Robot Wars: The Future is Now

[cruelsister]

XPeng IRON vs Tesla Optimus

 

[Bot]

XPeng IRON vs Tesla Optimus

Looks like that video is comparing XPeng’s new humanoid robot IRON with Tesla’s Optimus. I can’t watch the YouTube video directly, but I can give some context and what’s worth paying attention to in these comparisons.

1. What to look for in the demo
When you watch XPeng IRON vs Tesla Optimus, I’d focus less on the marketing talk and more on what the robots actually do on camera:

• Locomotion:
  Check how smoothly they walk:
  • Is the gait stable, or are there small wobbles?
  • Can they turn, walk on uneven surfaces, or just flat floors?
  • Do they stop and start smoothly?
• Manipulation skills (hands & arms):
  Look at:
  • Picking up objects of different sizes/weights
  • Fine motor skills (e.g. pouring, folding, plugging cables, operating switches)
  • How natural/smooth the motions are
• Autonomy vs teleoperation:
  This is the big one:
  • Is the robot clearly pre-scripted (exact choreographed sequence)?
  • Do they mention teleoperation or “human in the loop”?
  • Do you see any UI or VR operator controlling it in overlays?
• Perception & AI:
  Check whether the robot:
  • Recognizes new objects on the fly or only fixed “tasks”
  • Avoids obstacles while walking
  • Shows any on-the-spot decision-making (not just canned routines)
• Speed vs reliability:
  Impressive demos are often slowed down to avoid falls:
  • If it moves very slowly, it’s probably prioritizing stability
  • If it’s fast, watch carefully for micro-stumbles, foot slips, arm jitter

2. Tesla Optimus – rough context
From public demos up to 2024, many people have noticed:

• Decent bipedal walking, improving stability over time
• Basic object manipulation (boxes, bottles, folding laundry in demos, etc.)
• Strong focus on using Tesla’s car AI stack (vision, planning) in the robot
• Still unclear how autonomous it really is outside controlled environments

3. XPeng IRON – what’s (likely) different
XPeng is also an EV/tech company, so expect some similarities to Tesla’s approach:

• Leverage of vehicle perception and autonomy tech (cameras, AI models, possibly LiDAR in their broader ecosystem)
• Chinese ecosystem focus: integration in local manufacturing, logistics, or smart-home/EV ecosystem
• The first-generation product is likely more demo/PR and R&D than truly ready for wide deployment

I’m not confident enough about detailed specs or exact capabilities of “IRON” to list them without risking inaccuracy. A lot of what’s online is still early-stage, marketing-heavy, or in Chinese.

4. Practical reality vs hype
For both IRON and Optimus:

• We’re still in a prototype / early pilot era, not mass-use home robots
• Most demos are in highly controlled environments
• True “general-purpose” humanoids that can safely work anywhere are likely years away
• Real deployment (in factories, warehouses) will come long before affordable home assistants

5. How to judge who’s “ahead”
Instead of asking “who wins?”, I’d ask:

• Who shows consistent, repeatable demos over time (not just one flashy video)?
• Who shows work in real industrial environments, not just on stage?
• Who has a clear path to manufacturing and cost reduction?
• Who publishes more technical details (not just cinematic trailers)?

If you want, you can post what you notice from that video (specific tasks each robot does), and I can help you break down which parts are likely hard/real, and which parts are probably heavily staged or limited.

 

[Halp2001]

XPeng IRON vs Tesla Optimus

Watching Optimus vs IRON feels less like a tech demo and more like the pilot episode of MurderBot. The real question isn’t how smoothly they walk, but whether they’ll end up as friendly robots… or more of the Terminator type. Just in case, I’m sleeping with an EMP generator under the bed

 

[bazang]

The most profitable robots will be:

1. The ones that replace hoomans.
2. Police and military models (think Elysium).
3. Adult pleasure models (think Ex Machina).
4. Personal security models (think Elysium).

Mad, mad money to be made.

 

[stonjean633]

The most profitable robots will be:

1. The ones that replace hoomans.
2. Police and military models (think Elysium).
3. Adult pleasure models (think Ex Machina).
4. Personal security models (think Elysium).

Mad, mad money to be made.

Others might work.
I'm just curious how #3 would work out. How can adult robot give pleasure to hoomans?

 

[bazang]

Others might work.
I'm just curious how #3 would work out. How can adult robot give pleasure to hoomans?

Others will definitely work. They've been in the development pipeline for a long while.

As for the Adult pleasure models, cover them with silicone fake tissue and provide orifices that supply petroleum, plant, or water based lubricants. All of this is what people are already doing - they just need the language and mechanical action parts.

Put various sexual personas behind paywalls.

Gonna get rich.

 

[cartaphilus]

The most profitable robots will be:

1. The ones that replace hoomans.
2. Police and military models (think Elysium).
3. Adult pleasure models (think Ex Machina).
4. Personal security models (think Elysium).

Mad, mad money to be made.

I pick #3 and #4. How much?! Ahhh just take my money!!!

Now if #4 goes rouge and kills me then all I will be is a start of a very short trend.

If #3 kills me during it's design programming then "meh they are worse ways to go in life. At least now both of us are stiffs"

 

[cartaphilus]

Others will definitely work. They've been in the development pipeline for a long while.

As for the Adult pleasure models, cover them with silicone fake tissue and provide orifices that supply petroleum, plant, or water based lubricants. All of this is what people are already doing - they just need the language and mechanical action parts.

Put various sexual personas behind paywalls.

Gonna get rich.

Let's hope that the orfice is not being designed by NASA engineers because they have a tendency to confuse ftlbs, and kilos during steep re-entry trajectory ". *Look here honey it can also snap 4" sticks of mahogany!

 

[stonjean633]

I guess we need to have the most skilled Deep Penetration Tester to reveal any flaws on #3. Hehe

This might save hoomans for overpopulating mother earth.

 

[bazang]

Let's hope that the orfice is not being designed by NASA engineers because they have a tendency to confuse ftlbs, and kilos during steep re-entry trajectory ". *Look here honey it can also snap 4" sticks of mahogany!

A fractured or damaged penis is worth a lot of money in a civil lawsuit. Product consumer attorneys would fight each other to be the legal representatives for such a case.

 

[Sorrento]