419962B%40gmail.com.

[Chris Albertson]

> On May 8, 2026, at 10:48 AM, Robert Hasslen <rhas...@gmail.com> wrote:
>
> I'm new to this group, but looking forward to becoming more involved, even though I moved out of the immediate area.
>
> This humanoid robotics craziness feels like what happened with General Magic. Brilliant and visionary, but lots of lost investment. Hard to fathom 100's of millions of dollars being invested in this.

It is premature. Right now, we can build nice machines, but the current AI technology that is based on LLMs is just not up to the task of controlling the robot. So we have nice machines that can only do demo movements and remote control. Kind of semi-automated puppets really.


Intellectually, this should be very boring. What makes them interesting is that we humans are genetically programmed to think that a human-looking object works internally like a human. We, and all social animals, use this to predict and influence the actions of others. We think “they are like me”. Building the robot in the shape of a human triggers this. So the average person thinks the humanoid robot is vastly more capable than it is, even if all they observed is a machine that can only stand and walk.


Some research has shown that we can trigger this human trait with something as simple as drawing a happy face on a paper plate and gluing it to a stick. This is also why cartoons “work”. Even simple line drawings are enough to evoke the “it looks human and therefore it must think like me.” Six million years of evolution literally burned this into our DNA.


So we can’t help but like human-shape robots.

[Dave Everett]

On Sat, 9 May 2026 at 04:30, Robert Hasslen <rhas...@gmail.com> wrote:

I'm new to this group, but looking forward to becoming more involved, even though I moved out of the immediate area.

Don't be worried about that Robert, I'm 7500 miles away. 

Dave