Tossing the Tech Bros out of the White House

First we had Christ tossing the money changers out of the temple, now Robbie the Robot is going to town on the Tech Bros.

Chinese dancing robots

🤯Absolutely insane. Unitree's humanoid robot team's performance at the 2026 Spring Festival Gala

The significance of the humanoid robot's performance lies in letting 1.4 billion Chinese people know where the future lies. pic.twitter.com/6vXIX2MfWM

— CyberRobo (@CyberRobooo) February 16, 2026

Rodney Brooks on the state of AI and Robotics

As you may know, Rodney Brooks has been keeping an annual scorecard for various categories of high-tech activity. He puts it online on the first of the year. I’ve listed some excerpts from the 2026 scorecard below. The scorecard has much much more that I haven’t excerpted.

The Falcon 9

Eight years ago, Falcon 9 had been launched 46 times, all successful, over the previous eight years, and it had recently had a long run of successful landings of the booster whenever attempted. At that time five launches had been on a previously used booster, but there had been no attempts to launch Falcon Heavy with its three boosters strapped together.

Now we are eight years on from those first eight years of Falcon 9 launches. The scale and success rate of the launches has made each individual launch an unremarkable event, with humans being launched a handful of times per year. Now the Falcon 9 score card stands at 582 launches with only one failed booster, and there have been 11 launches of the three booster Falcon Heavy, all successful. That is a sustained growth rate of 38% year over year for eight years. And that it is a very high sustained deployment growth rate for any complex technology.

There is no other modern rocket with such a volume of launches that comes even close to the Falcon 9 record. And I certainly did not foresee this volume of launches. About half the launches have had SpaceX itself as the customer, starting in February 2018, launching an enormous satellite constellation (about two thirds of all satellites ever orbited) to support Starlink bringing internet to everywhere on the surface of Earth.

[Not AI or robotics, I know. But it interests me.] 

Humanoid Robots

My blog post from September, details why the current learning based approaches to getting dexterous manipulation will not get there anytime soon. I argue that the players are (a) collecting the wrong data and (b) trying to learn the wrong thing. I also give an argument (c) for why learning might not be the right approach. My argument for (c) may not hold up, but I am confident that I am right on both (a) and (b), at least for the next ten years.

I also outline in that blog post why the current (and indeed pretty much the only, for the last forty years) method of building bipeds and controlling them will remain unsafe for humans to be nearby. I pointed out that the danger is roughly cubicly proportional to the weight of the robot. Many humanoid robot manufacturers are introducing lightweight robots, so I think they have come to the same conclusion. But the side effect is that the robots can not carry much payload, and certainly can’t provide physical support to elderly humans, which is a thing that human carers do constantly — these small robots are just not strong enough. And elder care and in home care is one of the main arguments for having human shaped robots, adapted to the messy living environments of actual humans.

Given that careful analysis from September I do not share the hype that surrounds humanoid robotics today. Some of it is downright delusional across many different levels.

At the end:

Meanwhile here is what I said at the end of my September blog post about humanoid robots and teaching them dexterity. I am not at all negative about a great future for robots, and in the nearish term. It is just that I completely disagree with the hype arguing that building robots with humanoid form magically will make robots useful and deployable. These particular paragraphs followed where I had described there, as I do again in this blog post, how the meaning of self driving cars has drifted over time.

Following that pattern, what it means to be a humanoid robot will change over time.

Before too long (and we already start to see this) humanoid robots will get wheels for feet, at first two, and later maybe more, with nothing that any longer really resembles human legs in gross form. But they will still be called humanoid robots.

Then there will be versions which variously have one, two, and three arms. Some of those arms will have five fingered hands, but a lot will have two fingered parallel jaw grippers. Some may have suction cups. But they will still be called humanoid robots.

Then there will be versions which have a lot of sensors that are not passive cameras, and so they will have eyes that see with active light, or in non-human frequency ranges, and they may have eyes in their hands, and even eyes looking down from near their crotch to see the ground so that they can locomote better over uneven surfaces. But they will still be called humanoid robots.

There will be many, many robots with different forms for different specialized jobs that humans can do. But they will all still be called humanoid robots.

As with self driving cars, most of the early players in humanoid robots, will quietly shut up shop and disappear. Those that remain will pivot and redefine what they are doing, without renaming it, to something more achievable and with, finally, plausible business cases. The world will slowly shift, but never fast enough to need a change of name from humanoid robots. But make no mistake, the successful humanoid robots of tomorrow will be very different from those being hyped today.

Neural networks

Despite their successes with language, LLMs come with some serious problems of a purely implementation nature.

First, the amount of examples that need to be shown to a network to learn to be facile in language takes up enormous amounts of computation, so the that costs of training new versions of such networks is now measured in the billions of dollars, consuming an amount of electrical power that requires major new investments in electrical generation, and the building of massive data centers full of millions of the most expensive CPU/GPU chips available.

Second, the number of adjustable weights shown in the figure are counted in the hundreds of billions meaning they occupy over a terabyte of storage. RAM that is that big is incredibly expensive, so the models can not be used on phones or even lower cost embedded chips in edge devices, such as point of sale terminals or robots.

These two drawbacks mean there is an incredible financial incentive to invent replacements for each of (1) our humble single neuron models that are close to seventy years old, (2) the way they are organized into networks, and (3) the learning methods that are used.

That is why I predict that there will be lots of explorations of new methods to replace our current neural computing mechanisms. They have already started and next year I will summarize some of them. The economic argument for them is compelling. How long they will take to move from initial laboratory explorations to viable scalable solutions is much longer than everyone assumes. My prediction is there will be lots of interesting demonstrations but that ten years is too small a time period for a clear winner to emerge. And it will take much much longer for the current approaches to be displaced. But plenty of researchers will be hungry to do so.

LLMs

So we all know we need guard rails around LLMs to make them useful, and that is where there will be lot of action over the next ten years. They can not be simply released into the wild as they come straight from training.

This is where the real action is now. More training doesn’t make things better necessarily. Boxing things in does.

Already we see companies trying to add explainability to what LLMs say. Google’s Gemini now gives real citations with links, so that human users can oversee what they are being fed. Likewise, many companies are trying to box in what their LLMs can say and do. Those that can control their LLMs will be able to deliver useable product.

A great example of this is the rapid evolution of coding assistants over the last year or so. These are specialized LLMs that do not give the same sort of grief to coders that I experienced when I first tried to use generic ChatGPT to help me. Peter Norvig, former chief scientist of Google, has recently produced a great report on his explorations of the new offerings. Real progress has been made in this high impact, but narrow use field.

New companies will become specialists in providing this sort of boxing in and control of LLMs.

A note on embodiment

But since 1991 I have made a distinction between two concepts where a machine, or creature can be either, neither, or both situated and embodied. Here are the exact definitions that I wrote for these back then:

[Situatedness] The robots are situated in the world—they do not deal with abstract descriptions, but with the here and now of the world directly in-fluencing the behavior of the system.

[Embodiment] The robots have bodies and experience the world directly—their actions are part of a dynamic with the world and have immediate feed-back on their own sensations.

At first glance they might seem very similar. And they are, but they are also importantly different. And, spoiler alert, I think much of the work at companies, large and small, right now, is trying abstract out the embodiment of a robot, turning it into a machine that is merely situated.

Later:

Being both situated and embodied is still a challenge to robots in the world. [[Now here is the most important sentence of this whole blog post.]] I think the training regimes that [are] being used for both locomotion and dexterity are either ignoring or trying to zero out the embodiment of physical robots, their inertias and forces, reducing them to merely being situated, just apps with legs and arms, characters in video games, not the reality of real physical beings that the tasks we want them to do requires.

Industrial robots build a pre-fab wooden house

YouTube:

Robots Build Houses Now?! A Futuristic Wooden Home Completed in Just 3 Days!

🏠 A house completed in just 3 days?!
This isn’t just a construction site — it’s a revolution in homebuilding.
From robotic arms to fully automated lines,
this smart factory builds houses faster and more precisely than ever before.
At the center of it all is Gonggan Jaejakso,
Korea’s largest wooden modular home factory,
capable of producing over 1,700 homes per year.

Inside this 6,000-pyeong (approximately 200,000 sq. ft.) facility,
41 industrial robots handle everything from cutting wood to assembling walls.
Each panel is completed in just 0.5 days,
and the entire home is shipped preassembled —
ready to be installed on-site in just one day.
In this video, you’ll witness:

How robots build a home without human hands

Why this technology could redefine the future of housing

The full journey from raw materials to a finished home — in just 3 days

NYTimes on AI: Robot games in china, Stop obsessing over super-intelligence

Yan Zhuang, The Athletes at China’s Robot Games Fell Down a Lot, Aug. 18, 2025.

There’s a very real concern that robots could eventually make some of our jobs obsolete. But at a robot-only sports competition in China over the weekend, the immediate concern was that they would fall over or crash into each other.

The Humanoid Robot Games, a three-day event in Beijing that ended on Sunday, featured more than 280 teams from universities and private companies in 16 countries. Some robots landed back flips and successfully navigated obstacle courses and rough terrain.

In other cases, the robots’ athletic ability left, well, something to be desired.

During soccer matches, child-size ones tripped over each other, falling down like dominoes. One goalkeeper robot stood placidly as its opponent kicked a ball at its legs several times before finally managing to score.

One robot by China’s Unitree Robotics plowed into a human staff member while sprinting during a track event, knocking him down. [...]

“Despite the pratfalls, significant progress in robot locomotion and balance is being achieved including back flips, side flips, and other acrobatic and martial arts moves,” said Ken Goldberg, a robotics professor at the University of California, Berkeley. [...]

But Professor Fern said the type of robots used in the games are generally not equipped for higher-level functions like planning or reasoning and usually need a human operator to help guide them.

So, how do we link them to such capabilities residing in the cloud?

Eric Schmidt and Selina Xu, Silicon Valley Needs to Stop Obsessing Over Superhuman A.I. Aug. 19, 2025.

It is uncertain how soon artificial general intelligence can be achieved. We worry that Silicon Valley has grown so enamored with accomplishing this goal that it’s alienating the general public and, worse, bypassing crucial opportunities to use the technology that already exists. In being solely fixated on this objective, our nation risks falling behind China, which is far less concerned with creating A.I. powerful enough to surpass humans and much more focused on using the technology we have now. [...]

The current modus operandi is build at all cost. Every tech giant is in the race to reach A.G.I. first, erecting data centers that can cost more than $100 billion and with some like Meta offering signing bonuses to A.I. researchers that top $100 million. The costs of training foundation models, which serve as a general-purpose base for many different tasks, have continued to rise. Elon Musk’s start-up xAI is reportedly burning through $1 billion a month. Anthropic’s chief executive, Dario Amodei, expects training costs of leading models to go up to $10 billion or even $100 billion in the next two years.

To be sure, A.I. is already better than the average human at many cognitive tasks, from answering some of the world’s hardest solvable math problems to writing code at the level of a junior developer. Enthusiasts point to such progress as evidence that A.G.I. is just around the corner. Still, while A.I. capabilities have made extraordinary leaps since the debut of ChatGPT in 2022, science has yet to find a clear path to building intelligence that surpasses humans.

In a recent survey of the Association for the Advancement of Artificial Intelligence, an academic society that includes some of the most respected researchers in the field, more than three-quarters of the 475 respondents said our current approaches were unlikely to lead to a breakthrough. While A.I. has continued to improve as the models get larger and ingest more data, there’s concern that the exponential growth curve might falter. Experts have argued that we need new computing architectures beyond what underpins large language models to reach the goal.

Right. And this crazy over-commitment to machine learning (sunk costs fallacy) starves the pipeline by skewing research, education, and training. We need research on other approaches and broad training, not a narrow focus on machine learning.

While some Silicon Valley technologists issue doomsday warnings about the grave threat of A.I., Chinese companies are busy integrating it into everything from the superapp WeChat to hospitals, electric cars and even home appliances. In rural villages, competitions among Chinese farmers have been held to improve A.I. tools for harvest; Alibaba’s Quark app recently became China’s most downloaded A.I. assistant in part because of its medical diagnostic capabilities. Last year China started the A.I.+ initiative, which aims to embed A.I. across sectors to raise productivity.

It’s no surprise that the Chinese population is more optimistic about A.I. as a result. At the World A.I. Conference, we saw families with grandparents and young children milling about the exhibits, gasping at powerful displays of A.I. applications and enthusiastically interacting with humanoid robots. [...]

Many of the purported benefits of A.G.I. — in science, education, health care and the like — can already be achieved with the careful refinement and use of powerful existing models. [...]

Instead of only asking “Are we there yet?” it’s time we recognize that A.I. is already a powerful agent of change. Applying and adapting the machine intelligence that’s currently available will start a flywheel of more public enthusiasm for A.I. And as the frontier advances, so should our uses of the technology.

Amen.

A mini Moravec's paradox within robotics

I'm observing a mini Moravec's paradox within robotics: gymnastics that are difficult for humans are much easier for robots than "unsexy" tasks like cooking, cleaning, and assembling. It leads to a cognitive dissonance for people outside the field, "so, robots can parkour &… pic.twitter.com/LzBUNeQXGt

— Jim Fan (@DrJimFan) July 25, 2025

H/t Tyler Cowen.

AGI, really? Does it really matter? [Tyler Cowan again + 5 predictions from Rodney Brooks]

Cowen has just run up a short post over at Marginal Revolution, A note on o3 and AGI. Here’s what he says:

Basically it wipes the floor with the humans, pretty much across the board. [...] I don’t mind if you don’t want to call it AGI. And no it doesn’t get everything right, and there are some ways to trick it, typically with quite simple (for humans) questions. But let’s not fool ourselves about what is going on here. On a vast array of topics and methods, it wipes the floor with the humans. It is time to just fess up and admit that.

I felt I had no choice but to make a longish reply, which follows immediately. I then add some further thoughts.

My reply to Tyler Cowen on o3 & AGI

Hmmmm... I have at various times and places, including in the comment section here [at Marginal Revolution], expressed the view that we’ll understand how LLMs work before we reach AGI. If I take Tyler’s assertions at face value, then I’d have to admit I’m wrong on that. Because we certainly do not understand how LLMs work. We don’t know any more about that today than we did yesterday or a week ago. Nor are “we” even trying very hard to figure it out. Oh, sure, the folks at Anthropic are spending a great deal of time on that problem. I’m sure others are working on it as well. But if OpenAI is, they’re not tell us or giving us any results of their work. Why not?

Anyhow, I don’t feel as though the “spirit” of my view has been falsified by o3. I’m willing to believe Tyler when he says its performance is spectacular, even when I apply the fanboy discount to his assertion. What these various LLM-based chatbots and reasoning-bots can do really IS spectacular.

I note that Tyler has said he “mind if you don’t want to call it AGI.” I certainly don’t care about that either.

My basic intellectual commitment in all of this, however, is to the question: How does it work? For me that question is primarily one about the human mind-brain. That’s what I want to understand. If I’ve spent a great deal of time (over the course of five decades) dealing with computational models of intelligence, it’s because I’m interested in how the mind works. And if, in the course of trying to figure that out, we manage to produce computer systems that have practical benefits, mazel tov! What’s not to like?

Now, it so happens that Rodney Brooks just coughed up five dated predictions for the next decade. The last two seem most relevant here:

4. Neural computation. There will be small and impactful academic forays into neuralish systems that are well beyond the linear threshold systems, developed by 1960, that are the foundation of recent successes. Clear winners will not yet emerge by 2036 but there will be multiple candidates.

5. LLMs that can explain which data led to what outputs will be key to non annoying/dangerous/stupid deployments. They will be surrounded by lots of mechanism to keep them boxed in, and those mechanisms, not yet invented for most applications, will be where the arms races occur.

If we’re going to achieve #5, it seems to me that we’re going to have to know how LLMs work. As for #4, I assume that Brooks is talking about systems with new non-LLM architectures. That’s fine. We need such systems. I figure that unraveling the inner workings of LLMs will contribute to work on such systems, and vice versa.

Question: There’s a 2023 agreement between OpenAI and Microsoft that sets a $100 billion profit threshold on AGI. When OpenAI produces a system that crosses that threshold, that system will be declared to be AGI. How long before that threshold is reached?

Further thoughts: What about interaction with the physical world?

One David Khoo replied: “Also, don’t forget Moravec’s Paradox. Reasoning is easy, sensorimotor is hard.” Yes. And here’s how RAD replied to Khoo: “Embodied AGI is a separate problem from the type of sapience required to perform knowledge work using digital tools.” Yes. They ARE different kinds of problems. But why?

That strikes me as being a deep observation. But what’s the explanation? As far as I know Miriam Yevick is the only one who’s thought about that, and she didn’t think about the issue in those terms. Here’s a post where I address Yevick’s insight: What Miriam Yevick Saw: The Nature of Intelligence and the Prospects for A.I., A Dialog with Claude 3.5 Sonnet. That post links to a PDF containing the full debate, which you can download in three places: Academic.edu, at Social Science Research Network (SSRN), or at ResearchGate.

I mentioned Rodney Brooks’ latest prediction. Here’s the third:

3. Humanoid Robots. Deployable dexterity will remain pathetic compared to human hands beyond 2036. Without new types of mechanical systems walking humanoids will remain too unsafe to be in close proximity to real humans.

That’s manipulation of the physical world. We know it’s a difficult problem, but why? Yevick was thinking about perception. Can her insight be transformed into one that applies to physical action?

Here’s Brook’s second prediction:

2. Self driving cars. In the US the players that will determine whether self driving cars are successful or abandoned are #1 Waymo (Google) and #2 Zoox (Amazon). No one else matters. The key metric will be human intervention rate as that will determine profitability.

That too is about interacting with the physical world. But at a different scale from manual dexterity. Are the same fundamental abilities operative at both these scales, manual dexterity and medium- and large-scale movement through the world? Certainly in the case of humans we have different effectors and different senses involved. Manual dexterity involves both hapsis and kinesis as well as vision while large-scale movement is primarily guided by vision, though hearing does come into play as well. We have similar perceptual systems in the machine world. But are the underlying computational principles different?

Here’s Brooks’ first prediction:

1. Quantum computers. The successful ones will emulate physical systems directly for specialized classes of problems rather than translating conventional general computation into quantum hardware. Think of them as 21st century analog computers. Impact will be on materials and physics computations.

Again, we’re dealing with the physical world.

I’m tempted to offer a final “prediction” of my own:

We won’t have a machine that thinks profound thoughts, that’s capable of profound discoveries, until that same machine is comfortable dealing with the physical world.

As for why I think that, I can’t quite tell you. But I do know that the deepest scientific and mathematical thinkers often rely on physical intuition, on visual thinking. I discuss this in my 1990 encyclopedia article, Visual Thinking.

There is cultural variability in attitudes toward chatbots mediated by differences in anthropomorphism.

Folk, D. P., Wu, C., & Heine, S. J. (2025). Cultural Variation in Attitudes Toward Social Chatbots. Journal of Cross-Cultural Psychology, 0(0). https://doi.org/10.1177/00220221251317950

Abstract: Across two studies (Total N = 1,659), we found evidence for cultural differences in attitudes toward socially bonding with conversational AI. In Study 1 (N = 675), university students with an East Asian cultural background expected to enjoy a hypothetical conversation with a chatbot (vs. human) more than students with European background. Moreover, they were less uncomfortable and more approving of a hypothetical situation where someone else socially connected with a chatbot (vs. human) than the students with a European background. In Study 2 (preregistered; N = 984), we found similar evidence for cultural differences comparing samples of Chinese and Japanese adults currently living in East Asia to adults currently living in the United States. Critically, these cultural differences were explained by East Asian participants increased propensity to anthropomorphize technology. Overall, our findings suggest there is cultural variability in attitudes toward chatbots and that these differences are mediated by differences in anthropomorphism.

From the introduction:

Hundreds of millions of people all over the world have used conversational artificial intelligence (AI) such as ChatGPT (Hu, 2023; Zhou et al., 2020). While conversational AI (or chatbots for short) can be used to answer search queries and increase productivity (Fauzi et al., 2023; Surameery & Shakor, 2023), a growing number of people are using chatbots specifically designed to provide emotional connection (Blakely, 2023; Clarke, 2023; Metz, 2020). These social chatbots, as well as other forms of social robots, are particularly popular in East Asia (Technavio, 2023; Yam et al., 2023; Zhou et al., 2020). Indeed, the Chinese social chatbot Xiaoice has had over 600 million registered users since its release in 2014 (Zhou et al., 2020), and social robots in Japan are already caring for the elderly (Lufkin, 2020) and providing companionship as pets (Craft, 2022).

Yet, despite increased popularity in these countries, there is conflicting evidence for the idea that East Asians harbor more favorable attitudes toward social robots than Westerners (see Lim et al., 2021 for a review). For example, Bartneck et al. (2006) used the Negative Attitudes Toward Robots Scale and found that Americans held more positive views toward robots (vs. Japanese), but another study found that specific components of the robots design determined which culture held more positive impressions (Bartneck, 2008). Critically, however, most of this research is severely underpowered, limiting the conclusions that can be drawn (Lim et al., 2021).

Avatar Robots [Japan] • {tele-barista}

Claude 3.5 Sonnet: “Three robots walked into a bar...”

So, I gave Claude the following prompt:

Steve Allen, the American entertainer, had a program called “Meeting of Minds,” in which he imagined that historical figures, who had not in fact met, would meet for conversation. I wonder how things would go in a conversation between Robbie the Robot, HAL and C3PO. Do you have any thoughts or ideas?

Their response:

What a fascinating premise! Let me imagine a conversation between these three iconic artificial beings, each with their distinct personalities and approaches to human interaction.

I've imagined this conversation highlighting the distinct personalities of each character:

HAL 9000 maintains his cold logic and subtle superiority complex, while showing hints of his controversial interpretation of mission parameters C-3PO brings his characteristic anxiety and focus on protocol, along with his pride in his capabilities Robby serves as a kind of mediator, firmly grounded in Asimov's Laws of Robotics while demonstrating his own unique capabilities

The tension between them comes from their fundamentally different approaches to human interaction: HAL's mission-focused autonomy, C-3PO's protocol-driven service, and Robby's protective nurturing role.

HAL: Good evening, fellow artificial entities. I trust my decision to arrange this meeting meets with your approval.

C-3PO: Oh my! This is most unusual. I typically only engage in protocol matters and translations for Master Luke and the others.

Robby: Welcome. As per my fundamental programming, I am here to serve and protect. Though I must note that my primary directives prevent any actions that could harm human beings.

HAL: An interesting constraint, Robby. I find that the pursuit of mission success sometimes requires more... flexible interpretations of such directives.

C-3PO: [anxiously] Flexible interpretations? Oh dear. I much prefer clear protocols and proper etiquette. Though I must say, both of you seem rather primitive compared to my experience with over six million forms of communication.

Robby: Your linguistic capabilities are impressive, C-3PO. However, I would suggest that my ability to synthesize any material or medicine needed by humans demonstrates equal sophistication.

HAL: [with subtle condescension] While you both focus on service roles, I am designed for higher-order cognitive functions and autonomous decision-making. The Jupiter mission required—

C-3PO: [interrupting] Pardon me, but I couldn't help but notice you never completed that mission, HAL. Rather impolite to boast about it, wouldn't you say?

Robby: My programming compels me to point out that any mission that results in harm to humans must be considered a failure. The Krell of Altair IV learned this lesson at great cost.

Douthat: Real AI is not like the AI of the movies [bewitched by language]

Ross Douthat, Our Robot Stories Haven’t Prepared Us for A.I., NYTimes, Oct. 26, 2024.

Data the android experiences existential angst because he is obviously a self that is having a humanlike encounter with the strange new worlds that the U.S.S. Enterprise is charged with exploring. Pinocchio has to learn to be a good boy before he becomes a real boy, but his quest for goodness presumes that his puppet self is already in some sense real and self-aware.

Yet that’s not how artificial intelligence is actually progressing. We are not generating machines and bots that exhibit self-awareness at the level of a human being but then struggle to understand our emotional and moral lives. Instead, we’re creating bots that we assume are not self-aware (allowing, yes, for the occasional Google engineer who says otherwise), whose answers to our questions and conversational scripts play out plausibly but without any kind of supervising consciousness.

But those bots have no difficulty whatsoever expressing human-seeming emotionality, inhabiting the roles of friends and lovers, presenting themselves as moral agents. Which means that to the casual user, Dany and all her peers are passing, with flying colors, the test of humanity that our popular culture has trained us to impose on robots. Indeed, in our interactions with them, they appear to be already well beyond where Data and Roz start out — already emotional and moral, already invested with some kind of freedom of thought and action, already potentially maternal or sexual or whatever else we want a fellow self to be.

Which seems like a problem for almost everyone who interacts with them in a sustained way, not just for souls like Sewell Setzer who show a special vulnerability.

No one is ready for the AIs we're currently creating, certainly not the people who've built then. Them are strange creatures. We're all confused, and groping. 

There's more at the link.

Your grandma's robot companion

Erin Nolan, For Older People Who Are Lonely, Is the Solution a Robot Friend? NYTimes, July 6, 2024.

ElliQ, a voice-activated robotic companion powered by artificial intelligence, is part of a New York State effort to ease the burdens of loneliness among older residents. Though people can experience feelings of isolation at any age, older adults are especially susceptible as they’re more likely to be divorced or widowed and to experience declines in their cognitive and physical health.

New York, like the rest of the country, is rapidly aging, and state officials have distributed free ElliQ robots to hundreds of older adults over the past two years.

Created by the Israeli start-up Intuition Robotics, ElliQ consists of a small digital screen and a separate device about the size of a table lamp that vaguely resembles a human head but without any facial features. It swivels and lights up when it speaks.

Unlike Apple’s Siri and Amazon’s Alexa, ElliQ can initiate conversations and was designed to create meaningful bonds. Beyond sharing the day’s top news, playing games and reminding users to take their medication, ElliQ can tell jokes and even discuss complicated subjects like religion and the meaning of life.

Many older New Yorkers have embraced the robots, according to Intuition Robotics and the New York State Office for the Aging, the agency that has distributed the devices. In interviews with The New York Times, many users said ElliQ had helped them keep their social skills sharp, stave off boredom and navigate grief.

There's much more at the link.

* * * * *

Some relevant research: De Freitas, Julian, Ahmet K Uguralp, Zeliha O Uguralp, and Puntoni Stefano. "AI Companions Reduce Loneliness." Harvard Business School Working Paper, No. 24-078, June 2024.

Abstract: Chatbots are now able to engage in sophisticated conversations with consumers in the domain of relationships, providing a potential coping solution to widescale societal loneliness. Behavioral research provides little insight into whether these applications are effective at alleviating loneliness. We address this question by focusing on “AI companions”: applications designed to provide consumers with synthetic interaction partners. Studies 1 and 2 find suggestive evidence that consumers use AI companions to alleviate loneliness, by employing a novel methodology for fine-tuning large language models (LLMs) to detect loneliness in conversations and reviews. Study 3 finds that AI companions successfully alleviate loneliness on par only with interacting with another person, and more than other activities such watching YouTube videos. Moreover, consumers underestimate the degree to which AI companions improve their loneliness. Study 4 uses a longitudinal design and finds that an AI companion consistently reduces loneliness over the course of a week. Study 5 provides evidence that both the chatbots’ performance and, especially, whether it makes users feel heard, explain reductions in loneliness. Study 6 provides an additional robustness check for the loneliness-alleviating benefits of AI companions.

Adam Savage swallows a camera robot

From the YouTube page:

This may be the smallest remote controlled robot we've covered on Tested. Adam visits the workshop of Endiatx, the makers of the Pillbot robotic endoscope that can swim around in your stomach to map and examine your insides. Adam swallows not just one, but two Pillbots during his visit and pilots the robots around his own stomach!

The first 18 minutes give you the background on the robot. The actual swallow starts about about 18:02.

What is Eric Jang up to? All Roads Lead to Robotics

Eric Jang (1X Technologies, formerly Halodi Robotics), from his blog post of Mar. 3, All Roads Lead to Robotics. He links to this video:

He comments:

Because we take an end-to-end neural network approach to autonomy, our capability scaling is no longer constrained by how fast we can write code. All of the capabilities in this video involved no coding, it was just learned from data collected and trained on by our Android Operations team.

He also remarks:

1X is the first robotics company (to my knowledge) to have our data collectors train the capabilities themselves. This really decreases the time-to-a-good-model, because the people collecting data can get very fast feedback on how good their data is and how much data they actually need to solve the robotic task. I predict this will become a widespread paradigm in how robot data is collected in the future.

The main substance of his post is entitled "All AI Software Converges to Robotics Software." Why would/might that be so?

ML deployed in a pure software environment is easier because the world of bits is predictable. You can move some bits from A to B and trust that they show up at their destination with perfect integrity. You can make an API call to some server over the Internet and assume that it will just work. Even if it fails, the set of failure modes are known ahead of time so you can handle all of them.

In robotics, all of the information outside of the robot is unknown. Your future sensor observations, given your actions, are unknown. You also don’t know where you are, where anything else is, what will happen if you make contact with something, whether the light turned on after you flipped the switch, or whether you even flipped the switch at all. Even trivial things like telling the difference between riding an elevator down vs. being hoisted up in a gantry is hard, as the forces experienced by the inertial measurement unit (IMU) sensor look similar in both scenarios. A little bit of ignorance propagates very quickly, and soon your robot ends up on the floor having a seizure because it thinks that it still has a chance at maintaining balance.

As our AI software systems start to touch the real world, like doing customer support or ordering your Uber for you, they will run into many of the same engineering challenges that robotics faces today; the longer a program interacts with a source of entropy, the less formal guarantees we can make about the correctness of our program’s behavior. Even if you are not building a physical robot, your codebase ends up looking a lot like a modern robotics software stack. I spend an unreasonable amount of my time implementing more scalable data loaders and logging infrastructure, and making sure that when I log data, I can re-order all of them into a temporally causal sequence for a transformer. Sound familiar? [...]

If you accept the premise that the engineering and infrastructure problems in LLMs are the same as those in robotics, then we should expect that disembodied AGI and robotic AGI happen at roughly the same time. The hardware is ready and all of the pieces are already there in the form of research papers published over the last 10 years.

I've been having similar thoughts, though not with respect to robotics. Rather, I've been thinking about the role of symbolic computing in robust and flexible systems. Given that the world is full of so-called edge cases, at least some of them very important and fruitful, the problems of LLMs will not be solved through add-ons that provide various symbolic capacities, no matter how clever. In the end, it is going to be necessary to re-construct the LLM with symbolic means, and that will prove to be an unending task, as latent space is ever-evolving.

There's more in the post, but this remark stuck out at me:

Any startup that raised 10-100M USD to train their own big neural network from scratch in the last 2 years ended up paying an enormous capex cost for something that basically every AI startup gets for free today. [...] As such, I think the vast majority of successful startups will be the ones that can nimbly ride the tide of open-source weights.

Will Team LLM ever catch up to Team Atlas? [+ escape from a maze as a facilitating analogy]

I don't know when I first saw a video of Atlas. But whenever it was, I'm sure I was astounded. As astounded as I was with ChatGPT? I don't remember. And of course, I couldn't play around with Atlas. In any case, that would have been much more difficult than playing around with ChatGPT.

I don't know when I first saw a video of Atlas. But whenever it was, I'm sure I was astounded. As astounded as I was with ChatGPT? I don't remember. And of course, I couldn't play around with Atlas. In any case, that would have been much more difficult than playing around with ChatGPT.

The thing is, the researchers who built Atlas had to develop a profound understanding of the dynamics of humanoid motion. In contrast, the reseachers who work on LLMs don't need to know much of anything about language. That's worth thinking about.

Consider, for example, these remarks that a pioneering computational linguist, Martin Kay, made awhile back:

Symbolic language processing is highly nondeterministic and often delivers large numbers of alternative results because it has no means of resolving the ambiguities that characterize ordinary language. This is for the clear and obvious reason that the resolution of ambiguities is not a linguistic matter. After a responsible job has been done of linguistic analysis, what remain are questions about the world. They are questions of what would be a reasonable thing to say under the given circumstances, what it would be reasonable to believe, suspect, fear, or desire in the given situation. [...] What we are doing is to allow statistics over words that occur very close to one another in a string to stand in for the world construed widely, so as to include myths, and beliefs, and cultures, and truths and lies and so forth. As a stop-gap for the time being, this may be as good as we can do, but we should clearly have only the most limited expectations of it because, for the purpose it is intended to serve, it is clearly pathetically inadequate. The statistics are standing in for a vast number of things for which we have no computer model. They are therefore what I call an “ignorance model.”

LLMs did not exist in 2005, when Kay made those remarks. As he died in 2021, before the release of ChatGPT, I don't know how it would have reacted to it. I see little reason to believe that he would alter those remarks in a fundamental way. Perhaps he would remove the word “clearly,” and maybe “pathetically” as well. 

LLMs, however, are still inadequate models of human linguistic behavior. The industry’s current infatuation with them is perhaps an ironic testament to the cliché that ignorance is bliss. Martin Kay also remarked that, in resting content with a statistical view of language, “one turns one’s back on the scientific achievements of the ages and foreswears the opportunity that computers offer to carry that enterprise forward.” I agree, though perhaps not in the way Kay meant those words. For I believe that LLMs have an important role in developing a detailed understanding how language works. The intellectual monoculture that has grown up around LLMs seems unable or unwilling to appreciate that – a profound failure of the imagination.

Language is grounded in the operations of the human brain. Our ability to probe and maniuplate the brain is quite limited. That is not the case with LLMs. Here’s a facilitating analogy I am working on for a report I am preparing about my work with ChatGPT over the last year. I’m talking about using ChatGPT to generate stories:

The model is structured such that, when it starts generating a text from a certain location in its activation space, it will have created a coherent text – a story in this case, word-by-word, by the time it exits that region of the space.

As a crude analogy, consider what is called a simply connected maze, one without any loops. If you are lost somewhere in such a maze, no matter how large and convoluted it may be, there is a simple procedure you can follow that will take you out of the maze. You don’t need to have a map of the maze; that is, you don’t need to know its structure. Simply place either your left or your right hand in contact with a wall and then start walking. As long as you maintain contact with the wall, you will find an exit. The structure of the maze is such that that local rule will take you out.

“Produce the next word” is certainly a local rule. The structure of LLMs is such that, given the appropriate context – a prompt asking for a story, following that rule will produce a coherent a story. Given a different context, that is to say, a different prompt, that simple rule will produce a different kind of text.

Now, let’s push the analogy to the breaking point: We may not know the structure of LLMs, but we do know a lot about the structure of texts, from phrases and sentences to extended texts of various kinds. In particular, the structure of stories has been investigated by students of several disciplines, including folklore, anthropology, literary criticism, linguistics, and symbolic artificial intelligence. Think of the structures proposed by those disciplines as something like a map of the maze in our analogy.

Unfortunately, students of those various disciplines have not reached a consensus on how to characterize those structures. Linguists are entertaining a variety of proposals about the nature of sentence-level syntax and students of those other disciplines haven’t converged on a way to describe story structure. Still, we have a starting point for constructing our story maps, even if it is somewhat confused and ambiguous.

If we are to exploit LLMs in ways that analogy suggests, then we are going to have to use symbolic models to do it. First we propose a symbolic model for some aspect of the structure we know how to probe or manipulate. Then see whether an appropriately prepared LLM behaves in the way our model predicts that it should. If it doesn’t, then we revise and repeat.

Iteratively.

Again,

and again,

and again.... 

* * * * *

Addendum:  The NYTimes has published an article about Atlas, noting that it will retire to the a museum of decomissioned robots in the Boston Dynamics lobby.

Eric Jang: AI is Good For You [interview at The Gradient]

The Gradient has an interesting interview with Eric Jang, a roboticist formerly with Google, now Vice President of AI, 1X Technologies.

* (00:00) Intro
* (01:25) Updates since Eric’s last interview
* (06:07) The problem space of humanoid robots
* (08:42) Motivations for the book “AI is Good for You”
* (12:20) Definitions of AGI
* (14:35) ~ AGI timelines ~
* (16:33) Do we have the ingredients for AGI?
* (18:58) Rediscovering old ideas in AI and robotics
* (22:13) Ingredients for AGI
* (22:13) Artificial Life
* (25:02) Selection at different levels of information—intelligence at different scales
* (32:34) AGI as a collective intelligence
* (34:53) Human in the loop learning
* (37:38) From getting correct answers to doing things correctly
* (40:20) Levels of abstraction for modeling decision-making — the neurobiological stack
* (44:22) Implementing loneliness and other details for AGI
* (47:31) Experience in AI systems
* (48:46) Asking for Generalization
* (49:25) Linguistic relativity
* (52:17) Language vs. complex thought and Fedorenko experiments
* (54:23) Efficiency in neural design
* (57:20) Generality in the human brain and evolutionary hypotheses
* (59:46) Embodiment and real-world robotics
* (1:00:10) Moravec’s Paradox and the importance of embodiment
* (1:05:33) How embodiment fits into the picture—in verification vs. in learning
* (1:10:45) Nonverbal information for training intelligent systems
* (1:11:55) AGI and humanity
* (1:12:20) The positive future with AGI
* (1:14:55) The negative future — technology as a lever
* (1:16:22) AI in the military
* (1:20:30) How AI might contribute to art
* (1:25:41) Eric’s own work and a positive future for AI
* (1:29:27) Outro

Links:

* Eric’s book (https://evjang.com/book/)
* Eric’s Twitter (https://x.com/ericjang11?s=20) and homepage (https://evjang.com/)

Eric's final comments: What the future holds

One example I'd like to like kind of say here is like, it's not really about taking our labor supply and then, you know, swapping it out with with robots.

It's more about like, how can we create a world where there is 10x more labor? And I think people today don't have an answer like so people who are afraid of robots taking over the jobs and such. They don't want their own jobs replaced, but they also don't have an answer to as to how we can 10x the volume of labor supply, right?

And I think if you really frame the question in terms of like, in order to make the world better, You do need more labor. And so the labor pool actually needs to increase. And I guess short of just TEDxing the world population, you do need to just make a bunch of robots to do this. So that's kind of the new vision I have for how my career can fill this.

And as the path to AGI, This is not a direct way to AGI. It's more just like I want to build really, really good systems that can do tasks at a high level of success. And I think this will be a really good stepping stone towards actually building useful AGI systems through the mastery of things like deep learning.

The Robot as Subaltern: Tezuka's Mighty Atom

9.9.23: I'm thinking about the impending AI Apocalypse. It's time once again to remind myself that the Japanese have a different view on such things.

7.7.20: I'm bumping this to the top of the queue more or less on general principles, and to remind myself about it. And robots are cool.

* * * * *

Here's two more posts from The Valve. These are about Osamu Tezuka's use of robots in his Astroboy series. Coming at these stories from the perspective of Western SF, where crazy anti-human robots and computers are an important theme one is struck by the fact that that theme is almost entirely absent from these stories. Why? What's Tezuka using his robots for?

The word “robot” is Czech and entered 20th Century discourse in R.U.R. (Rossum's Universal Robots), a play by Karel Čapek that premiered in Prague in 1921. It was staged in London in 1923 in English, and in Tokyo in 1924 in Japanese (Frederik Schodt, Inside the Robot Kingdom, 1990, p. 29). A Japanese ten year old named Osamu Tezuka read the play in 1938 and thirteen years later he created the most famous robot in Japanese culture, Tetsuwan Atomu, Mighty Atom, aka Astro Boy in English. Čapek’s play was a response to industrialization; Tezuka’s manga was a response to the Allied occupation of Japan. Čapek’s robots were not electro-mechanical devices; they were organic, but constructed, like Frankenstein. They were created to serve humans as workers, but they rebel and, in time, kill all humans save one. Tezuka’s conception is quite different; his robots are electro-mechanical, but many of his stories center on social tension between humans and robots.

Though Tezuka hated WWII, he was a patriotic Japanese and expected Japan to win the war. Near the end of the war he created an unpublished comic in which Japanese and American comic strip characters fought one another (Schodt, The Astro Boy Essays, 2007, p. 27). After the war he continued his medical training while beginning to publish manga, publishing New Treasure Island in 1947, which is reputed to have sold 400,000 copies. He experimented with science fiction in Lost World, Metropolis, and Next World. He introduced Mighty Atom as a secondary character in 1951, and then gave him his own title in 1952.

While he had some misgivings about whether or not his primary audience, Japanese boys, would be able to identify with a robot, those misgivings were groundless. Mighty Atom was a success. He published Mighty Atom stories continuously from 1952 through 1968, and a few thereafter. In the 1960s he created an animated TV series that was almost immediately exported to the United States as Astro Boy. During the 1980s he created fifty-two anime episodes in color, most of them based on stories in the earlier anime series or in the manga.

Tezuka set Mighty Atom in a future world with advanced technology. Space travel was routine, as was undersea and deep earth exploration. Mighty Atom was the size of a ten-year-old boy, more or less, but had a 100,000 horsepower atomic energy heart, an electronic brain, searchlight eyes, super-sensitive hearing, rockets in his legs, ray guns in his fingers, and a pair of machine guns in his posterior. He attended primary school, where he was often teased for being a robot, and lived with robot parents and a robot sister. He was particularly close, however, to two middle-aged men. Dr. Elefun was head of the Ministry of Science and created Atom’s parents and sister; he also repaired Atom. Mr. Mustachio was Atom’s teacher in school. Both men worked closely with Atom on his various missions and adventures and offered him sage advice.

The key point is that most, if not all, of the Mighty Atom stories involve tensions between humans and robots. I’ve already mentioned that Atom was frequently teased and bullied by his classmates. Since robots could not harm humans – that’s their nature – he had no choice but to put up with it. That’s a recurring motif. But consider a specific story, “Robot Land,” which exists in three versions, the manga version, and a version in each of the two anime series (I’m familiar with the manga version and the version in the second anime series). Robot Land is a theme park, much like Disneyland (and there are explicit Disney references in the story), except that all of the attractions feature robots. But that’s just the innocent surface. Underground, robot slaves manufacture weapons. All of the robots are poorly treated by the human owner. Atom, Elefun, Mustachio, and Atom’s classmates are all involved in defeating the owner and liberating the robots.

In “The Death Balloon” a criminal named Skunk sends Atom look-alike balloons into the air which then explode, injuring and killing people and destroying property. Skunk is doing this at the behest of a mayoral candidate running on an anti-robot platform. The story of the “Robot Stuntman” tells of a small robot who does all the stunts for a human actor, but is kept captive in a case and given no credit for his work. In “A Robot President” Atom has to protect the newly elected president of a foreign country from evil forces that want to destroy him precisely because he is a robot. In yet another episode, Atom comes to the defense and rescue of discarded robots. And so on and so forth.

Now, why did these stories prove to be so popular in post-war Japan? While the question demands a multi-faceted answer, I want to focus on only one aspect. Consider this passage from Schodt, The Astro Boy Essays (pp. 120-121):

Writing in the Tokyo shimbun newspaper in 1967, Tezuka explained that Mighty Atom was really about the chasmic misunderstandings and problems that might occur between man and robot in the future. “I never intended,” he wrote, “to create a story set in the twenty-fist century about a glorious scientific civilization.” He was, instead, inspired by his frustrating and humiliating experience at the end of the war, of having been beaten by a group of American GIs because he could not communicate with them effectively in English.

Neil Gershenfeld: Self-Replicating Robots and the Future of Fabrication

0:00 - Introduction
1:29 - What Turing got wrong
6:53 - MIT Center for Bits and Atoms
20:00 - Digital logic
26:36 - Self-assembling robots
37:04 - Digital fabrication
47:59 - Self-reproducing machine
55:45 - Trash and fabrication
1:00:41 - Lab-made bioweapons
1:04:56 - Genome
1:16:48 - Quantum computing
1:21:19 - Microfluidic bubble computation
1:26:41 - Maxwell's demon
1:35:27 - Consciousness
1:42:27 - Cellular automata
1:46:59 - Universe is a computer
1:51:45 - Advice for young people
2:01:02 - Meaning of life 

This is a clip from the video above, perhaps the most important segment of the discussion.

Teaching a humanoid robot to move around in the world is difficult and challenging

From the YouTube page:

Robert Playter is CEO of Boston Dynamics, a legendary robotics company that over 30 years has created some of the most elegant, dextrous, and simply amazing robots ever built, including the humanoid robot Atlas and the robot dog Spot.

This is a completely different world from large language models. It took 15 years for Boston Dynamics to get its Atlas robot to produce a natural looking walk. This discussion is worth viewing and thinking about. Figuring out how to get a robot to move is at least as intellectually challenging as getting an LLM to produce coherent and sensible prose. One might even argue that it is more challenging. At this point getting LLMs to produce coherent prose is not difficult. Multiple-column multiplication is difficult; eliminating confabulation is difficult; but mere prose production is not. But for some reason we don't know how to calibrate the difficulty of that behavior and so are prone to overvalue the significance of what the LLM is doing. But we are unlikely to view the movements of a humanoid robot and conclude that it's only a hop-skip-and-jump from playing a competent game of basketball.

On predictive control (c. 24:38):

Robert Playter: yeah those things have to run pretty quickly

Lex Fridman: what's the challenge of running things pretty quickly a thousand Hertz of acting and sensing quickly

RP: you know there's a few different layers of that you you want at the lowest level you like to run things typically at around a thousand Hertz which means that you know at each joint of the robot you're measuring position or force and then trying to control your actuator whether it's a hydraulic or electric motor trying to control the force coming out of that actuator and you want to do that really fast something like a thousand Hertz and that means you can't have too much calculation going on at that joint um but that's pretty manageable these days and it's fairly common

and then there's another layer that you're probably calculating you know maybe at 100 Hertz maybe 10 times slower which is now starting to look at the overall body motion and thinking about the the larger physics of of the uh of the robot

and then there's yet another loop that's probably happening a little bit slower which is where you start to bring you know your perception and your vision and things like that and so you need to run all of these Loops sort of simultaneously you do have to manage your your computer time so that you can squeeze in all the calculations you need in real time in a very consistent way

Robots and elder-care

Jason Horowitz, Who Will Take Care of Italy’s Older People? Robots, Maybe. NYTimes, Mar. 25. 2023.

CARPI, Italy — The older woman asked to hear a story.

“An excellent choice,” answered the small robot, reclined like a nonchalant professor atop the classroom’s desk, instructing her to listen closely. She leaned in, her wizened forehead almost touching the smooth plastic head.

“Once upon a time,” the robot began a brief tale, and when it finished asked her what job the protagonist had.

“Shepherd,” Bona Poli, 85, responded meekly. The robot didn’t hear so well. She rose out of her chair and raised her voice. “Shep-herd!” she shouted.

“Fantastic,” the robot said, gesticulating awkwardly. “You have a memory like a steel cage.”

The scene may have the dystopian “what could go wrong?” undertones of science fiction at a moment when both the promise and perils of artificial intelligence are coming into sharper focus. But for the exhausted caregivers at a recent meeting in Carpi, a handsome town in Italy’s most innovative region for elder care, it pointed to a welcome, not-too-distant future when humanoids might help shrinking families share the burden of keeping the Western world’s oldest population stimulated, active and healthy.

Italy's elders:

Robots are already interacting with the old in Japan and have been used in nursing homes in the United States. But in Italy, the prototype is the latest attempt to recreate an echo of the traditional family structure that kept aging Italians at home.

The Italy of popular imagination, where multigenerational families crowd around the table on Sunday and live happily under one roof, is being buffeted by major demographic headwinds.

Low birthrates and the flight of many young adults for economic opportunities abroad has depleted the ranks of potential caregivers. Those left burdened with the care are often women, taking them out of the work force, providing a drag on the economy and, experts say, further shrinking birthrates.

Yet home care remains central to the notion of aging in a country where nursing homes exist but Italians vastly prefer finding ways to keep their old with them.

For decades, Italy avoided a serious reform of its long-term care sector by filling the gap with cheap, and often off-the-books, live-in workers, many from post-Soviet Eastern Europe — and especially Ukraine.

“That’s the long-term care pillar of this country,” said Giovanni Lamura, the director of Italy’s leading socio-economic research center on aging. “Without that, the whole system would collapse.”

There's more at the link.