We at Purism want to revolutionize robotics and are seeking your support to fund this research and development. We believe robotics, AI, and technology as a whole need a new direction from what Big Tech is building. We consistently deliver on revolutionary technology, and with your support will invest to revolutionize humanoid robotics. We are Purism, a Social Purpose Company, who consistently challenges the norms of Big Tech by providing innovative products and services that respect your digital rights.
Interesting! I’m hoping people will be interested in this campaign, though I don’t see a personal use-case. This may be a campaign aimed more for niche use-cases like home assistance (people with disabilities) or for light industrial use, though I’m not very imaginative here.
The current job postings do not have any description other than to email Todd. Maybe due to the secrecy of the campaign until now, the descriptions were hidden and will be filled out soon.
2 Likes
In my opinion the idea is interesting and it is probably good to cross Big Tech’s plans early to avoid that every future robot is under their control.
As I can’t take part in development I may come in when I see that the robot can do some basic every task well, e.g. pick an object from one place and bring it to another safely.
3 Likes
OMG, this is … scary 
… and exciting 
at the same time !
The remote control from the L5 would be wild !
I have a friend who worked many years at Aldebaran/Softbank robotics on NAO, and he talked to me a lot about the issues they had to deal with (grip of the hands, bipedial stability with every gestures, hardware failures - better make a good and adapted choice for the motors, taking vocal command while robot is “speaking”, automatic stand up when falling, some ethical issues with some behaviors, etc…)
I don’t want to be a downer, but it seems to me 1.200.000$ is a very very low budget for a full body robot (head only, or head on wheels, seems more realistic)
But hey… you made it to the end with the L5 with also a low budget, soo why not …
I have multiple questions :
- Do you have someone who already worked on this kind of product hardware wise ?
- What kind of AI are you planning to use ? online or/and offline ? your own ?
- Translation services online or offline ?
- Have you planned personnalized vocal commands ? (“hey mimi do X”, then my_script.sh is executed)
- Big up for the clear change of “No refunds” policy at it should have been for the L5
, but what if the campaign doesn’t reach the goal ? No refund ?
I think it should be “No refund” after the goal is reached
what is the rational to keep the money if the project is not even officially started ?
Some remarks :
- “Subscription by default” + “An operator (OP) and owner must register for an account to operate a MiMi” for remote control is a BIG ICK
But at the same time “There will be an option to self-host but would be the exception to the default”, I’m not sure I fully understand what’s going on here - “Cellular Subscription by Default” for ??? (probably alert on access) is a BIG ICK
- “The owner will register the MiMi to verify an active subscription and account.” is a BIG ICK
- Or maybe I missed something, are those accounts on a 3rd party server, or inside the MiMi ?
2 Likes
I certainly have quite a few concerns.
Engineering a robot, let alone one that uses libre software is an incredibly demanding task. Consider how long it took Boston Dynamics to get their robot to walk. Can it gently grab things? Can it even reach things? Else, I am picturing a toy that waddles at 0.01 km/h.
Vision seems to be lacking, I still don’t have a clear idea of what Purism is building after reading all their material. Are they trying to crowdsource AI training data for completion of particular tasks. Can the robot operate well on its own? What exactly is being crowdfunded? Is it a response to Amazon Astro (which I would note is on wheels instead of bipedal)? If so, it doesn’t seem to come close to matching it’s feature-set. Also, I don’t understand if it is on wheels, why it doesn’t also just vacuum my house while it’s at it.
Focus on existing and parallel market segments I feel is also lacking. There is a lot to be improved on the hardware, firmware, and software. Librem 5 took 6 years to fully get out the door, and 4G radios are expected to be taken offline in the next 6-10 years. The Librem 11 doesn’t seem like a proper Purism device without HKS, and with that weird mask on the back, doesn’t have an easily replaceable battery. Crimson is missing support on many devices, and Trixie will already be available sometime next year, leaving devices 2 major releases behind.
8 Likes
I don’t understand how $1.2m would be even close to being the budget for R&D for something like this. $120m probably isn’t enough, if we’re honest. And that would be for regular robotics, making the software free and open source will only add to the cost. I understand the plan is to do multiple funding rounds, but none of this makes sense. Amazon, Google, Boston Dynamics, and many others have struggled in this space. Putting the fact that Purism has zero expertise in this area aside, I just don’t see how $1.2m even scratches the surface on preliminary research and design. (If we’re all honest with ourselves, Canonical’s decision to not pursue building an Ubuntu Phone a decade ago after it couldn’t raise $32m was the right one; the $12m wasn’t enough and rather than pretend it was, Canoncial shut down the project.)
The decision to pre-sell via coupons rather than traditional preorders is slightly better than selling a promise for a product that doesn’t exist and probably won’t exist, but the cynic in me sees this as a way for Purism the company to try to get some free cash flow to continue existing operations, rather than a viable avenue that the company is investing in for the future. In fact, the more I look at this whole funding scheme, the more concerns I have.
I personally do not expect this robot to ever ship in any form (and certainly not in any form that matches the computer generated mockup in the photos and video) and the best case scenario I can see for anyone who backs this would be for Purism to tell them in a year (because coupons can’t even be redeemed until 12 months have passed) that “hey, we killed the robot project, but you can use that coupon towards any of the antiquated items in our shop.”
If this was a real area where Purism really thought they could build and ship a consumer (or even developer centric) robot, the correct way to fund the R&D would be via venture capital, research grants or from traditional lenders. And even putting aside how unlikely it would be for traditional lenders or VCs to invest in Purism given the past lawsuits over private investment and the payday loan sharks, if this is an area you really think you can make a move in, you just start a new company and fundraise under it. Having this tied to the Purism name makes me incredibly, incredibly skeptical.
People can obviously do whatever they want to do with their money. But I do think it is worth thinking about if the people who have been unable to sustain a business around a phone project somehow now have the know how, expertise, and budget to build in a much more expensive and much more complicated space.
6 Likes
Open Source Robots are nothing new and already exist for a decade. I guess Purism can make use of many things that are already published.
I ask myself more what’s the real use case of this beside being a toy and if this is not another project that eats money (but at least the L5 has a real and important use case). At the end I am not much interested in, but we will see how many people want to spend money.
4 Likes
I understand why they are making the connected service a paid subscription. Maintaining servers and the associated software costs money and I do think that moving away from a paid product with “free” service is an overall good move.
In section M, “The local owner will have the option to self-host or operate the MiMi without a paid subscription.” tells me that there will be clear instructions for self-hosting (if licensed under either the GPL or AGPL section 6 ““Installation Information” for a User Product means any methods, procedures, authorization keys, or other information required to install and execute modified versions of a covered work in that User Product from a modified version of its Corresponding Source.”; I hope that this server software is licensed under the AGPL instead of the GPL). With the given information, I’m currently relieved that MiMi can work offline for those that don’t with to pay for a subscription or wish to use it without remote connectivity.
I agree, I don’t understand why a cellular subscription has to be included. At least it has the option for WiFi.
If someone does not wish to use the subscription service, this shouldn’t be required to do according to Purism’s prior statement “The local owner will have the option to self-host or operate the MiMi without a paid subscription.”.
If there is an active subscription, there will need to be an account registered and I assume this would be on Purism’s servers. I don’t know if they are using a hosting provider like DigitalOcean or AWS but it seems likely they don’t manage their own server racks.
I agree. Purism’s relatively “small” budget for the Librem 5 might have factored in ongoing laptop sales, which included the Librem 13 and 15 at the time. Robotics, let alone “personal” robotics, is still a developing field and I can only assume that Purism might be relying on some sort of reference design to reduce cost while they can focus on its software. Even then, it seems like Purism may need to create the software mostly from the beginning as I don’t see how PureOS can be suitable for the robot itself.
Purism does appear to be taking a very conservative approach here to not promise a product for providing money to them. This looks like a workaround to the refund issues Purism has with the Librem 5 where people were promised a Librem 5 device. At least here, the terms are clear and are not too dissimilar to other crowdfunding platforms like Indiegogo or Kickstarter because they’re not “stores” and you’re not buying products but rather ideas and concepts that are hopefully evolved into a thing the backers can use.
Looking at their projected timeline, they estimate it can take up to 18 months for each iteration, which would be about 6 years, a timeframe I still find relatively short given the lack of experience Purism has developing a product such as this. There are a number of usability and durability concerns I have such as a need for this to be relatively durable to blunt force and avoiding liquid damage if one of the use-cases will be within a home.
2 Likes
Wow interesting announcement. It reminds me of Buck Rogers. It will be ahead of its time if launched before the 25th Century. Twiki-Speak == MiMi on youtube The promo video with Todd and David is cool.
Is this project running in parallel to releasing Crimson for Librem 5? Are past ambitions, Fir phone still planned? Interledger and paid apps? Monero?
2 Likes
Well best luck in all of this
1 Like
I don’t think it does. Disclaimer: We are speculating about a product that is some years away and exactly what it looks like in terms of all attributes at delivery can’t be considered known at this stage.
The full text:
The MiMi will include by default a Cellular subscription bundled with the OP App subscription, even though the MiMi can connect via WiFi for OP App use. The local owner will also have the option to do WiFi only or utilize their own prepaid SIM to operate the MiMi outside the default cellular subscription.
So it seems clear:
- do WiFi only, OR
- do WiFi + cellular but with your own SIM (although I don’t understand why it has to be prepaid), OR
- do WiFi + cellular with (presumably) AweSIM
I guess there are use cases for remote access, hence it makes sense to have cellular support. You wouldn’t want the robot freezing because it went outside your WiFi range. This is the flip side of being operator-controlled rather than being fully autonomous. The robot needs fairly continuous communication with the operator. The AI “assistance” can only provide so much functionality.
Since AweSIM, at this time, is only offered in the US, it seems likely that the first two options must exist if Purism intends to sell this outside the US.
2 Likes
This is… something. A mini robot is many things: a toy, a curiosity, a hobby project, a tech development project, a distraction. I can’t see it as useful or beneficial - especially at scale. But maybe that’s not needed and niche is ok. This could lead to know-how that leads to more useful (and commercial) things like safe and privacy respecting robot vacuums (those have a lot of sensors and data gets sold) or open and controllable AI in linux tech etc.
Humanoid form is cute but it’s a human ego flaw to assume that’s an efficient form - especially in the short and mid-term. I hope the cuteness factor appeals to enough people to make this financially sound. Big tech can afford these vanity projects. I would have done the robot-vacuum (or hybrid with other tasks) step first before advancing to this but I guess it’s a PR thing for the company as well.
Pic from Toyoda museum: a robot band. The most useful of the bunch is not a (semi-)humanoid musician. [edit: The pic is just for illustration - I happened to visit there some time ago. The top of the line violinist robot was precise enough to play a regular violin but a walkman could outperform all these robots.]
6 Likes
True, the app subscription and cellular plan could be a potential bundle like with the Librem 5 + AweSIM.
This is an interesting perspective as it would seem logical with a humanoid robot being a seemingly niche product. The underlying technology research and experience might benefit many other tech stacks where the parts might make up more than the sum of them.
The concept of Purism’s promotional video make me think of the telepresence robot that never seemed to take off. I suspect this would be more flexible in its practicality though. One of the proposed use-cases of offering a more personal connection with people on both ends might be more difficult to achieve; just seeing reports mentioning how the Apple Vision Pro seems stagnant right now and sales are perhaps not what Apple would like, despite the impressive engineering, makes me think telecommunication could be a challenging sell outside of B2B.
2 Likes
Or drone?
That’s an interesting discussion point but I guess there’s a trade-off between efficiency and ability to sell. If people want and expect humanoid form (too many sci-fi movies?) then there is some merit in meeting the market demand. Worth also linking to: Uncanny valley - Wikipedia
Actually I would like an open source power loader (a la Alien film franchise). 
That would be my hope.
4 Likes
I do think Robotics + AI are poised to take off. I’m not so keen on the anthropomorphism angle, but generative AI combined with physical manipulation abilities will be a huge change. Exponential growth of progress is always surprising.
However, I don’t see Purism being able to keep up. I also don’t see what they bring to the table.
4 Likes
“Rapid iterative milestone” development sounds like a future train wreck of reduced capabilities and missed deadlines. What’s the rush? Can we do a thoughtful separation of concerns, produce verifiable class models, then translate them into working code instead?
2 Likes
The MiMi will be a home computer assistant. It will be able to read the news, read your messages, give you a recipe, search for useful information from voice command, read an e-book, play your favorite music or have so many other useful applications.
It will also be able to surveil your house while you’re away, let you connect to it to move around and see through it’s eyes, alert from an intrusion, alert from a fire and more… all in a privacy respecting way.
The MiMi will also revolutionize remote social interactions. Your friends, family, partner will be able to spend time with you, physically interact with you or play with your kids, your pets all remotely. With a VR headset, one would be able to see what the MiMi sees in 3D. Imagine the immersion!
And this is only the tip of the iceberg. The MiMi is such a creative machine, fully open, with so many sensors and capabilities, there is so much one can imagine doing with it.
7 Likes
Regarding the crowdfunding and financing of the development, as the topic came across this thread, I wish to comment about it.
Development willl be incremental and the first target is not about having a complete robot, but about having a dev-kit with the board, sensors and software, which will already be a great base for moving forward with the MiMi and for anyone wanting to do their own robotic project.
From the experience we have in making a phone from scratch, we believe that the first crowdfunding target will be enough to finance this first development goal.
We don’t know the final cost of the product yet and it is why we are using coupons during the crowdfunding campaign.
3 Likes
As a technology professional, this would be a fun project to be involved in (as long as “social purpose” extends to management style); as a consumer, I would much prefer such a device from a social purpose corporation…
…but as a Librem 5 customer, I keep hearing a huge sucking sound. Hopefully, the software work is more additive than subtractive.
6 Likes
I encourage everyone to read the fine print in regard to what you are buying. You are not buying a MiMi or a MiMi related product. You are buying a non-refundable coupon that can be used to buy a MiMi related product if one were to ever exist. While there is a value associated to the coupon it is only for a MiMi related product and there is no promise or representation about what the MiMi (or related product) will cost. [e.g. Whatever MiMi related product you might want might end up priced at $4,000 and you have paid $2,000 into your non-refundable coupon, you’ll still need to pay an additional nearly $2,000 (you will get 10% off).]
i.e. Purism holds no risk at all and really isn’t committing to anything.
3 Likes