I have not done Zone2 workouts since I used to d triathalons. More than a decade ago? What happened to the time….
I started doing zone-2 heartrate based workouts for all my ‘easy’ workouts the last few weeks. My hard workouts have suddenly gotten very easy and outputs that used to be a struggle are suddenly sustainable.
The bottleneck for my performance apparently had something to do with my energy supply and was limited by my ability to use/oxidize fat into energy. Doing training that improved this bottleneck lead to an improvement in performance. Training harder would never have done that since it would engage the bottleneck without improvement.
Often we can not improve something by doing more of it. Focusing on one aspect of what we do can make us blind to what we can do to improve it. This is especially true when we engage with complex systems. All systems have capacity constraints. Focusing on the final output is not offten illustrative of these constraints.
I stumbled back into zone-2 by allowing myself to engage in biking in a different way. Just suggesting you allow yourself the same opportunities.
I have heard and read several people saying you don’t want to bet against Zuck. I have no doubt he is brilliant, knowledgeable, and capable. I’m not betting against him, but here is why I might.
Facebook has had an empirical innovation deficient. They react to competitor innovation and do so effectively. Their performance management and culture efficiently motivate and focus large group of employees on specific outcomes. I also believe this ability of myopic targeting contributes to their social/political issues and lack of creativity.
Zuck has attributed current issues: TikTok, Apple, and the AR/VR future. TikTok is taking attention/users, Apple is reducing tracking/attribution and impacts ads, and the future medium is AR/VR requiring investment to be a key/dominant player. Meta is very profitable and can be so for the next decade without any transition. Zuck does not want to be the founder/CEO of IBM.
The previous transition to mobile, which was also deemed uncertain, was riding a wave of technology that Facebook was not developing. Users were transitioning to mobile. Parts of the world were mobile-first. Apple and Google were developing the new human/computer interface. Facebook ported onto them. Meta wants to own the next human/computer inference platform instead of transitioning to it. The investment is heavy, and the current capability is higher than any existing product suggests.
The competition for this space is already Microsoft, Apple, and Google, plus all the gaming companies that are meta-verse adjacent. Developing and iterating in uncertain spaces has not been a Facebook/Meta core competency. I do not see the first-mover advantage for the company with the significant investment. My current perspective is the future requires a large amount of innovation and discovery, and the company is terrific on short-term, concrete accomplishments. In my estimation, Apple and Microsoft have more abilities in this area.
However, Meta is great at fast-follow. Current investment in technology and employees allows the company to do its best with any new and successful innovations: copy. There is still the issue of the developers, another area that Facebook has not excelled at.
I am certainly not betting on Zuck, but also one I wouldn’t bet against him. It would be amazing to see him pull it off.
Meta’s earnings report resulted in a 20% drop in its stock price. I recently noticed, not being a very active Facebook user, that when I look at my feed I get an ad in the first post, then in the 4th post, then every 3 posts after that. I am also getting a number of red badges on the watch tab trying to get me to get engaged. That has an ad every two results in my feed. At a high level, it seems like Facebook is saying, “Don’t like seeing what you’re friends and family are posting? Here are the most interesting things on Facebook.”
I imagine the more time I spend in the apps, by interactions or attention, a better understanding of what engages me forms. I also imagine that increases their ability at targeting ads towards me. Right now that’s pretty poor. For months I’ve only seen Remarkable2 ads (which I clicked on one back in Sept and never purchased) and online education courses (which I have not purchased for years). I feel so misunderstood.
This imagination leads me to consider that the more time I spend on Facebook will lead to more ads, the ads should be better positioned, and my expected click rate increases. This draws a direct line between time/attention/engagement and Meta’s bottom line.
The ATA changes have affected the ability to target ads, competition from TikTok and others has captured attention and production, and the company has divided attention between launching/growing copies of competitive products and developing the next paradigm of human/computer interfaces. This is happening while in a competition for talent with Apple and Google, both beating expectations with stock price increases.
It seems to me that there is a potential for a downward spiral. I am very interested in what happens next.
The Bark Gift Shop Ltd. case from HBR has an example of a CFO setting sales targets for stores with a bonus incentive based on exceeding targets. I got curious about the optimal policy a manager would use in this scheme. I had to make some assumptions about how increased effort reduced the marginal increase in sales, how the market grew, and what managers valued. I made it simple and said they maximized bonus /effort. According to my python script, managers would not uniformly exert extra effort if targets grew faster than the market. The optimal policy was to throw some quarters away and exert extra efforts in others. This seems consistent with known issues around hurdle and ratchet effects. Despite the intention to motivate store managers to increase sales, the result was managers increasing compensation.
I know very little about inflation. Today I am only writing to get some thinking out so that it can be a base on which to build. My current knowledge can be summarized by, “Too much money chasing too few goods”.
My current thinking is that companies like Amazon and TSMC making capital investments that are larger than several years of previous investments will increase supply. US consumers spending more on services with goods purchasing starting to decline will reduce demand on constrained supplies. I suspect that some of the inflation fears will be abated.
On the other hand, there are more jobs than people looking for jobs, and corporate profits are stronger than before the pandemic. Wages don’t drop over time. Some company’s profits have come from rising prices more than rising costs. Both of these observations seem to suggest sustained costs/inflation.
My current forecast is that inflation will be higher than the previous decade going forwards, but not as large as current CPI estimates and not are large as feared. I suspect the recent sell-offs are overreactions, and there are buying opportunities.
But I am well aware that I don’t know what I’m talking about, and I’m not placing money on the forecast.
I was attending the first session of an AWS certification course. The instructor was giving some examples of AWS tooling and said something that I found surprising. The gist of the statement was that AWS Lamba + Cloud9 is just like Google Colabs. To be clear – I am not calling out the instructor. I am very happy with the first sessions. Just as someone who has used Notebooks extensively for analytics, I found the statement absurd. It was clear that many of the workflows I have used would not operate (well) in the model the instructor was describing. If the model is writing and testing functions in an interactive way, I can see the correctness of the statement (Real SWEs might disagree with me – this is fine). If the model is generating analytics from data, I do not see the similarity.
Instead of this being a tooling question, but had to do with evaluating a business, the instructor would have missed what differentiates Colab from AWS Lambda + Cloud9. On one dimension they are very similar, while on another dimension they are almost incomparable. The market and the opportunity would be evaluated differently. It has me wondering how many places my perspective is leaving me blind to differences that matter. It also has me wondering how many places my perspective is letting me see what others do not.
My favorite part from my Quantum Field Theory class from grad school was the idea that particles are excitations of a quantum field. Take your body, and all the cells in your body, and all the molecules in those cells, and all the atoms in those molecules, and all those electrons in those atoms, and know that they are all electron waves of some universal electron ocean. That your electrons, and my electrons, and the Sun’s electrons are all a part of the same thing.
I have an intent in my mind to move my fingers and type this sentence. There were electrons that moved that hit keys, cascading more electrons that traveled to a server that saved data to a database that is written on a disk. You’re reading a screen powered by moving electrons, hitting eyes that propagate the information to your brain with electronic motion. And all these different things, at different times, that we construct as distinct. They are all part of the same ocean.
It was a great class – though electronic oceans were not on the final.
Chamath Palihapitiya said a few months ago that the cheapest inflation hedge would have been to go long on Google and short on Netflix. At the time I heard this on the All In Pod, I believe Google was around $2900 and Netflix was at $680. Today, after Netflix earning and interest rate hikes, they are around $2680 and $510, respectively. A equal long/short would net about 17% return in about 2 months. I will be watching where this goes over the next few months.
I have had a few conversations and listened to a few podcasts where some people are confused by the crypto-markets dropping along with stocks in response to interest rates and inflation. I’m not an expert, but I am willing to play one today.
My mental model for this is simple – capital asset pricing model (CAPM) / efficient frontier + a belief that institutions invest in crypto. Institutions investing in crypto, I have strongly suspected, resulted in price increases of all coins. If these institutions have a target expected return, CAPM says that each investor set has a mix of risk-free returns and the market (including crypto) to get that return at the lowest possible risk. As risk-free rates go up, the slope of the CAPM line reduces, and the needed risk for a target return reduces. This drives asset allocations from riskier assets to risk-free and less risky assets. Hence the drop in demand for crypto by institutions and large price drops.
No idea if this is a true story. It is just my story for now.
I started this section on the blog because I was inspired. There are several Twitter threads I have seen about writing and everyone one of them has mentioned that the great writers were consistent. I just want to get less bad. I suspect practice still applies.
The action-initiating inspiration was from Seth Godin’s podcast Akimbo. The episode “Paul has a practice” has Seth’s take about “The Beatles: Get Back” documentary shows examples about how the habits, dynamics, and personalities lead to a practice that resulted in one of their albums and best songs they created. Seth calls this out as an example of practice facilitating the transition from bad to good. Seth also calls out how the constraints and the situations, which were engineered, enabled the practice.
This is what inspired me. I needed a constraint to facilitate my practice. And I needed to give myself permission to do it badly. We all learn the same way – We start out doing badly, then we get better with practice.
My partner introduced me to the “Huberman Lab” podcast this weekend. The first episode I listened to is about the science of setting and achieving goals. It goes into great detail the neurology and psychology of perception, feelings, stop/go actions, and value assignment impacting our ability to meet any goal. This was timely since goals are something I think about often since I see them get set and fail so often. I am currently at a failure point with my weight goal (which is only a small part of my health goals).
His discussion of value and stop-and-go actions was particularly resonant with me. I have found that stop-actions are the most important for weight gaols. My experience is that in a given moment we have multiple goals along with our current perceptions. Meanwhile, we have a single value function mapped onto our action space. When I have cut weight I will find myself in a state of boredom, notice I’m hungry and want to eat. My immediate value function is minimizing my boredom and discomfort. I have, in these moments, force myself to imagine a future where I do not meet my goal, where I get grief from my friends for not meeting the goal I told them about, and suddenly I value not eating more than eating. When I have failures of imagination, I find myself eating. Personally, I have found that imagining the downsides increases compliance more than imagining the benefits. This is more universal than me.
My response to a consistent failure in making progress in a goal is to identify a major failure point, make a concrete behavior change, and make success and failure as clear as possible. A major failure point for me with cutting weight is the time between my son’s bedtime and my bedtime. I find myself snacking during this time. My crystal clear behavior change is no eating after 6:30pm with my last meal being veggie & protein-based. I also hang a post-it of my goal weight on the fridge and pantry. It is very easy to know if I am being compliant, and the visual cues help my imagination stay active. It has been effective short term. I will follow up with the long-term results later.
The google data science blog had a recent post, “Uncertainties: Statistical, Representational, Interventional.”
Statistical Uncertainty is the fact that the true value may be different from what is estimated. Representational Uncertainty is the fact that what you measure way not fully represent what ultimately want to impact. Interventional Uncertainty is that the full impact is often different from what you measure.
I think a great example of these ideas has to do with health goals. There are multiple dimensions of health. The broadest being mental and physical. Losing weight can improve both and it is common for people to choose weight as a goal. Weight is representing health, but it is not capable of fully representing health.
Weight is a highly volatile metric since it depends on what and how much we consume. I know my weight throughout the day can vary 3lbs – 5lbs, but I do not view myself as being heavier or lighter in terms of health based on this variation. Drivers of the statistical uncertainty are from the scale (usually small), from recent consumption, and from water retention.
The main driver of short-term weight loss is calorie deficits. This can be done through a combination of reduced consumption and or increased exercise. All combinations lead to an increased cognitive load from a new behavior and can affect recovery, sleep, and mood.
The intervention of diet and exercise to meet the goal of better health as measured by weight has a negative impact on how we feel during the intervention. Probably why so many people quit – it sucks and it is unclear if it is working.
I read Moxie’s “My First impressions of web 3.” Everything he described in his experience in building a dApps and NFT matched my experience and intuition from setting up geth to exploring analytics on the entire Etherum blockchain without relying on the “centralized” platforms.
Aside – I was surprised by the lack of privacy. ENS + Twitter + Network Analysis = Mindblown.
Moxie’s take about centralization early on being a bad sign about decentralization is interesting to me. I agree in broad strokes, He also highlights the path dependant nature. I wonder if this is a permanent change in the equilibrium position of the internet.
The internet started with the idea of being completely decentralized with protocols. It was centralized in universities that had infrastructure that made it easy to participate. Large companies like CompuServe, Prodigy, and AOL emerged as centralized points in the consumer experience. There was a time when Netscape was pretty much the only way to “surf the web” from points like Yahoo, AltVista, and Ask Jeeves. Web1, as it is called now, was very centralized.
The decentralized nature of the protocols made it possible for any job being done by a company to be done by someone else. Partly supported by new technologies. Partly supported by new framing, changing markets, or new business models.
I strongly suspect that web3 will be the same. The early centralization will happen with the enthusiasm to participate. The decentralized nature of the protocols and the acceleration of technology and business models will lead to increased opportunity/competition with every early job done by the initial cohort of centralized web3 companies. Plus as the market grows, the jobs that need to be done change.
I agree with Moxie’s take. I also think that equilibrium has shifted. There is opportunity is surfing the web3.
I had a virtual doctor’s appointment scheduled at the same time I potentially had to go from California to Arizona. When I mentioned this to the doctor, the doctor told me that I could not be treated if I was in Arizona. The license only allowed treatment in California. A few inches, feet, or miles do not make a meaningful difference in digital communication. The construct of a state or a license, however, affected both of our decisions. California is not real in the way the Arizona desert is real. Yo do not see the boundary in real life. It is an idea, a story, we all agree to act as real.
I was listening to a podcast where a question was asked by the host. I immediately had a thought and a follow up fantasy response. It was something immature. I did not identify with the thought. I started to think that is something teenage Bryan would have thought. This lead to musing about how young me is still in there. Wherever there is. I have a construct of Bryan and I have a physical brain. That brain produced thoughts, but some of thought are observed by Bryan and labeled as Bryan’s thoughts, and others are not. I am sure its not arbitrary labeling. Still, it sure seems that Bryan is only as real as California.
Several of my MBA courses covered the HBR case “Trouble at Tessei”. This case covered the before of an organization turn-around.
The company responsible for cleaning the trains is in a downward spiral of employee mistakes, customer complaints, high turnover, and injuries. The average experience of employees is declining as a larger fraction of employees are part-time. Teruo Yabo is the protagonist who is appointed to the organization around.
What Teruo Yabe did was recently pointed out to me. There were organizational changes made that did not significantly increase operational costs. He took “cleaners”, a low stats vocation, and elevated their position.
Their uniforms became bright red so the employees stood out in the station. They were now responsible for hospitality in the station, which included cleaning. Their interactions with the riders/customers increased. The very demanding cleaning of the trains was branded as the “7-minute miracle”. Employee recognition became constant and commonplace. Suggestions from the employees for improvement were implemented. All of the training and support for employees were operationalized in alignment with these ideas.
The scope of work increased, but the position of it changed how the employees and customers saw the employees. This increased pride and responsibility. It increased engagement with the work along with the outcomes of the business.
Yabe improved the experience of working, and the work done became better. Yabe helped elevate the work, and the work felt more meaningful. He did it with attention and care, not cost.
I typically do a long peloton ride (iambrybry – incase you want to fact check me) on Saturday morning. Most recently I aim to do the Saturday morning power zone ride + old rides until I hit 90 mins to 120 minutes. I usually just do the routine as describe and adjust to the high, middle, or low end of the prescribed zones based on my sleep and activity throughout the previous week. Today I decided just to stick to a zone two heart rate instead.
Half-way through the 90 minute ride I noticed I was not sitting in the seat. My butt was in the seat. If you asked me last week, while sitting in the identically position, I would have told you I was in the sitting. Today it was clear as day that I was not sitting despite the seat supporting me. I sunk into the seat. It was not just supporting me. It was my support. I found I could pedal in a slight different way. I turned up the resistance so I could pedal 80-90 to practice this new way of pedaling. My output was the same but my heartrate dropped 10 p points. After a while I decided to go back to 100 cadence. Heart rate was still 7 bmp lower than the start of my workout with the same cadence and output.
If I would have stuck to the goal – do the workout as prescribe – I would not have noticed this. I would not have observed something new. I would not have given myself the space to improve in a new way. I would have achieved a goal without improvement. It is a dangerous trap.
A class I am taking highlighted an experiment from the ’70s done by Lepper et. al. where they experimented with pre-school students. They observed the students and their propensity to choose to engage in an activity, drawing with magic markers, then divided them into three groups:
- No Treatment
- An unexpected award after engaging in the activity
- Telling the students they would receive an award if they engaged in the activity
After the treatment, it was observed that the first two groups played with magic markers just as often as before the treatment. The group that was enticed with the reward engaged with magic marker significantly less after the treatment.
Lepper wanted to build evidence for the overjustification effect. From the paper:
In a self-perception analysis, this outcome is simply the result of a self-directed inference process. In the low-justification conditions, the subject infers from his behavior and the lack of apparent external pressure that he must have wished to act as he did; while in the high-justification conditions, the subject infers that his behavior was determined by the external pressures in the situation.
The Wikipedia entry on this highlights a number of counterpoints and perspectives to the narrative presented in the study. The effect is true. It has been reproduced repeatedly in several contexts. The mechanisms are still unclear to me.
It reminds me that we have to be careful with each other. If we try to change someone, we just might. Just not in the way to expect.
Get new content delivered directly to your inbox.