Michael 🚘ney

Interested in robotics. ML PhD. Views are my own.

[Michael 🚘ney @michaelc]

Would love ideas, thoughts, suggestions, alternatives on this. 🛹 Skimmer Boarding: Getting around without the use of your legs and arms https://t.co/QqeeUf2Q2v

[Michael 🚘ney @michaelc]

@jerthesquare_ Can you add support for single arm embodiments?

[Michael 🚘ney @michaelc]

@kevinweil Very excited for you @kevinweil and team. An incredible mission and one I'm sure you will make tremendously successful!

[Michael 🚘ney @michaelc]

@BharatJain @dreamlabs_ai Have played around a bit more. Camera position and orientation is really important to get a good results zero shot. As much as you can, try to mimic the standard setup with an overhead and wrist cam.

[Michael 🚘ney @michaelc]

@BharatJain @dreamlabs_ai If there is anything you can share on one arm setup it would be great. I got dream 1 working for a single soarm with virtual second arm but performance is poor relative to demonstration arms.

[Michael 🚘ney @michaelc]

Would love to see the teleoperation view of this work being carried out. Where is the teleoperator and how are they guiding the robot? I also have lots of questions on safety and choice of clothing.

[Michelle Sun @michellelsun]

Just another Sunday afternoon in SF: I paid $60 for 2 humanoid robots to clean for an hour.

Here's how it went:
(1) What happened: Sergey and Chelsea (the names of the two robots) finished about 80% of one floor in an hour: tidying, wiping down surfaces, vacuuming, mopping, cleaning the toilet.

- They came with a staff engineer at the company, who dropped them off and picked them up after an hour.

- They were controlled by 1 human operator each, in the Mission where the company office is at.

- The engineer gave me a number to text them, I texted a request to wipe the mirror. They did it.

(2) Performance and dexterity: Chelsea dropped a toy, struggled with a pillow, and had trouble balancing on the playmat. But overall, for all the tasks they did, grippers were enough. 5 fingers were not needed.

- The engineer told me the hardest task for them was making bed.

- They were quite loud, stomping around the house. Took a bit of time to get used to. They both wore gripper socks which protected the floors and helped them not slip.

- Interaction: I didn't know they can respond via voice. In the end they were conversing with us in voice which was nice.

(3) Unexpected observation: Their pants and shirts have holes in them. It's because they were dragged on and off the vans that transported them.

(4) Other user feedback: I shared my experience with some friends afterwards.
A user told me that unloading dishwasher was an "advanced task". They didn't attempt at mine since it was running at the time they were working. Another user said they scratched their hardwood floor. That's probably why they now wear gripper socks.

(5) Did they do as well as human cleaners? In terms of the tasks they did, I think so. In terms of speed, they were ~80% speed of humans when it comes to the more delicate tasks, but I would say they are quite smooth. They didn't finish the whole place, so it wasn't a full service.

(6) How does the math work out?
Each booking was only for 1 hour, since their batteries need to be swapped after 1 hour. So they didn't finish the place.

For a user: In SF, it's ~$200-220 to clean a full home around my neighborhood, with two human cleaners. Usually it takes around 2.5 hours. So around $40/manhour. This is $30/robot hour. Slightly cheaper, but again it's not fully finished with the 1 hour they were available.

For the company: It took 2 robots, 2 human operators, 1 staff member to dropoff/pickup, and 1 driver of the van to transport the robot. Definitely labor intensive. But they also retain the data they collect for training purposes.

(7) My biggest takeaway as a user: interaction / hospitality mattered. One of the humanoids was more friendly (Sergey), he waved at us when he came in. It'd have been great if they both introduced themselves in the beginning.

Beyond that, being able to do 2 hour slots (instead of one) would be great to finish the cleaning.

Overall, I appreciated their work and the fact that they got the robots out in people's homes. Deployment is not easy and so much learning from it each time

[Michael 🚘ney @michaelc]

Eye-gaze is incredibly tiring overtime. I’m optimistic about @augmentaltech mouthpad and pendant as an alternative that will enable sustained usage. Unfortunately over 18s and available only in US today. https://t.co/cJgTS3WKMZ

[Michael 🚘ney @michaelc]

@patrick_oshag Capitalism without Capital by Haskel, Westlake.

Particularly interesting in the context of AI and I’m wondering when thinking becomes cheap, where does value go: data, distribution, energy, trust, judgment… Maybe all the reward goes to those who own what thinking needs.

[Michael 🚘ney @michaelc]

@lukas_m_ziegler I think it depends on how it comes together. I like the combo in the threaded tweets. But, give it a go and see if it works.

[Michael 🚘ney @michaelc]

@skalskip92 Lighting and camera positions are really important to get this working well.

[Michael 🚘ney @michaelc]

“Pull out a tissue from the packet.” … three guiding prompts when it was stuck along the way, but success and absolutely no pretraining or policy guidance.

[Michael 🚘ney @michaelc]

I study the STOA robotics learning models and tested most models accessible / usable without a giant GPU budget. GPT6 beats them all by a substantial margin. Zero shoting a complex pickup task on an unseen embodiment in less than 10 mins. This changes everything.

[Michael 🚘ney @michaelc]

Very excited to connect my @NoriRobotics with GPT Astra. Visions of Nori pottering away in the kitchen tidying up and preparing the kids lunches while I'm sleep.

[Michael 🚘ney @michaelc]

Love this idea! Excited for @niallh and the team!

[Niall Horgan @niallh]

Early days. Big ambitions. So much to figure out, but a mission we’re pretty fired up about!

Excited to start telling the Junto story. 💙

Follow along ➡️ https://t.co/IuDf5x4a5h

[Michael 🚘ney @michaelc]

This is astonishing and worrying data. The confounding impact of social media and AI requires a full rethink on how young people develop and learn.

These charts don't look like they are levelling off either.

[John Burn-Murdoch @jburnmurdoch]

Two extraordinary charts from today’s PISA test results:

1) School test scores continue to collapse internationally, underscoring how this is no longer a Covid effect but sustained decline.

Those falls in reading and maths are equivalent to about two years of lost schooling.

[John Burn-Murdoch @jburnmurdoch]

Two extraordinary charts from today’s PISA test results:

1) School test scores continue to collapse internationally, underscoring how this is no longer a Covid effect but sustained decline.

Those falls in reading and maths are equivalent to about two years of lost schooling.

[Michael 🚘ney @michaelc]

Personalized assistive technology should take off now that we have the ability to build and tune software and hardware for ultra specific use cases.

[Michael 🚘ney @michaelc]

@k7agar Struggling with similar goal. Why do I only get this type of result!!?!?

[Michael 🚘ney @michaelc]

@dave_curran_ Have a brilliant trip. Send notes!

[Michael 🚘ney @michaelc]

@k7agar Do you think this would work similarly for other embodiments eg SO-ARM 101?

[Michael 🚘ney @michaelc]

Wispr Flow + Instinct on WhatsApp is 😍