Learning in Public: 8/13/2021

Bits and bytes

Great read. Highly recommend it. Read it for the story of a young LBJ.

Notes on LBJ:

  1. A few false starts in work/business before college. Was obsessed with status and money (having had a taste of both when his father was a respected politician, and later, knowing what it was like to have nothing and be a nobody).

  2. Singular focus (obtain power and the presidency) from a young age (college)

  3. Not especially book smart, but clearly incredibly people smart. Over and over he won the respect and help of women and men in power by acting like their son and becoming so beloved that they would do anything to help him.

  4. He could and would work like crazy to achieve his goals. His motto: If you do everything, you will win.

    • He would work his staff to death. And he knew how to dole out just enough praise for them to continue to love him. Though it sounds like generally, he treated them very poorly.

    • He was a master of leveraging mail (the communication medium of the day) — looking for any reason to send someone in Texas a piece of mail, so they would know he was thinking of them, representing them, helping them (even if not true).

  5. He “re-invented” Texas politics — spending 10x more to win than had ever been done before.

  6. He appears to have no morals — willing to do or say anything to get what he wanted. His actions in WWII, if true, would surely have made him unelectable (dodging military service as long as possible, when he said he would do the opposite pre-war, then going on one flight as an observer yet claiming he had participated in many battles and had really been “in the trenches”.

This book should be required reading in US history classes.

Not worth it. This book covers the next decade of LBJ’s life — when he gets his business interests off the ground through selling his political power to do favors for businesses or secure government rights that no one else could for himself (this is how he created his radio station empire). He went from having tens of thousands of dollars to being a millionaire. Why did he do it? He thought the path to the presidency for him was over, so he instead focused on amassing wealth.


  1. LBJ and Coke R. Stevenson were opponents running for the Senate. When Stevenson would accurately accuse LBJ of something, LBJ would instantly counter by accusing Stevenson of the same thing. This was done in different cities so who said what first was unclear to each audience. Truth didn’t matter and people simply believed what was in the newspapers and on the radio. But the headlines and messages themselves can and were bought — by LBJ. LBJ took this tactic further by vastly outspending Stevenson in buying airtime on the radio — getting his accusations against Stevenson to be the ones heard and known.

  2. LBJ’s team was listening in on Stevenson’s phone calls (via the switchboards). This allowed them the counter Stevenson’s strategies immediately or even pre-empt them — getting their message to the press first.

Modern Poetry:

Binance is a cryptocurrency exchange, headquartered nowhere, that offers highly levered cryptocurrency futures to retail traders. The way it apparently works is that you put in $1,000, and if the price of Bitcoin goes up by 1% you make $1,000, and if the price of Bitcoin goes up by 20% you make $20,000 and think you’re the best trader in the world, but if the price of Bitcoin goes down by 1% you lose all your money.

Source: : https://www.bloomberg.com/opinion/articles/2021-07-12/goldman-doesn-t-want-mercenaries

Matt Levine and his daily Money Stuff newsletter is an absolute delight.

Quick Hits:

  1. 1% of cars sold in the US today have a manual transmission, down from 35% in 1980 (link)

  2. Story of Baseball Advanced Media (BAMTECH) - great thread and fun to look back at their origin and how they honed in on building a very successful tech company (doesn’t hurt to have a monopoly on MLB) and now critically important streaming company. What a steal Disney got in buying this streaming stack — it will take the rest 5-10 years to catch up.

  3. 60% of US consumers bought groceries online in the last 12 months (link)

  4. Average U.S. android user spent 24.5 hours per month on TikTok, compared with 22 hours on YouTube and 17.5 hours on Facebook. That’s ~3 days per month!?!(link)

  5. Theme: Buying a customer -> Microsoft to acquire AT&T’s Network Cloud technology  (so they can run the 5G services on Azure Cloud for AT&T and other TelCos in the future)

  6. Gas station in space (link)

  7. Self-flying gas tank (Drone gas tank for the Air Force)

  8. Videos of Amazon pickers & packers


Learning in Public: 6/29/2021

Bits and bytes

Notes on Crypto

  1. You can earn interest on your crypto holdings?! WTF is happening?

    Everyone is encouraged to never sell Bitcoin. But if I never sell my Bitcoin or stock, how do I use it to buy stuff? The answer is: you get loans against your BTC (assets) as collateral.

    For example, BlockFi will loan you $10k at a 9.75% interest rate if you put up 0.43 BTC ($20,321 at time of writing) of collateral. Meaning, if you don’t pay back your loan, they will begin selling parts of your BTC to pay for your loan.

    Short version: You get this loan because you think the value of BTC is going UP UP UP — and you don’t want to miss out on those gains. Read more here with a detailed example: http://huntermonk.com/2021/02/11/why-so-high.html

  2. There are cryptocurrencies that “pay” you to hold the currency. Uniswap and SushiSwap are examples of automated market makers (AMMs), a type of decentralized exchange (DEX). They enable the exchange of tokens on the Ethereum blockchain (e.g. I want to convert token X into token Y. Uniswap and SushiSwap have different governance and forms of incentivizing users to hold the currency. One “burns” a fraction of a token after each exchange — meaning that given a fixed number of tokens — the value of all remaining tokens should increase. The other, takes the fees from doing the exchange and applies a percentage of them to BUY the token on the open market. Read more here.

  3. Lofty.AI - Invest in fractions of A.I. vetted real estate

    Own fractions of tokenized real estate for only $50 and sell any time. Become an investor in less than 5 minutes.

    Is this real life? Yes.

    For me, this is one of the first uses of the blockchain that makes sense, touches the real world, and transforms an already existing market/asset class. I’m impressed by both the drop in fees AND the mechanism for “instant” liquidity. Further, token holders in this house get their fraction of the rental income distributed on a daily basis and it is the blockchain that makes this viable.

Quick Hits

  • Are Offline Policy Evaluations the new A/B tests?

    • Great article: https://edoconti.medium.com/offline-policy-evaluation-run-fewer-better-a-b-tests-60ce8f93fa15 — gets actually good at the 🔎 OPE methods & examples section.

    • Bill summary:

      • When running a service in which you have significant logging of user and application state — instead of running an A/B test, look back through the logs to find users that fit the test condition you would otherwise A/B test and see what happened. Saves time. Reduces risk of A/B testing something “Bad”.

      • Log everything — keep what is logged highly structured and well documented, and now you have incredible potential to ask and answer questions you hadn’t even thought to consider a priori.

    • Paper: https://arxiv.org/pdf/1811.00260.pdf

  • SaaS Asset Backed Securities

  • Artificial Intelligence is really Artificial Time

    Specifically, modern AI is better understood as AT — “Artificial Time” that can be prosthetically attached to human minds. And highly capable computing systems are best understood as existing in superhistory rather than embodying superintelligence. I think this is genuinely an interesting shift in perspective, not just a fun bit of idle speculation for time nerds like me.

    This is a must-read in understanding AI. Full piece here: https://studio.ribbonfarm.com/p/superhistory-not-superintelligence

  • We Only Ever Talk About the Third Attack on Pearl Harbor

  • Automated container unloading

  • Satellite trails

  • Waymo Self Driving car — truly self-driving. Great look at the state of the tech and what can and did go wrong driving around Phoenix


Learning in Public: 6/22/2021

Bits and bytes

Cloud Kitchen: Virtual Brands

Great article covering the developments in Cloud Kitchens over the last year: https://www.theverge.com/2021/6/1/22456930/chicken-wings-delivery-virtual-brands.

A few notes:

  1. Chicken Wings are the ultimate Cloud Kitchen food. Easy to prepare, high margin, travels + reheats well.

  2. A new breed of asset-light Cloud Kitchen companies have popped up. They push existing restaurants to launch a 2nd, 3rd, and 4th Virtual Restaurant brand.

  3. The various models:

    1. Existing restaurant chain launches a 2nd restaurant brand to sell via the apps (DoorDash, Uber Eats). Examples: Buca di Beppo has a wings brand, Chuck-e-cheese has a pizza brand.

    2. Cloud Kitchen company runs their own kitchens - Cloud Kitchen company has a kitchen and staffs it with their own employees to cook from a variety of their OWN restaurant concepts. The savings here is having just a kitchen — no front-of-house staff, no tables, etc.

    3. Cloud Kitchen company provides kitchen space and placement into the apps - Chefs/Cooks/Restauranters then rent space and give a commission on orders sold via the apps.

    4. Cloud Kitchen company launches a major brand and has orders fulfilled at 100s of random existing restaurants - Example: Mr. Beast Burger. Very powerful when it works — the Cloud Kitchen company now has a direct relationship with the customer, instead of going through the apps.

    5. *NEW* model Cloud Kitchen company partners with existing restaurant, providing one or more Virtual Brands and driving orders via the apps - Value to the existing restaurant is increased utilization via more orders — but at a steep price — literally and figuratively. Literally: ~20% commission. Figuratively: The restaurant has effectively no connection to the end customer. The apps stand between the Cloud Kitchen company and customer — and the restaurant stands one further step removed. While the brand is intended to be memorable, the algorithms on DoorDash and Uber Eats decide who gets seen. In effect, these restaurants will be forced to continually buy their next customer — even if that customer really liked the food last time. This really seems to be the model in which the Cloud Kitchen company has the least skin in the game.

  4. The ordering apps (DoorDash, Uber Eats) have launched their own cloud kitchens — partnering with existing, significant brands AND launching their own restaurant concepts. I believe over the next few years, just as we’ve seen Amazon go after high-value verticals with Private Label brands — we will see the same on the food apps — each launching the most popular food categories AND being able to self-promote in the app to ensure a steady and high rate of orders — thereby maximizing kitchen/employee utilization while keeping customers happy — and in the process out-competing other sole proprietor restaurants in the same food category (e.g. Thai).

Quick Hits

New Pasta Shape: Cascatelli


  1. Cascatelli really does improve sauceability. We had some a couple of weeks ago and I look forward to eating it again. Learn more here: https://www.sfoglini.com/products/sporkful

  2. Theme: The Internet makes it possible to find your 1000 (or in this case 10,000+ fans). As of this writing, they have a 10k+ long order backlist they are slowing fulfilling — and that just covers the orders through the end of March! For all the talk of the creator economy, this is an excellent example of a person pursuing their passion — and creating a story + product that many want. We will see more of this as the concept of a job transforms from a 9-5 to include many more side projects.

Learning in Public: 5/2/2021

Bits and bytes

M-Pesa & the IndiaStack: The Internet Country

Go read: https://tigerfeathers.substack.com/p/the-internet-country

M-Pesa is a mobile phone-based money transfer service, banking, payments, and microfinancing serving consumers in Kenya, Tanzania, Mozambique, DRC, Lesotho, Ghana, Egypt, Afghanistan, and South Africa. M-Pesa offered banking in countries where there was none. And now it is the standard — as in just like China has the SuperApp of WeChat, M-Pesa could be considered the super infrastructure/app for the countries above. M-Pesa leapfrogged traditional/physical banks. As the article above details, Kenya had mobile person-to-person money transfers via cell phone in 2007. The US didn’t get that until 2009 — and STILL today, there is no defacto answer — Venmo? Cash App? Bank to Bank? I highlight M-Pesa for the extreme value they have brought to their users/countries.


  • In 2009, only 17% of Indian adults possessed a bank account.

  • By 2018, almost 80% of Indian adults possessed a bank account.

How? The IndiaStack:

Just as the US Government funded the highways, India in 2009 started building an Identity Layer, a Payments Layer, and now a proposed Data Layer.

For the identity layer, ~400M Indians in 2009 had no id — as in no Social Security Number equivalent. And today there are 1.27 billion Aadhaar (identity) cards. More importantly, these cards brought down the cost of KYC to pennies from tens of dollars — hence enabling the payment layer.

Again, read the essay above, but the interesting bit is the payments layer. The government of India made a common piece of core infrastructure — the National Payments Corporation of India — through which all digital payments transact. This ensures that new companies can join and participate — and that no one company can get a stranglehold on the payments infrastructure — the way that the Credit Card companies have here in the US — with a ~2% tax on transactions. It also means sending money from person A to person B — at any bank connected to the NPCI will just work.

We need more boring infrastructure projects like this. What are the data/internet projects the US Government is currently funding that will have a significant impact 50 years from now? What are the platforms and infrastructure being built? I ask this honestly and with no snark. Most of what I read about has to do with revitalizing the ailing infrastructure of years past.

Lying for Money by Dan Davies

Recommendation: Only read if this is your type of thing. I read it a month ago and don’t really remember much — which doesn’t speak too highly.


  • Frauds basically prey on the greedy, the desperate, people who don’t know what they are doing.

  • All fraud falls into one of four categories: “long firm,” counterfeiting, control fraud, and market crimes.

    • This is probably the most interesting bit from the book. Just go look up the 4 types and you are done.

Quick Hits:

Learning in Public: 3/24/2021

Interesting bits and bytes

Share Links and Learnings

Not Born Yesterday by Hugo Mercier

I do not recommend this book.


  1. People are really hard to convince / sway on issues where they are knowledgable AND OR where they stand to benefit/lose significantly.

  2. Most often, when we look for an explanation we are post hoc explaining / rationalizing. So blaming FaceBook, Trump, people being gullible — really that is just us looking for the “easy” explanation.

    1. And this is why this book was so unsatisfying. It gave the complex answer — when, what I wanted was the easy answer.

Quick Hits:

FUN: Rally car pit crew fix wrecked car in 30 minutes

It is always fun to watch masters at work.

The Suez Canal

Ever Given ship info: https://www.vesselfinder.com/vessels/EVER-GIVEN-IMO-9811000-MMSI-353136000

Photo source

FUN: Mars helicopter releases from the bottom?!

I knew there was a helicopter/drone — but I had assumed it would be released from the top.

Artist of the week: Andrew WK


Loading more posts…