omenti

creating the arrow tips

To start off you need to subscribe (free) to the Crazy Stupid Tech newsletter from Om Malik and Fred Vogelstein.  A couple of paragraphs from their introduction gives a sense of where they're going:

 

Together we have followed Silicon Valley’s innovation engine for decades. We've seen a lot. But one observation stands out: The best ideas - the ones that launch meaningful companies - need to seem crazy and stupid at first.

We get that tech makes many of us more nervous than hopeful right now. But as Silicon Valley’s influence extends deeper into all corners of our economy and lives, there also have never been more interesting, crazy, stupid new and interesting technologies out there. They are not getting enough attention. So our goal is to tell you about them.

 

I've been involved in a variety of technologies from application to tools, but I'm a bit of an outlier.  My background is fundamental physics - something I've been able to mostly keep up since grad school while doing other things and what consumes most of my time these days.  It's non-linear and the payoffs often don't arrive for decades and rarely for those who did the work.  But over time it's where almost all technologies come from as well and a number of powerful tools.  It's difficult to quantify the payoffs, but most numbers exceed one hundred time return in investment after factoring in inflation - and that includes a large number of paths that didn't work out.  It may come as a surprise, but this is a case where the liberal arts win over time.¹

 

A secondary effect is the training.  Get to the point of a postdoc in physics and you begin to think like a physicist.  The same for chemistry and probably many other areas.  Sometimes this ability to 'think different' is useful in more applied areas - particularly when it comes to stepping outside just a bit.  Of course you probably don't want a room full of physicists, chemists and mathematicians working on finance .. well .. unless you happen to be Jim Simons, but he was probably an outlier. 

 

This kind of research, often called curiosity-driven or discovery research, has been increasingly under fire in the US and EU for about thirty years. Funders are much more interested in some kind of payoff - more of an applied research (although that term is used for something else in industry).  This has happened before and I suspect we'll see a swing the other way as China takes over as the leading scientific power.

 

I'm not discounting applied research.. more of it needs to be done and I suspect a better focus is required.  The enormous sums of money and numbers of people going into computer science for gAI strikes me as mis-targeted and wasteful given more pressing challenges around global warming, inequality in healthcare and so on.

Finding the right mix of people and research areas at the fundamental end demands a great deal of good taste.  A few organizations have been very successful: Cambridge was behind in sorting out issues of telegraphy and the likelihood of radio, but decided to work on fundamental issues of thermodynamics.  They managed to assemble some of the best minds and created the modern research laboratory.  They created the spark that launched quantum mechanics which exploded in the 1920s.  In the teens Caltech was a sleepy little engineering school that found itself with a new mandate.  They decided not to work on  the transmission of electricity like everyone else.  By going deeper they managed to become one of the more relevant pure research organizations for the next century.  There are other examples like Bell Labs which had about six decades.  To be fair it was mostly applied research, but there was a core that changed the world.

 

There's an interesting form of teamwork at these places. People tend to talk to each other a lot.  Places like the Perimeter Institute have blackboards in nearly every room and hallway.  Each table in the cafeteria has notepads and pencils to facilitate the discussions that flow across groups and through the building. Setting direction is an art that is often based on a perception of beauty. It's similar, but slower in the experimental laboratories and there's something of a competition and balancing act between theory and experiment.   It may seem like herding cats and it's messier and more inefficient than the so-called scientific process, but it turns out to be the most efficient way to create possible futures.  It's an act of faith that learning more about the Universe is important to humanity.  Of course it takes time along with the insight of people with different skills to turn that into reality and Crazy Stupid Tech will probably give some insight into that direction.

__________

1 The physical science and math are part of the liberal arts as opposed to professional and technical fields. 

 

 

choking under pressure

There's this story of a student, having just given his thesis defense talk, freezes at the first question with his chalk describing an arc as he faints.  The same story with different details has been told and retold since the 50s.  But there is real pressure.  It's rare, but some students fail their dissertation defense.   Just last week a friend told me about a failure at the University of Michigan. 

 

Temporary choking is much more common.  I've seen it in at least a third of the defenses I've been to.  Usually it's minor and people recognize it happens.  Questions are rephrased or breaks take place and things come back to normal.  It's just part of a rite of passage for many.

 

Choking is more recognized in sport. People talk about choking and 'the yips' - I'll stick with choking here was it's more common and easier to deal with.¹ I've seen it in several events at the Olympic and National and World Championship levels when a lot is on the line. Recent work published in Neuron shines some light on a neurological basis.  The experimenters monitored the motor cortex in the frontal lobe and watched a simple task requiring manual dexterity when the stakes were low, medium and high.   What they found was a big drop in activity as the stakes went up - the subjects suffered a big performance hit.  The effect was non-linear..  subjects didn't perform well at low reward levels - there was a sweet spot.  It's part of becoming an elite athlete to learn how to deal with the pressure, but it still happens on a regular basis. 

 

I've been interested in what the dynamics are in a team sport - specifically the simplest example with only two  athletes on a team.  Beach volleyball is a textbook example as there isn't a lot of overlap in responsibilities and players need to trust each other.  Over time - say an Olympic run - the best teams tend to have stable partnerships. When one player begins to crumble, it's up to the other to shoulder the load somehow and communicate - often with a few words or a bit of body language like a glance. It's an impressive skill to pick up on this in a couple of seconds in the middle of a point and figure out how to deal with it.   Interestingly there's evidence of sexual dimorphism in both style and success dealing with choking.  Sarah summed it up: men fight to bond, women bond to fight.  On average women may be more effective team players.²  Casual observation tells me this applies outside of sport. 

__________

¹ Some use the terms interchangeable, but differ in severity and treatment. Choking often occurs when an athlete over focuses on mechanics under pressure and lose the fluidity, grace, ability to analyze etc, that  they've spent their career building. Their effective skill level falls dramatically.   The yips, on the other hand, take place when anxiety is so high that it becomes difficult to execute even simple actions.  It can be chronic requiring help from specialists.

² For this reason I tend to prefer women's team sports to men's.  Men's sports are more athletic, but the women often show superior team tactics and strategies. 

 

 

(way) beyond black swans

a mnipost

 

Fundamental change comes from discovery - something that science is wired to do.  The process of science is admittedly messy and a bit inefficient, but it works and has vastly expanded our understanding in so many areas. There's this myth about how it works -  the scientific method kids had to memorize in school is a not very accurate model of how things really operated - at least the areas I know. All of the pieces are there, but they're generally non-linear and can spread out over long periods of time.   What drives people are the rare magical moments.  For most of us they're rather small, but the pleasure in discovery is enormous.  It's very hard work as it's so easy to fool yourself.  Some of the most dramatic discoveries come when you're looking for something else - if you're careful enough to recognize them. 

 

One of my favorite examples is a 24 year old astronomy grad student who fundamentally put a dent in astronomy and physics.  Her discovery involved a huge amount of careful work.  Her advisor was awarded the Nobel Prize in Physics while she got nothing -after all - she was a woman and a student in the 60s.  I've been lucky enough to have had dinner with her - she sees her lack of the Nobel underscoring how sexist science and the Nobel committee are.  

 

The interview is excellent and something to show to kids for inspiration as well as dealing with the imposter syndrome.  She's done so much for women in science.

 

airship quadball

a minipost

Every field has its special teachers - those who have figured out how to inspire and raise the level of the students. About five years ago I wrote about Woodie Flowers and his passion for teaching.  One of the most playful and play inspiring teachers I know about (the video on the page is a good intro).  

 

Woodie had a famous course for engineering undergrads.  I know of a few other similar, but different flavored, courses - mostly in mechanical engineering.  ME72 a Caltech is one of these focused play classes and the final exam is a competition.  The most recent exam was Airship Quadball.  Here's a description with a cute linked video.

Some schools set up cross departmental collaborations. One physics dept I know matches physics students with creative writing students and art majors to create science fiction graphic novels.  When it works you need students who have an interest in the other two areas.  The original spark was producing physics students who could communicate more clearly, but it turned out to be fun and educational for each group.

 

 

 

creativity in film and beyond

Unburdened by the plastic wrap so common to art books, it was sitting  in the bookstore and was reminded how much I wanted it.  A copy of The Art of WolfWalkers. Ten minutes of delight, but it was too spendy for this trip.  

 

 WolfWalkers is the third in a set devoted to Irish myth from Cartoon Salon - a tiny studio in Kilkenny, Ireland that has won respect from Pixar, Disney Animation and Studio Ghibli.  It's on Apple TV+ if you like amazing animation. 

 

Animation studios and art schools are amazing places.  Years ago they seemed like black boxes that took stories and turned them into animation.  More recently I've been involved with one and am wobbled by the amount of creativity and work play that goes into the most simple moments.  It's always a team effort spread among disciplines and years.  

 

After a few months I finally bought the book. Just wow!   Here's a video of a roundtable the before the film was released.  It was my introduction to the film (although not to the myth or its historical offshoot as Kilkenny is an ancestral home). 

 

It's worth writing about as (a) some of you may enjoy and even love it and, (b) creativity in the era of generative AI  is something I've been spending time on this past year an animation studio and two universities.  Signals have emerged. 

 

The New Yorker published a must-read piece on AI and Art by Ted Chiang.  Given my experience with animation studios, a music conservatory and physics and math, I'm in conditional agreement.  I do see a place as a tool, even in the creative arts, but more as a tool to relieve drudgery rather than become 'thought'.  I use it to organize references in papers as well as to create simple programs that involve freshman level programming for simple but time-consuming tasks.  I stay far away when precision and accuracy are needed and I don't see the creative spark. 

 

Film studios, musicians, and artists worry about wholesale theft of their work in the short term and the potential, if these tools improve, that we'll see more 'art' of lower caliber largely because human-created art is too expensive.  I think the later isn't possible with generative AI - at least not for quality work.  That doesn't mean real art won't be forced out by economic pressures. 

 

As for schools I worry teenagers may take the easy way out in pursuit of easy grades and free time. They'll be cheating themselves.   The world has no need for more high school essays, but that isn't why they're assigned. On the other hand student can learn about tasks these tools can help them with.  That's going to be more important in college.  I've been on two task forces and some schools are developing solid measures as well as the ability to change with time.  The guiding light is "are you teaching them how to think?"  

 

Clearly we're in a fascinating time.  I think we'll see some real carnage in the industry and use cases before signals we probably haven't thought about emerge.  But in the meantime we have beautiful art.   If you haven't seen WolfWalkers, you're in for a treat!

 

 

a legend

Adm Grace Hopper is of the heroes of early computing.  In addition to making fundamental contributions to software, she was one of the few who started to realize what technological change impacted people.  Recently some of the security folks - notably Bruce Schneier and Steve Bellovin - learned about a general talk she gave at the NSA in 1982.  They hounded the NSA to declassify and publish the talks and here they are!  They give a sense of what the world was like then and you can sense her brilliance.   This was a year before ARPANET cutover to TCP/IP, which many consider the real start of the Internet (although it goes back 13 years earlier in one form or another).  Perhaps there's something you can take away.  

Side story: About 12 years ago I found myself spitting into the wind trying to deal with climate deniers who lacked a technical background.  One of her papers gave a good description of what we now call sealioning. It was a good enough argument to convince me to move on and try other channels.  She also wrote about dealing with people like Trump who use bullshit asymmetry - what is known as Brandolini's law these days.  A very wise woman indeed!

 

 

 

playful active learning

A few months ago I had another in what has become a series of chats with a neurology professor.  Both of us are (well) over 60 and, given the two aging presidential candidates at the time,  the subject of preventing dementia came up.  He noted the standard things - exercise is a powerful tool, avoiding inflammation (you really don't want Covid), etc.  There's a genetic component, but that's more in dispute now as it breaks with twin studies - perhaps there's a bad habit component.  But actively learning is important.  

 

It seems scans of mentally sharp people with the sort of neural damage you see in dementia are able to continue to do well as they've created a web of alternate pathways - it's like the original idea for the Internet - route around the damage. Even the elderly can do this.  The trick is working hard at  something that really forces you learn.  Doing puzzles or reading is fine, but doesn't do the neural expansion. Learning a new language, how to build things, how to do something that's very confusing and hard at first creates a novel wiring overlay.

 

I've been known to plunge into new things for the fun of it.  I'm generally untalented, but love the experience of seeing some progress in an area I knew nothing about. My neurologist friend does some of the same. He likes playing board games so he's trying to create one.  It's given him a lot of respect for those that do it.  It's helpful to try something that might be fun as play is a powerful motivator.  If you like music, learn how to write something.  Try writing poetry.  The hardest part may be finding the time - but if you watch tv, perhaps it's the easiest. Turn on the curiosity.  I'd love to hear the projects some of you have.  It may be worse than what ChatGPT can do, but ChatGPT isn't giving you richer neural pathways. And there may be gains in your current level of creativity. 

 

Just now he sent a video that inspired him..  It's Xyla Foxlin.  She graduated from CWRU about five years ago and makes her living taking on unusual projects - (the rocket powered Christmas Tree is a personal favorite) Here she builds a mobius strip violin.  She admits she's not an instrument maker and goes on to use real luthiers for helping her with the bridge and a few other issues. It's a fun video!  And cheers to her for mentioning Hillary Hahn!

 

the story of the soil

a minipost

My Dad worked with three of the Native American Nations in Montana for much of his life. In the process he became close with the Blackfeet people and was a friend of Chief Earl Old Person.  He told a story of walking with the chief along an experimental area scientists and environmentalists had tired to return to its native state after decades of a near mono-culture dry farming.  The project, overrun with non-native weeds, was a failure. 

 

The experts had carefully identified what wasn't native and laboriously removed it.  Then they brought in once native plants and seeds and expected nature to take over.  Initially the  plants did well, but over a period of a few months they began to fail.  The next year saw them struggling and plants they had eradicated along with non-native weeds began to appear.  Three years out, when Earl and my Dad were walking through the area, it was mostly non-native weeds and a few examples of plants they had removed.

 

Earl explained, with some frustration, this wasn't an isolated case.  Western environmental science 'knew' much more than the Blackfeet who had been in the area for thousands of years.  He paused and offered:

They never asked the soil for its story. 

Years later my Dad learned some soil scientists were brought in and discovered the mono-culture had changed the makeup of the microorganisms in the ground. In fact it would be impossible for the original plants to thrive in the otherwise rich soil.  The original group didn't know the story of the soil.

 

This pretty much sums up my frustration with techno-solutionism.  People develop hammers, see everything as a nail and rush in armed with wads of  money and a singular focus.  But they don't know the story of the soil.

 

 

the correct tools for the job

Every now and again I'm asked to be an outside member on a thesis committee.  Every school I know of requires a member from outside the department and often outside the university.  They're a lot of work, but they can be a lot of fun.  At first you run into the imposter syndrome, but you can ask questions to clarity from the candidate as well as their advisor and even other people in their department.  It's a neat window into an area you didn't know much about.

 

But you're there to bring your expertise and offer constructive criticism

 

About 20 years ago I was on doing this at a nearby Ivy.  The candidate was excited by the work and very articulate, but as I came up to speed it was clear their statistical interpretation was just wrong.  A few questions revealed they didn't know the basics and they should have had at least a year of statistics and a good probability course.  Not knowing how to drive, they used a commercial statistics package.  (the graphs were very pretty) I told them they'd have to do it right or not have it in the thesis.  This turned out to be a problem.  The defense was in three weeks.  Their advisor agreed and the candidate spent the next four or five months learning statistics and reworking the thesis. The final defense was flawless, but I could tell they weren't happy having lost all that time,    Then, a few years ago, I got a call out of the blue.  It was the candidate, now a professor at UCSD.  He thanked me for holding his nose to the grindstone - "it was one of the best lessons I had in college.."  You don't get calls like that often.

 

In the last few weeks I've spent time with some other folks interviewing professors and students to get a sense of modifying the school's honor code as well as general instructions to students and faculty.  It's clear this is an on-going effort as people learn and technology changes, but the experience reminds me of being an outside examiner. It's also clear the purpose of education at this school is to inspire students to follow a path of mastery rather than achievement.  It's pretty clear that a lot of  AI (I use the term loosely as it's mostly a marketing term at this point) is being used without much thought of costs vs rewards.  Sometimes these tools are appropriate and sometimes they're wrong.  Here's an analogy I proposed:

Almost everyone has a microwave oven now. Most of us use them to cook frozen meals, reheat leftovers, boil water, pop popcorn etc and some interest uses like cooking corn on the cob have emerged. They’re extremely useful in these domains. When they were less common - the  period from the 60s through the mid 80s - manufacturers pitched them as general cooking tools that save time. Cook pot roasts, chickens, pies, etc.. there were special tools like temperature probes and browning trays, but the physics just doesn’t work and people generally gave up after a few attempts. Currently AI, in particular GAI, is like the early microwave ovens. It will take time for people to discover the best tasks for it. We’re still at the cooks your whole Thanksgiving dinner - or at least it will next year stage.

Without going into my personal experiences I offer an example of a serendipitous learning when a wrong tool turns out to work.  At least it was serendipitous to me.. the person who pulled it off knew exactly what she was doing.  First find a recording of the Preludio of Bach's E Major Partita for Violin.  It's very famous so there are any number of good recordings - I like Isabelle Faust's interpretation, but your mileage may vary.     After you've listened, try this version by Tobie Miller.   She takes some liberties, but you can tell who wrote it and her playing is lovely.

 

 

improv and duets that skate the edge of the imposter syndrome

I've written about a wonderful set of idea sessions I had at Bell Labs and then AT&T Labs that had a ten year run.  After lunch on Fridays three of us would get together and talk.    Talking with others is important - there's something about organizing ideas into a stream of words for others.  All three of us were somewhat different and somewhat similar.  Two physicists - although from very different areas - and a mathematician. Often a fourth person would join.  Ideas were presented and talked about.   Thinking about these sessions brought out several points:

They were playful and enormous fun. We had a rule that we'd stop after four hours, but that only worked about half the time.

There was no sense of rank although one held a higher position in the company.  

It was allowed - our management didn't seem to mind. Perhaps because it generated a few good ideas.  I believe all of us thought we were getting away with something.

We had to explain ourselves clearly to the others.  Any one of us could easily "snow" the other two and generate the imposter syndrome.  We had a goal of making it clear enough that anyone with our common background would get it.

One of us (Norm Schryer for those who knew him) was an amazing listener and synthesizer.  I've never seen better.  He would listen to the discussion and at some point say "let me see if I understand this" and give synthesis that often was more in depth and clarity than the individual pieces.

The number of participants was small and we knew each other.  We didn't run into the sociological issues that render brainstorming somewhere between useless and counterproductive.  Since then I've come across several studies that look at ideal group size for generating new ideas in fields like math, physics, chemistry and biology...  It's rarely one and usually between two and four, falling fast as the group grows.

Expert visitors would show up by invitation or serendipity.  We all knew each other so it seemed like a natural extension.

Everyone could think on their feet and often in different ways.  It led to mutual respect.  We used quick approximations rather than accurate calculations to keep the flow up.  Often knowing something to a within a factor of two or even an order of magnitude is sufficient. 

We used yes and ...?  rather than well but...  Ideas, even obviously flawed ones, had to go around a few times by adding to them and learning a bit before attacking them.  It wasn't until I took Alan Alda's course on science communication that I realized this is core improv. 

We realized there are limits to the type of question this works for.  You learn that by experience. 

Even though the treats were less than desirable (Norm loved chocolate covered coffee beans), it was great fun.  I can't overstate the importance of fun.

I'm sure some, maybe a lot, of this is generally applicable.  So how to capture this lighting in a bottle?  I've had similar experiences one on one with some people working on specific problems.  That playfulness is central.  At one point I thought of my thesis advisor's requirement of being able to explain what you're doing to a smart high school student.  It's still a good idea and some of us have been successful spreading the idea.  Starting a few years ago videos with an expert explaining something about their speciality to four or five people at various degrees of sophistication.  Some of these are great. I encourage the practice, but the people tend to be great speakers and teachers. 

 

I'm not a gifted explainer or teacher, but I've found a useful variation that touches on some of the points in the old Bell Labs sessions.  When I'm working on something there are three or four friends in very different fields I talk to.  It forces me to communicate clearly enough for them .. the best way to do this is one on one and being able to see them or, if you really know them, hear their voice.  The rule is they have to call out when they can't follow.  As I learned in Norm's group there are limits, but I've seem in work on some rather abstract ideas.  Sometimes the greatest insight comes from chatting with someone with a very different background.  Here's the variation -  the person on the other end has to do the same with me with something they're working on.  Skate close as far as one can, but staying away from where the imposter syndrome causes the ice to shatter.  The basic improv rule of yes and ..?  is important.   I find myself using Norm's "let me see if I understand ..."       When it works you have a duet of two different instruments - it's wonderful!  I suspect this approach might work in a number of domains.