Recently, collaborating on a paper, I was convinced that there was a problem in the algorithm we were presenting, and got together with a colleague to discuss it. He saw some of the problems, but had a different take on others, and kept coming back to a minor point about our use of a method in one step.
As we talked, we slowly realized the problem I was raising and the comment he was making, while seemingly unrelated, were actually two sides of the same coin. A minor tweak in the use of a published algorithm, seemingly made just out of necessity to make a demo work, was actually a key, load-bearing innovation that made everything downstream in the algorithm work.
We made the change, and suddenly everything in the paper started to fall into place.
But we'd never have gotten there if we hadn't taken the time to listen to each other.
-the Centaur
Pictured: Nola, the night of another great conversation with a friend.
SO! I'm back at the ranch after my business trip, back home after being laid off while traveling for work, back in my office slash library where I worked remotely for much of the past two and a half years.
BUT I was surprised at the exhilarating sense of freedom when I stepped back into my library. I spent fifty years collecting these books and a decade of that organizing them into a kind of external brain, a structuring which only really worked once we had this much larger space on the East Coast, but I never fully had time to take advantage of all of this ... this INFORMATION SPACE, because of the pressures of work.
Well, work hasn't come to an end, even though I've been laid off, because as a researcher it's in my best interest to continue the collaborations I had on papers in flight. But now it's just focused on the important stuff, and the rest of my time is focused on much needed improvements. And as I stood in the library, the outcome of years of work organizing - prompted by my wife, who long since adopted "everything has a place" in her studio, and knew the effect it would have - I felt like, "I can DO this now."
Already today, I read up on probability theory, practiced piano, debugged problems with my bass guitar, and, yes, collaborated on an ongoing paper. (I also continued to recuperate from food poisoning, worked on taxes, let the plumber in to fix the plumbing, picked the brain of the landscaper, and took care of house and cat things for my wife, who's having migraines). And I still have hours left in my day.
Who knew getting laid off could be so liberating?
Maybe it was time.
Back to work!
-the Centaur
Pictured: My "work" workstation, and the "information space" behind it.
SO, I was so sick after the food poisoning (and soreness from retching, and the general lack of sleep on the red-eye) that I slept for over thirty hours. I think I may have gotten up in the middle of that for an hour and blogged or something, but I basically crashed from ~10am Tuesday to ~5pm Wednesday. By the evening I was well enough to do a grocery store run and have a light dinner, but I am still wiped out, and I canceled all my meetings and am going back to bed. Here's hoping rest does the trick.
-the Centaur
Pictured: This just popped up "From Three Years Ago" in my photos app. Thank you, Google Photos, for this truly bizarre picture, which I cannot even find again due to the way they clear the recommendations once you look at them (again, thanks, Google Photos, for this truly bizarre anti-user feature). I present to you ... a robot alligator, atop a Coke Zero, with ... a pen for scale? Perhaps this was nine years ago, from the Google Objects project? I really don't recall. Enjoy.
You know, I posted this a few days ago, but didn't post what I thought of it. I do think that it's pretty crummy that some people think that aid should be needs-tested. The rationale, mostly on the right, is that giving out assistance after someone misbehaves creates a "moral hazard", that is, a chance people will exploit the help in an attempt to behave without consequence. And this, of course, is absolutely true - there are some people who do exploit help to allow themselves to avoid taking consequences. Any system that is created can be exploited, and some people live by, literally have as the method of their living, exploiting others.
But someone who's starving and cold on the street is starving and cold on the street, and it's also true that we have an obligation to our fellow human beings to help those who need help. When someone collapses in front of you, you don't know whether that person was on drugs or just had a heart attack: you need to help them. Our human society works because we are social animals who factor the needs of others into our decision-making. (This is as useful for robots as it is for humans, so I suspect this is more than just a human-bleeding-heart thing, and is instead a universal property of successful intelligent civilizations).
And, not surprisingly, I know many rich people who understand this. It isn't a big issue; many members of my family are well off and they didn't need to be told to contribute to the needy or to step up providing food and water when there was a disaster, they just did it.
SO, back to the quote:
What is it about rich people that makes them think they can starve poor people into good behavior?
David Wong
I know the answer to this.
Because that kind of thinking works on them.
I know a lot of people who are affluent. Some from inheritance, some from hard work, some from a combination of the above. Not all of them managed to keep it. The ones that did were frugal - perhaps not all the time, not on everything, but when it came down to it, they were always worried about running out of money, even if they had a lot of it.
They knew if they behaved the way that "the money" wanted them to, they'd have none of it left.
Many rich people behave responsibly only because they fear the consequences of their own misbehavior. And, I think, that's why they're so concerned about "moral hazard:" they know if they didn't have a backstop, they would behave terribly irresponsibly.
Well, fine. Good for them, if that keeps them being responsible.
But that's no excuse to project their experience out to the whole world.
If someone is sick, homeless, or otherwise in need, we should help them. That's what humans do. That's how our civilization has become such an amazing success: we live in a world where life events can sweep away our entire foundation, and when those things happen, we all help each other stand back up.
Food poisoning and a red-eye flight: two great tastes that go great together. Or not. I will remember this one later, in, like, my life and stuff. No more blog for you today.
In the "beyond the last place you'd look for it" department, I found my wife's laptop. It disappeared during her last trip to renovate the old house, and she could not find it, any place she looked. I couldn't find it either, until we had to re-do the floors and I had to move everything out of all the side rooms. Before I put everything back, I staged it into another room and started methodically going through every box. No dice. But then, while moving some of the spare suitcases we'd left here, I noticed one of them was strangely heavy. Huh. What's this I feel in here? Could it be ... a laptop?
So it turns out my wife apparently had used the above small suitcase to transport her laptop and all of its accoutrements (charger, case, etc) but ... perhaps forgot that's where she put it, in all the chaos of moving things from room A to B to paint stuff, only to move it from B to C to fix things, then from C back to A again. Regardless, it was in the wrong room, with the empty suitcases, so it doesn't surprise me that it was hard to find. But I found it ... by methodically searching every place, whether it made sense or not.
Busy working on a revision of a paper for the HRI Workshop in Academia and Industry, so enjoy this picture of our renovations instead. My other task for the day was working on our house out here in California, which we have to fix up if we wish to sell, rent, or even just really live in it. Not much going on in this picture, but earlier today I was crawling all over the floors with waterborne Color Putty, filling gaps in the slightly dodgy wood flooring. The installers left some, um, pretty substantial gaps ...
... but were nice enough to come back for free and spend several hours fixing most of it, leaving me with jars of the product to fill in any gaps we found later. As they explained, the gaps we were seeing were natural to this product and we can only see down to the tongues of the material, but, still, there's a pretty marked difference between the gaps we see on this new flooring and the tightly joined hardwood floors in our new house, or even the damn near hermetic pergo floors in the rest of the original California house, and we don't think "Oh, just don't ever spill anything, ever" is a reasonable answer. So I'm going to go over them carefully before putting our boxed belongings back into the rooms ... one crack at a time.
Oh, joy. Don't get me started on the work I had to do to try to rescue the path beside the house, which nature firmly decided it wants to reclaim ...
-the Centaur
P. S. I promise all this work is necessary, and is not elaborate avoidance behavior of the manuscript, as my subconscious hunts for other things to work on in an attempt to hide my writer's block from myself.
Restarting the numbering a little bit so we're capturing 'blog per day' (day of year: 44 - blog series: 23 = 21 behind). Up late doing various stuff, so here's a neat shot from a parking garage in downtown Berkeley, after I did my traditional artist date "visit a couple of cool bookstores, get some nice food, find a coffeehouse to work on my book." The road in question is just chock full of theaters and other artistic venues, so there's often quite the interesting crowd milling about when I'm heading from the garage to grab some food.
SO! I just turned *AHEM* a year older and decided to go to my favorite restaurant, Nola, in Palo Alto. Nola and I go way back - logically speaking, I must have first gone there in something like summer of 1997, near the end of my internship at SRI (formerly, the Stanford Research Institute) on hierarchical planning. Nola was the first place I ever got "drunk", or more honestly, slightly buzzed from a very powerful margarita, the first on-the-rocks margarita I had ever had. I had even ordered it by accident; normally, at the time, I was drinking the equivalent of watered-down alcohol slushies, and ordered on-the-rocks on accident. My tastes in drink have ... considerably evolved ... since then, though I still stick to the one-drink-per-day limit.
Nola's a truly magical place. I'd put it in a novel ... if I hadn't already. (LIQUID FIRE, as the vampire-friendly restaurant our heroes retreat to after a nasty dustup with some evil firespinners). Look at their decorations for Mardi Gras! Beautiful.
Welp! One more spin around the sun. Plan to keep doing those as long as I am able ...
Another neat little place in downtown Palo Alto. It's amazing how special a place downtown Palo Alto is; for being a part of a vast megalopolis, it's a charming downtown with a small-town feel and a surprisingly connected place. I ran into at least six people I knew in the short time I was down there tonight, and got an introduction to a robotics group at MIT just by sitting in a chair and talking to some friends.
This is kind of the experience I had when I first came out to the Bay as an intern, 25 years ago (more or less); I had just arrived, was hungry, but restaurants were busy, so I took a seat at a restaurant bar, the only space available ... but no sooner had I sat down than I got offered a job.
Well, technically, I sat down and cracked open a very technical book, and the person sitting next to me didn't offer me a job, but did give me their card and let me know their startup was hiring.
A colorful auto-generated effect from Google Photos showing the profusion of succulents in our xeriscaped yard. They've really taken off in the past couple of years, and are quite beautiful even when not blooming.
So, I'm proud to announce my next venture: Logical Robotics, a robot intelligence firm focused on making learning robots work better for people. My research agenda is to combine the latest advances of deep learning with the rich history of classical artificial intelligence, using human-robot interaction research and my years of experience working on products and benchmarking to help robots make a positive impact.
Recent advances in large language model planning, combined with deep learning of robotic skills, have enabled almost magical developments in explainable artificial intelligence, where it is now possible to ask robots to do things in plain language and for the robots to write their own programs to accomplish those goals, building on deep learned skills but reporting results back in plain language. But applying these technologies to real problems will require a deep understanding of both robot performance benchmarks to refine those skills and human psychological studies to evaluate how these systems benefit human users, particularly in the areas of social robotics where robots work in crowds of people.
Logical Robotics will begin accepting new clients in May, after my obligations to my previous employer have come to a close (and I have taken a break after 17 years of work at the Search Engine That Starts With a G). In the meantime, I am available to answer general questions about what we'll be doing; if you're interested, please feel free to drop me a line at via centaur at logicalrobotics.com or take a look at our website.
After almost seventeen years at Google, I've made the difficult decision to get laid off with no warning. :-) Working with Google was an amazing experience, from search to robotics to 3D objects and back to robotics again. We did amazing things and I am proud of all my great colleagues and what we accomplished together.
However, my work in robotics is not done, and I will still be pushing for better robot navigation, large language model planning, and especially social robot navigation and embodied AI. I'm spinning up an independent consulting business and will announce more details on this as it evolves - feel free to reach out directly though!
-the Centaur
P.S. Sorry for the delay - this has been up on my Linkedin forever. But for some reason I just wasn't ready to post this here. Avoidance behavior, however, has gone on long enough. Time to move on.
Pictured: me and Ryan at Sports Page, the traditional hangout you go to on your last day at Google. It was a blast seeing all the friends, thank you for coming!
Image apropos of nothing. Nevertheless, avoidance behavior has gone on long enough ... soon it comes ...
-the Centaur
Pictured: Another shot of the real place in Palo Alto which must have subconsciously inspired the Librarian's Favorite Ramen noodle shop, from an unpublished story.
It's been a difficult few weeks due to "the Kerfluffle" which I hope to blog about shortly (those on my LinkedIn have seen it already) but equally as much from a Stanford extension class I was taking on Deep Reinforcement Learning (XCS234 - speaking as an expert in this area seeking to keep my skills sharp, I can highly recommend it: I definitely learned some things, and according to the graphs, so did my programs).
Finally, that's over, and I have a moment to breathe.
And maybe start blogging again.
-the Centaur
Pictured: A mocha from Red Rock Cafe, excellent as always, and a learning curve from one of my programs from class (details suppressed since we're not supposed to share the assignments).
Been up super late chasing a paper deadline for the past few nights, so no real post for you; instead, enjoy this picture of a photogenic cat.
I do have to say, these past few nights pulling this paper together have proved over and over again that my ethic of "work just a little harder than you have to" pays off, as each night I wanted to quit, and each night I got a major chunk of forward progress done by working just a bit harder.
Now, sometimes that was 4 am, but still ... after spending a great deal of time with what felt like a giant incoherent mess of a paper (17 pages at last count), the exercise of pushing to submit a 6-page excerpt to this conference really crystallized the notion that, yes, we have done something good here, and, yes, if we take the time out to summarize it, it really does add something to our understanding of the problem.
Oh wait. This isn't an inside-the-firewall production. So, AHEM, "... it really does add something to our understanding of SOCIAL ROBOT NAVIGATION." Just to be clear on what we are doing.
... a computer with a fast fricking hard drive. I use a backup scheme in which one older computer has all my file mirroring services on it (Dropbox, Google Drive, etc) and then backs that data up to a local Time Machine backup. But the old iMac I had had long ago reached the point where it couldn't back up to local network storage and needed a directly connected USB drive, and eventually that, too, gave up the ghost, with Google Drive and Dropbox essentially strangling each other to death if you tried to load them simultaneously. In other news, unrelated except for the inexorable passage of time, my personal daily driver laptop had reached the point where half the keys skip and the battery life was down to roughly 1 minute.
SO! I bought a refurbished Apple Silicon MacBook Pro. Even though it is a gently used machine, way cheaper than the most recent models, this M1 Max screamer has downloaded most of Dropbox and a large chunk of Drive without breaking a sweat. Apparently, the larger, faster SSD of a 2022 MacBook Pro beats the heck out of the old spinny hard drive of a 2015 (or is it 2013?) iMac. Who knew? And it can serve as a daily driver until such time as I can afford a top of the line machine, if I even need one if Apple Silicon is as fast as they say.
... still falling behind. But I'm not giving up on getting back on track. The key problem is making sure that time is carved out in the day for blogging, taking walks, taking care of yourself; once you start to let that slip, six months have gone by and you haven't posted. And when you have a TODO list two inches thick, it really is hard to make sure you carve out that time ... it's easy to trick yourself that it's much more important to get the "important" stuff done ... even though, sooner or later, you must do "unimportant" things like laundry. Building new habits is like doing the laundry ... the world turns, and so must the dryer, and if you don't make time for it to get done, you'll be sad when it doesn't.
Call your wife, take a walk, post on your blog. The "important" stuff can wait.
-the Centaur
Pictured: A nice meal, somewhere (Iron Hill Brewery?) featuring fish and chips, a Moscow mule, and a book.
... still behind, but, whatever. Pictured: Left Bank in Menlo Park. Hadn't been there in years, but it was wonderful. Unexpectedly I turned out the other way from the bathroom and discovered a second stair with beautiful mosaic tile, sparkling in the late morning daylight in a way I never recalled seeing before.
Still pictures really can't do it justice, but it still is beautiful.
