Mark Twain had a deep appreciation of Lake Tahoe.

“As it lay there with the shadows of the mountains brilliantly photographed upon its still surface, I thought it must surely be the fairest picture the whole earth affords.”

Macro forecasting is an area where it is easy to be as right as the consensus, but very hard to be more right. This highlights that consensus macro forecasts provide no value—it doesn’t tell you anything everybody else doesn’t know so there is no information advantage.

Read Thinking About Macro by Howard Marks.

See also:

• Nobody grades an economist and they lack a published track record of their predictions to be believed.

A time crystal is a repeating pattern in the time dimensions which breaks symmetry with the law of conservation of energy. Besides the energy used to set up the experiment, the repetitions appear to require no other energy to loop forever.

See also:

• Video of the first recorded space-time crystal.
• Time crystals can be observed in soft matter systems
• Linking time crystals

Remote work seems like a natural extension of death-cult politicians to rail against remote work like they did against masks and vaccinations. Similarly, it seems like remote work should be an essential part of any liberal platform.

Why haven’t we seen either side take it up? What are we to make of this?

There is every evidence that intelligence can be embodied materially. Our intelligence and mind comes from matter. We know the rules that govern matter sufficiently well. We can even build and represent complex systems inside computers.

We tend to think of artificial intelligence as something that needs to be constructed or built.

Human intelligence is just a special case of artificial intelligence—the same rules about matter apply. However, we naturally evolved into intelligence rather than being constructed.

Watch the video with Frank Wilczek

See also:

• AI is usually described as singular, but really it will be a multitude of AIs.
• There is no special ‘power of mind’ that is not governed by physics, but thoughts do create downward causation.

Having ‘good taste’ is difficult to define. Taste is subjective, it is only confirmed by assent from others (from Immanual Kant) a kind of determinate negation.

Taste is also essential for building great products.

In Creative Selection, Ken Kocienda describes what it means at Apple. Taste is three things 1) judgment 2) balance 3) and wholeness.

A refined sense of judgment leads to good decisions repeatedly. Balancing important dimensions and trade-offs means building for people and not building to build. Finally, it’s pleasing as a whole—it all fits together and pleasing not because of a thin veneer, but because design is how it works.

See also

• A cathedral for creation a bazaar for growth explains why organizations have an impact when it comes to ‘good taste’
• Taste is important in the early stages of product development too. A minimum remarkable product is obviously better.
• There are suprisingly few product engineers which comes down to a matter of taste.
• Engineering managers should also expect their engineers to have a refined taste and it can’t be outsourced to designers.

Steve Jobs was quoted as saying “Design is how it works.” Design is not a step in the process to make it look nice after it’s already been built, but an imbuing process that happens as a matter of course. In that way design is substantial and not a gimmick or afterthought.

According to Creative Selection, this was a mantra repeated often during the building of the iPhone. In order for the whole product to be really great, how it worked and every little decision needed to be thoughtfully designed.

See also:

• Taste is the refined sense of judgement and finding balance that produces a pleasing and integrated whole

Because the COVID-19 Delta variant is much more transmissible (translating to a higher R rate of 50-100% higher than a year ago) and mRNA vaccines are 90% effective, we need 90% of the population to be vaccinated to achieve herd immunity. Unfortunately that seems unlikely due to the politicization of the vaccine.

Read Delta Variant - Everything you need to know.

In sales, the difference between hitting your numbers and not has more to do with the volume of your top of the funnel than anything else. Since a third of sales comes from doing nothing at all, simply increasing the amount of prospecting and appointment setting will improve your sales numbers and can make a big difference.

See also:

• Cold Calling Techniques talks about the importance of prospecting and how your current results is actually the culmination of prospecting weeks before.
• Founding Sales

In 2019, Americans spent an average of 55.2 minutes per day commuting. During the COVID-19 pandemic, remote workers have completely eliminated morning commutes which is like a 10% raise (or higher if you are like 10% of Americans that commute two hours per day). The monetary value of saved commuting time would be equivalent to the largest tax cuts for the middle class ever.

See also:

• 45% of jobs can be done remotely although this might increase inequality of remote work
• Two-thirds of remote workers want to continue to work remotely
• Remote work resulted in a 30% increase in hours worked and 20% decrease in productivity

Using codespaces, you can open a web-based VSCode session, open a terminal and install emacs.

sudo add-apt-repository ppa:kelleyk/emacs && sudo apt install emacs27

Then set the option for using the option key as meta in Settings.

Install dependencies for vterm:

sudo apt install cmake libtool libtool-bin

Then clone my .emacs.d:

git clone https://github.com/alexkehayias/emacs.d && mv emacs.d .emacs.d

Then open emacs (which installs packages with use-package):

emacs

Now there is a full emacs with my config and a vterm running in the browser using GitHub Codespaces.

Other tweaks

• Add your ssh key so you can push/pull to a remote git repository.
• Fix M-BKSPC not working like it does in macOS
• Fix magit can’t open the commit dialogue
• Install other OS dependencies needed for writing (e.g. efm-langserver, aspell)
• Go full screen CTRL-CMD-P (there’s still no way of removing the top heading, but it’s close to fully full screen)

Alternatives

It’s possible to run a Linux container and run it through the browser using Kasm (an easier remote desktop kindof thing). It can be customize to launch emacs directly. I haven’t tried it but seems promising for self-hosting and running privately behind tailscale.

Product engineers solve user problems, but why are there so few of them?

As recently as 1999, The Inmates are Running the Asylum told the tech industry that engineers were the reason websites and products were bad. Engineers were not trusted with knowing about users and their problems and yet, they are blamed for poor solutions to solving them.

The way we organize engineering doesn’t help. As companies grow, engineers specialize and are closely managed by project managers and product managers. It’s difficult to build anything significant that requires action across teams and departments.

Altogether that means we lack the trust in engineers to understand users and the experience pool to cultivate a strong product engineering discipline. Status quo preserving behavior will continue to keep engineering as they are, which doesn’t bode well for users.

See also:

• Most companies treat engineering as delivery.
• Startups have a much better shot at building capable product engineers because startups value generalists early
• Automating cooperation might help larger companies resist these tendencies

SQLite doesn’t have a FIND_IN_SET operator to test whether or not a string exists in set value. To work around it, you can do string comparison using LIKE.

However, to handle all cases of a string in a comma joined array of strings (e.g. a collection of tags related to a post aggregated using GROUP_CONCAT) you need to prepare the value by adding preceding and trailing commas.

For example:

select * from (select post_id, group_concat(tags.tag) as alltags from tags group by tags.post_id) where ','||alltags||',' like '%,food,%';

This returns the correct result whether the food tag was first, last, or somewhere in the middle.

In org-roam v2, a new requirement was added for all notes to have an org ID. This has a few notable downsides.

It duplicates unique identifiers. Notes already have a unique ID, the file name which is enforced to be unique by the file system. Changing a file name should be equivalent to changing a unique ID—you probably don’t want to do that (and it’s a sign you should probably delete it and start a new note).

It clutters the note. The preamble alone is a minimum of 4 lines compared to 1 before due to the addition of the :PROPERTIES: drawer.

It hurts portability. Previously, org-roam files used to ‘just work’ in Working Copy, including links to other files. It doesn’t anymore because of ID links. It also further locks the implementation in with org-mode which is lacking parsers to make it more ubiquitous (go-org is a good option, but many more are started and abandoned).

Professional employer organizations rely on co-employment to share responsibilities with their customers. A company enters into an agreement with a PEO where the PEO is responsible for statutory requirements (remitting payroll taxes and compliance with labor laws) and the company acts as the workplace employer that hires, terminates, and supervises. (Interestingly, PEOs also have the power to hire/fire, but it is seldom, if ever, used). Each state recognizes co-employment (to some degree).

Co-employment enables PEOs to file payroll taxes using their FEIN and state ID numbers (except client reporting states). They also can put their customers’ employees on their master health care plans and workers’ compensation insurance.

See also:

• PEOs are an underwriting business and need to manage insurance premiums and pass-through costs.
• PEOs taking care of everything is a myth
• Back office

SUTA dumping is fraudulent arbitrage of State Unemployment Tax where fraud-y PEOs put employees from high insurance rate companies into a new entity with a low insurance rate and pocketed the difference. This was made illegal by the SUTA Dumping Protection Act of 2004 which is why you typically get asked questions when signing up for payroll whether or not you are trying to do this.

PEOs charge between 2-7% of the dollar volume of the payroll for their services (more for GEOs) generating revenue between $1,200 per employee per year up to $4,000 (source: JP Morgan).

See also:

• Back office

Employee leasing was started in the 1960s and is a predecessor to professional employer organizations (PEO). It was used to exploit loopholes in the US tax system. The idea evolved from staffing agencies to include HR and worker safety services until someone coined the phrase PEO to sound more like HMO and clean up the image of employee leasing.

Their image was further cleaned up by the SUTA Dumping Protection Act of 2004 which prevented SUTA dumping.

See also:

• Employer of record
• Back office

A global employment organization (GEO) is an employer of record that is used to hire employees outside of the client’s country. They provide payroll administration, benefits administration, and HR services. As an EOR, they also take on the liability for statutory compliance.

See also:

• Back office

While government tends to be short-term oriented (the election cycle drives decisions so officials can be re-elected), an area the do feel comfortable thinking long term is infrastructure. With regularity, they will initiate projects to build roads, telecommunications, buildings, and other infrastructure that takes many years to build (the second avenue subway line in Manhattan started in 1972).

In the The Clock of the Long Now, they argue that the same infrastructure thinking can be applied nature such as terraforming. Maybe by trying to terraform Mars we can fix the Earth.

See also:

• Longer time horizons are a competitive advantage—applying infrastructure to natural systems might pay off 100 years from now when parts of the globe become uninhabitable due to climate change.
• People are bad at long-term thinking, governments are no different.
• Terraforming is another example of how space exploration and conservation are natural partners.
• Colonizing Mercury is a better prospect than colonizing Mars.

Notion has limited support for counting rows in a table using rollups. Unfortunately rollups don’t allow you to count by any other grouping (like month) out of the box (ala group by in sql). Instead you need to prepare your database by normalizing the data, extracting grouping data using a formula, and then modifying the rollup criteria.

Normalizing the field you want to group by

Let’s say you have a list of tasks that have different types and you want to count how many were completed by month.

You might start with a single database that looks like this (where Task type is a select column).

Now you want to count but a rollup field won’t work. You need normalize the data by changing Task type into a relational column that points to another database.

Now change the Task type column from a select column to a relational column that points to the Task Type database.

To count the number of tasks by type, add a rollup column where the relation is the Tasks database, the property is Task name, and choose count all.

Now there is a count for each task type. However, we need to add a few more columns if we want to count by month.

Extract the month using a formula

Add a formula column named ‘Month’ to the Tasks database to extract the month from the date completed.

if(empty(prop("Date completed")), "", formatDate(prop("Date completed"), "YYYY-MM"))

Now you will see this:

Counting by task type by month

This is the tricky part needed to work around the limitations of rollups. Add another formula column named ‘July 2021’ to the Tasks database that will indicate if the task was completed in that month.

Note: You will need to add a column for each month because rollups don’t allow grouping. Yes it’s very manual, but it works.

prop("Month") == "2021-07"

Now in the Task type database, we can count the number of tasks completed in July by adding rollup column (let’s name it ‘July’) where the relation is the Tasks database, the property is ‘July 2021’, and choose count ‘Checked’.

The result should look like this:

Now we have a count by type by month! You can follow this pattern to add more rollups for each month.

Learning is composed of two elements: surprise and memory. If you go about life and nothing is surprising then you wouldn’t have anything new to learn. Similarly, if you don’t retain the lessons from being surprised then you haven’t learned it.

See also:

• From the The Clock of the Long Now
• There’s always something more to learn. Experience is an illusion of system completeness
• Lecture - Learning to Learn

To summarize the difference in mindset between the two places—San Francisco is living in the future and New York City is living in the present.

San Francisco creates and embraces new technology so much so that it feels backwards when the same things are not present in other places you visit. Scooters, one wheel boards, electric cars, autonomous vehicles, are all part of the scenery.

New York City has a constant stream of things to do. Few plans need to be made and people have a wealth of choices about everything. Collective enjoyment comes from discovering new cool things to do and every other place you visit seems quiet.

See also:

• New York City makes you feel like a different person

Our senses are attuned to things that change quickly (like the lottery), but this causes us to miss what is actually going on (gambling means losing). If one is to cut through this illusion to what’s real, one needs to take the long view.

Not only that, the most important changes take a long time. Take climate change for example—changes in CO2 gases took four decades of measurement before noticing the problem. Similarly it will take decades to counter.

See also:

• From The Clock of the Long Now
• Societies live by decades, civilizations by centuries so solving problems at that threaten civilizations requires much greater long term thinking to solve. A 4 year election cycle or quarterly earnings doesn’t get us there.
• Civilization is composed of fast layers and slow layers to absorb shocks, but some of the greatest problems we face (like climate change) are caused by fast layers disrupting the slow layers.