Before launching Actuator, someone suggested that I create a weekly robotics newsletter. I was hesitant, because I was unsure if we could come up with enough content to keep it running. My concern wasn’t if there was enough robotics content in general, but if there was enough related to the field.
At TechCrunch, we are very selective about what content is featured on our website. We have a limited staff and a tight schedule, so not everything that comes across our desk can be published. I am personally responsible for overseeing our robotics section as well as all consumer hardware-related news and reviews.
An essential element of the job is curation. This includes careful examination, investigation, and deciding which stories would be most interesting to our readers. That is why much of our coverage about robotics focuses on startups and venture capital. We use this as a tool to understand technology in general.
If I told you that every week at Actuator HQ (a one-bedroom office managed by a mischievous lionhead rabbit mix) was full of great news, I would be lying. The pandemic has had an effect on both robotics and our coverage of it, so some weeks are better than others. Overall, I’m satisfied with the amount and caliber of news that I encounter.
Last week was a bit sluggish. This week, not quite so much. In addition to the typical roundup of stories, I asked a handful of investors a difficult question with a straightforward answer: What effect will the SVB events have on robotics investing and startups? You can find a few of those answers below, followed by a long and in-depth conversation with robotics pioneer, Rodney Brooks.
I intend to attend ProMat.
I have not been to ProMat before, but over the last few months I have slowly realized that I should go at least once. What is usually advertised as a trade show for supply chain and logistics has become one of the biggest robotics events of the year without anyone really noticing.
It can be difficult to properly explain the technologies I cover in an email. To get the full experience, I often have to travel to places like Boston, Pittsburgh, or Chicago for a trade show. Then I can make time for the Dali exhibit at the Art Institute, as well as my usual stops at Quimby’s and Myopic in Wicker Park. No matter what, I’m thrilled to return to one of the greatest cities in the United States after a five year absence.
The event is very close, but I’m still in the process of setting things up. If you want to get some time with me, make sure to follow me on Twitter and/or LinkedIn. I’m planning to spend a couple of hours having a casual chat with investors and startups, probably in one of the city’s coffee shops. If you’re interested in joining, please vote here. It’ll probably be on Sunday afternoon and possibly some other time in the week if enough people show interest. Further details are forthcoming (I am still trying to organize my timetable).
I took a break for a couple of days, and it seemed like the tech sector stopped altogether. I’m not claiming that my decision to make up for working extra hours caused the Silicon Valley Bank run, but it’s worth noting that the timing is interesting. This is definitely the biggest tech story of the week, month and maybe even year, providing that Elon Musk doesn’t buy Facebook.
I don’t have as much of a connection to SVB as most of my coworkers at TechCrunch, yet its reverberations have been felt far and wide in the past week. I’m focusing on the ramifications that this may have on robotics companies over a longer period of time. There are a few robotics and AI corporations in the San Francisco Bay Area, but it’s nowhere near the level of robotics development that can be seen in cities like Pittsburgh and Boston. Many, if not most, of the companies investing in the companies discussed here have significant operations in the South Bay and San Francisco area. Much of the money that is discussed here has been processed through Silicon Valley Bank at some point.
In the last few months, I have questioned many new businesses about the effect of the economic downturn on their ability to bring in investment. The majority of them said it has been difficult. This is a massive shift from the companies in the tech industry that were doing extremely well during the pandemic. This is going to be an obstacle. It is essential to have faith and be careful, particularly if you are an entrepreneur who can see their resources diminishing.
I wanted to know how the SVB events might affect robotics investing and startup businesses, so I asked a number of investors if we will likely see more closures or mergers and acquisitions. Here is the feedback that I received:
I cannot understand why venture capitalists chose to launch a strike against the businesses in their portfolio. There is a chance that Silicon Valley Bank will survive, due to people such as Ro Khanna, and if we take steps to restrict the usage of digital media for malicious purposes, it may not be so expensive for robotics companies to raise funds. In contrast, if we are accustomed to robotic banking, it could become chaotic. How could opponents interfere with robotic advancements? We have seen the power of a few tweets and emails to undo a well-known business that had been going for 40 years. Rather than using cyber attacks, it would be easier to cause damage to thousands of innovative businesses with a few properly written messages from seemingly reliable sources.
Kelly Chen from DCVC stated that robotics startups usually have more intricate banking relationships, connecting their equipment financing and other types of debt to their banking. Despite the administration’s good action with SVB, robotics startups are still hopeful yet careful, wanting to spread out their options and ensure they have the best quality even if it means not getting the best rate.
Abe Murray from Alley Robotics is still very hopeful about the robotics industry. Despite expecting the venture capital climate to be more hesitant, robotics firms are not going to be disproportionately affected. Due to the tangible gains they offer their customers and the healthy profits they make, these companies remain attractive investments for financiers in any kind of market. They have more varied methods of funding than the majority of those businesses backed by venture capitalists (including capital expenditure leasing and project finance), and there are still numerous organizations and inventive financing associates ready to join with fantastic businesses that deliver value.
Murielle Thinard McLane from Intuitive Ventures believes that the capitalization strategies for robotics will change. Robotics has always been a business that requires a large amount of money for hardware and software. Many startups in the industry have been using SVB’s venture debt offering, so now that it is gone, other strategies for obtaining capital must be considered. For instance, if a company wants to create a digitally native product like a surgical robot, they will need to find a new way to get the necessary funding. In order to bring a medical device to market, there is a need for investment in a comprehensive system of services, assistance, instruction, and market growth. As a result, robotics businesses will have to concentrate on utilizing capital wisely and connecting to investors who comprehend the requirements. I would suggest that strategic investors on the company’s capitalization table could assist robotics startups in achieving their objectives due to their special connections to the bigger system. Following the termination of SVB, the already sluggish funding situation will be even more sluggish. Investors will probably pause to observe how the market settles, so the fundraising climate will be difficult. Robotics investments will be affected the same as other areas and mid- to late-stage robotics startups that raised funds during the investment surge will be hit the hardest. Since robotics companies rely heavily on equity dollars, investors will be very particular about what valuations they accept and the competition for funding will be fierce.
If you are familiar with Actuator, you probably know I am not a tech optimist. My caution does not come from a lack of faith in the creativity of people (we can be very clever when faced with a challenge), but rather from a lack of incentive to use technology in ways that are just and helpful to those who need it the most.
I have been giving thought to Rodney Brooks’ 2018 post, “My Dated Predictions” this week. The article attempts to predict the rate at which certain technologies will be developed and used. He has stated that he will update the post every year until 2050, as he expects to be too worn out by then to debate whether he was right or wrong.
It is amazing to note that a 95-year-old person can still be so passionate about their beliefs that they will engage in an intense discussion about them. Brooks pointed out in the start of the article:
Recently, the initial optimism that the Internet would provide a voice to everyone and stop governments from controlling individuals has been replaced by a sense of disappointment that this hasn’t come to pass. It’s been acknowledged that the bright future that was envisioned is taking much longer to become reality than previously thought. This is indicative of the fact that the initial optimism surrounding the deployment and usage of the Internet didn’t come to fruition.
This is the optimistic side of believing that technology will bring about a utopian world. We can be too quick to assume that new inventions will be used in the way they were intended, when in reality they are often driven by financial gain. This is especially true when it comes to how new technology is adopted. We find ourselves repeating the same patterns that were supposed to be broken by the introduction of the internet.
Brooks’ practical approach emphasizes the boundaries of technology. He labels himself a “realist” instead of a “pessimist,” and I can understand why. Recently, I wrote an article regarding the hype surrounding ChatGPT and other similar tools. As I pointed out then, I had a firsthand experience of the original 3D printing craze, which has shaped my opinion on the over-expectation seen in later phenomena like blockchain/web3. I wouldn’t suggest having your faith entirely destroyed, but if you take something away from the experience without still being naive, you have been paying attention.
I believe that coming up with ideas is simple, Brooks declares in the 2018 post. However, making them a reality is difficult and even more so, making them reach a large scale. I think it is feasible to categorize the prospects of success of technology and technology implementation into a range that varies from simple to very challenging.
I think it would be a good idea for serial entrepreneurs to get a tattoo related to this topic. On the fifth anniversary of the initial predictions post, Brooks said he now believes it will take even longer than he originally thought. Who would have known that his initial thoughts were actually quite optimistic? I had planned to make a joke about how ‘life comes at you fast’, but I guess the lesson here is that life actually progresses much more slowly than we anticipate.
A person queried me during a time when news was sparse to find out if I would be interested in talking to Brooks in regards to him receiving the IEEE Founders Medal at the IEEE VIC Summit and Honors Ceremony in May. An overview of his achievements was sent alongside the inquiry, however, I, and likely you, are already familiar with Brooks’ contributions to robotics over the last 40 years, particularly during his time at MIT. He was the main actor in an Errol Morris movie featuring clever and captivating individuals.
I was more than willing to have a conversation with Brooks. To my surprise, we had never had an actual discussion before. Lora Kolodny interviewed him at the first Robotics event in 2017, while Devin chaired a panel including Brooks and Clara Vu at the event last year. This week, we had a pleasant discussion which can be seen in the following link.
TC: What is your typical daily routine as a Chief Technology Officer?
RB: Apart from the events of the past few days, I usually attempt to get a reaction from people.
No changes in that area.
I refer to my work as “making provocations”. I’ll write some code or design something, although it’s not necessarily something that will go into a product. I like to challenge engineers and ask them, “What if we tried to do this?” They usually tell me it’s impossible, so I’ll do a poor job of it. Then I’ll think it over and they’ll work to improve it and turn it into something great.
It is intriguing to begin with the idea of “impossible” considering that you are one of the more practical people in the profession.
If you don’t make an effort to do something difficult, someone else will beat you to it. You have to be determined, but you also need to be realistic. I’ve read some of your recent writings and I don’t think the humanoid race you talk about is achievable. It’s impossible to accomplish what they want to do in such a short period of time and with so few people. A company I recently encountered stated that they had 100 years of robotics knowledge between them. The last time I created human-like robots – and I have constructed more human-like robots than any other person in the world – I began with 1,000 years of robotics experience, and we have only improved a little bit.
I think a lot of businesses have been lingering in the background until Tesla’s announcement came out. Even though I’m not an expert on robotics, in my opinion, the most difficult aspect is creating something that is suitable for all purposes, rather than just humanoid robots.
It is difficult to design a general-purpose manipulator. We have been attempting to develop robotic hands since the 1960s, yet there has been minimal progress made since then. In warehouses, the most popular types of robot hands are parallel jaw grippers or suction cups for relocating objects. There is nothing close to general-purpose manipulation available yet.
Do you still have a positive outlook on the design of humanoid robots, given that you have built more of them than anyone else?
They are utilizing the same line of reasoning that I employed when I began the task back in 1992.
We should construct robots that are capable of functioning in the systems and frameworks that we have created for ourselves.
I no longer feel as strongly about that anymore. I observed that it was true that you can achieve a lot more in the long run – which for me means within the span of 50 to 100 years – by constructing machines that are designed for a particular purpose.
Rethink Robotics was involved in the development of robots that resemble humans.
Prior to the Cog, there had been human-like robots, but none with legs. Several firms have developed legged robots, but it is suggested not to get too close to them. Their way of moving is totally different from the way humans move, and when there is a slight instability, a great deal of power is released. If you are in the vicinity, you may get struck by the legs forcefully.
Rethink focused a lot on developing cobots, which are robots that can collaborate with humans.
We used our arms, not our legs, to do it. We instantly responded to anything that wasn’t suitable, and we got the servers to take the energy from the system quickly.
I believe that the majority of people overlook the fact that the robotic systems being employed in Amazon warehouses are kept in cages.
Exactly. You should not be roaming around near the automated machines. They have preventive measures that will turn them off if a person gets on the ground.
What was the outcome of Rethink?
There were a few takeaways. The primary explanation was the disagreement with China concerning trade. We were designing robots in America and then sending them to China, and then we encountered retaliatory tariffs from China, whereas our competitors in Europe did not. There was a deeper issue, which was probably something I caused, because I allowed us to make an expensive robot. Once it was a costly robot, people wanted to use it as a standard industrial robot, and they needed consistency. We required a significantly improved sales and training network to instruct the final users on how to employ the robot in a different manner, as it was based on force.
I still observe robots being used in universities and other research facilities quite frequently.
I believe we achieved a great level of artistry with our work. Unfortunately, we weren’t able to monetarily succeed. We revolutionized the way robots are presented at shows; previously, they were always caged, but now that is no longer the case. We had to struggle to get approval for that in 2013, in Chicago, and that is now the norm.
Melonee Wise informed me in the past that Fetch was still constructing robots similar to those of Willow Garage since they can be used as an effective way to attract new personnel; however, they cannot maintain a company with them.
At iRobot, we had a 10-pack version of the Roomba, which could be bought for $1000, that was meant to be used to teach robotics. This wasn’t a major part of our market though; it was more of a side venture. We believed that we needed to help create more people knowledgeable in robotics.
Over the years, I have inquired of Colin Angle and Helen Greiner the same question: How long did it take iRobot to create the Roomba, and what was the defining moment for them?
In 1997, I traveled to Taipei and was shown the ropes of manufacturing in China by a local. I spent a few weeks with him and we were tasked with making toys inexpensively. We ended up partnering with Hasbro and produced a product called “My Real Baby,” a humanoid figure, which was then sold in stores. We discovered a way to construct items cheaply, and the objective with the Roomba was to see how inexpensively it could be made. We didn’t try to make a complex design more economical. Instead, we asked people what they would pay for it. When the response was $200, that was the primary design criteria. It had to be sold for $200.
In recent years, there has been a concerted effort to create software solutions that are not dependent on specific hardware, in order to facilitate the use of robotics in businesses. Is Robust AI working in this realm?
We agree and disagree. We believe that a lot of these businesses are failing to recognize the potential of silicon. Silicon enables us to perform detailed calculations near cameras. This allows us to do a lot more compared to conventional mobile robots that mainly use lidar, which is a single laser scanner. Consequently, we don’t have access to as much information about the world. These cameras with their processors are amazingly inexpensive. You can execute numerous neural models and receive a highly detailed 3D and labeled explanation of what is in the environment.
When I looked back at your forecasts, I was glad to see the emphasis on robotics for the elderly. It’s been popular in Japan for a while, but I don’t think it’s been talked about enough in the U.S. It makes sense to me that this would be the way to make robots more commonplace in our homes.
I’m sure there will be a huge demand for anything that assists the elderly to remain in their homes with autonomy and respect. I’m not sure what form this will take, however you could say something like the Roomba helps people stay in their residences since it can clean the floors without much effort. This will be very sought after due to the large demographic shift across the globe where there are fewer young people to look after the elderly. Any assistance that can be provided there will be beneficial.
I understand that Helen would not share my opinion, however, the robot vacuum cleaner is still the most popular home robot. It has been a long time since the first Roomba was released, so why has it been so hard to replicate its success?
I wish I had the information, so I could act on it!
You may not be able to see the outcome, but you understand the challenge of reaching it as well as anyone else.
Homes are often full of obstacles, such as steps, which can be very challenging for robots with wheels. Even a single inch of a step can make it almost impossible for them to pass. I have seen some approaches to overcoming one-inch steps, but they cannot manage four-inch steps. This is due to how homes are constructed and altered over time, making it an incredibly difficult issue to solve. Stairs are a big issue for these robots.
There has been much dialogue regarding the connection between advances in smartphone technology and the progress of robotics. I believe the same can be applied to self-driving cars. The major difference is that, unlike smartphones, we have yet to develop true autonomous vehicles.
It is evident that our driver assistance technology has improved. The data is gathered and interpreted, but ultimately the human driver is responsible for the decisions, as it is impossible to have all the information needed to cover every unusual case. Historically, when the way of transportation has changed, the infrastructure had to change as well. A guarantee was made that each new system would replace an old one, and this has prevented a lot of disruption to the infrastructure.
It is assumed that fatalities are inevitable with humans at the wheel. If an autonomous car causes the death of a person, it will cause a major setback that could last for years.
In the U.S., there are 35,000 fatalities caused by car accidents each year. Some tech experts are of the opinion that if this number were to be reduced to 30,000 through the use of self-driving cars, it would be seen as a success. On the contrary, I believe that the maximum number of deaths that would be acceptable from self-driving cars should not exceed 10 per year. Even though 10 deaths are not desirable, it is a much better alternative than 35,000. It needs to be significantly fewer for self-driving cars, even though it may not make sense to some, that is what will happen.
What is your opinion on the current excitement around ChatGPT?
I believe that individuals tend to be overly optimistic. They are confusing performance with skill. When you see someone do something well, you can state that they are proficient in that. We are great at predicting how people will act, however, those same predictions do not work in the same way with technology. Even if an AI system performs really well, it doesn’t necessarily mean that it will be able to work with different data in a new environment.
Have you seen any advancements in robotics over the past few years that have really amazed you?
The results of deep learning and image labeling have been referred to as “perception,” but this is not the same as what a human would perceive. The ability to identify images is an impressive improvement, and we are taking advantage of it. As long as you keep in mind that you are not achieving a human-level of performance, but rather a unique form of labeling. As long as you understand the restrictions, you can create very impressive characteristics in robots that were not conceivable recently.
Verity has just obtained $32 million in their Series B funding round. Ikea is making use of 100 drones from the Zurich-based company in 16 different places in Europe.
Tolga ÃncÃ¼ of the Dutch Ikea holding company, Ingka, expressed that they are investing in technology to make their stores more capable of customer fulfillment and to transform them into omnichannel retailing centers. He added that introducing drones and other innovative tools, such as robots for collecting goods, is beneficial for everyone. It can enhance the wellbeing of their staff, reduce operational costs, and make them more affordable and accessible to their customers.
Verity was founded in 2016 by Raffaello DâAndrea, who was also a part of the team responsible for the inception of Kiva. This project was purchased in 2012 and served as the origin of what is currently known as Amazon Robotics. One of the more interesting uses of drones that I have seen recently is its role in warehouse stock tracking, as they are equipped with image processing capabilities to record items coming in and going out.
This week, Zipline presented a new drone delivery platform which they assert is able to transport freight up to 10 miles in 10 minutes. The P2 model takes a unique combined method, suspending when it gets to its destination and then deploying a small robot that can guide itself on the final leg of the journey to deliver the package.
Joseph Mardell, head of engineering at Zipline, explains that the droid will have a range of sensors such as GPS and vision sensors to help it navigate and make sure that there are no kids, dogs, or other obstacles at a delivery site. Co-founder Keller Rinaudo adds an extra touch of excitement to the project. We assert that we have designed the most similar system to teleportation ever devised – a seamless, ultrafast, convenient, and truly remarkable autonomous logistics system that provides equal service to everyone, no matter where they are located.
Zipline is looking to complete over 10,000 trials with 100 unmanned aerial vehicles this year and expects to inaugurate the system to its patrons by 2024.
This morning, it was announced that Nimble had secured an additional $65 million to extend its reach into third-party logistics services. The San Francisco-based business has stealthily opened several distribution hubs in the United States. CEO Simon Kalouche refused to provide the exact figure or locations, simply stating that there are “more than one but fewer than 10” and spread out geographically.
He stated that the change in direction was not something that was intended from the outset, but rather a consequence of gaining an understanding of customer requirements over the six years since the business was created. He said to TechCrunch, “It changed as we discovered more about the sector. I have visited hundreds of warehouses and with each one, I realized that people are automating almost all facets of the warehouse, however selecting is still the toughest part.” Until you introduce automation in the picking process, it is necessary to have people in the warehouse. It is essential to make the warehouses comfortable, safe and in compliance with OSHA regulations for people. When you automate the picking step, all of those requirements are no longer applicable.
The issue of greenfield versus brownfield has been a significant topic of debate in the industry. The greenfield camp maintains that the ideal factory model is one constructed from the beginning to feature automation, but the brownfield camp argues that this is very costly. Nimble’s new product embraces the greenfield approach, while also giving retailers the option to outsource their inventory needs to a rapidly expanding number of automatic warehouses.
Kate has all the facts about LexxPluss this week. The Japanese startup recently acquired $10.7 million in Series A funding, courtesy of Drone Fund, HAX (SOSV), Incubate Fund, SBI Investment, and DBJ Capital. Duncan Turner of SOSV gave the following statement about the company: “LexxPluss has a clear advantage over other warehousing operations.” Companies specializing in automation are taking advantage of the extensive technical team in Japan, which is well-known for both the industrial robotics market (37% of the worldwide market) and the automotive industry (35% of the American automotive sector).
LeaxPlus is looking to set itself apart from the competition by being more transparent about their hardware and software solutions. Masaya Aso, the company’s founder and CEO, commented to TechCrunch, “We provide plenty of technical information about our technology so that our partners can get a full understanding of how it works and how it can be implemented in their warehouses or factories.” They are capable of taking care of repairs on their own. We strive to make it simpler for everyone to work together by increasing product visibility.
ACR’s initial system takes care of the rebar tying, while the new one is meant to lift and transport it at the job site. According to the company’s founder Stephen Muck, the collaboration between TyBot and IronBot has led to a revolutionary technology that is projected to markedly reduce time and cost of reinforcing steel installation, thus completely revolutionizing the industry.
At Carnegie Mellon University’s Robotics Institute, they recently demonstrated a head-mounted device created to give autonomy to those with physical disabilities. The Head-Worn Assistive Teleoperation (HAT) allows them to use their voice to control a robotic arm, and it is installed in an actual hat.
Speech recognition that is captured by a wireless microphone worn by the person is used to choose between four different robot modes: drive, arm, wrist, and gripper, which is displayed in the image on the left. The head-worn system communicates with the mobile manipulator using Bluetooth and translates the mode the user is in into velocity commands for the robot’s actuators.
Neesha has announced that the TechCrunch Startup Battlefield 200 is now available to apply for. The two hundred best startups will be asked to attend Disrupt in September.
From the 200 startups in the Startup Battlefield, 20 will be chosen as finalists. They will be given the opportunity to showcase their pitches in front of the entire TechCrunch audience on the main stage, while receiving personal coaching and publicity. Additionally, they will be pitching to some of the leading venture capital firms such as Sequoia, Mayfield, and SOSV. Ultimately, the winner of the competition will be awarded a $100,000 prize and the highly sought-after Disrupt Cup.
I would like to observe some robotics businesses included. Make me proud.
We are powerful. Nobody can declare that we are incorrect. Investigating our feelings for a lengthy time. Both of us understanding that you should join Actuator.