quad

Tech notes

Open source was born in response to the fear that corporations were abusing copyright to enclose the creative commons of software. So, a movement of knowledge workers made a collective new deal to share freely the bounty of our intellectual labours. And with copyleft, some went further: we demanded that anyone raised up by standing on our shoulders must offer the same help to all, without bias or restriction.

Open source was born before the modern digital economy, back when the default was still to own things. But with the arrival (or return) of software-as-a-service, corporations like Facebook, Apple, Amazon, Netflix and Google found a loophole. They could enrich themselves without sharing. They could climb up and kick away the ladder.

And now, with the birth of Generative AI, corporations like Midjourney and OpenAI charge a pretty penny for the service of taking from the creative commons of all digital creative expression. If they deign to share anything, it's under rules and licenses friendly to their balance sheets and their bottom line, enclosed by their donations to politicians and their legal threats.

And so now everything is open source, whether we like it or not.

Scott Robinson CC BY-NC-SA 4.0

At least every year, colleagues express surprise and delight at how I give formal written feedback. That is, the kind of feedback that corporate employers request annually. I've meant to write down my process for over a decade; so, better late than never!

Formal written feedback has a few conflicting goals:

  1. Prompt (self-)reflection and conversations with our colleagues
  2. Find out how we're regarded by our manager and peers
  3. Help the company decide who to promote, keep, and let go

But, everywhere I've worked, goals 1 and 2 are stymied by goal 3. It's hard to be genuine when our bosses or human resources are eavesdropping! Is it surprising to hear that, too often, I've seen the most well-meaning of my colleagues write 250 words of filler; just enough to tick the “wrote feedback” box for the year?

Regardless of the status quo, I deeply believe these goals are all worthwhile.

Years ago, the first time a company gave me people to look after, I felt trapped between potentially being a manager from hell or wasting their time and my own. In desperation, I came up with a plan that worked so well that I've used it ever since, in both individual contributor and managerial roles.

  1. Of the intended recipient, I ask what they want “Is there anything in particular you'd like feedback on?”
  2. From that, I write my feedback, kindly and directly addressing their prompt with concrete examples
  3. I show what I’ve written “Here's what I got! [link to a doc]”
  4. I ask them to review it, looking for context and corrections “Am I missing or should I add something? Am I off-base? Anything I should remove? Anywhere I can be more clear?”
  5. I only share what we agree upon

In short, we have a conversation. We collaborate and arrive at a mutually satisfactory result, achieving goals 1 and 2. And goal 3 is still satisfied thanks to the written artifact.

I believe everyone should have a degree of control over their reputation. And I don't believe in saying anything about someone that I haven't already told them.

Thank you for reading this note. I'd love to hear your thoughts on it, publicly or privately!

(🙏🏾 James Adams, Arvin Aminpour, Mike Bain, Marvin Charles, Phoebe Lew, and Valerio Pastro for their feedback on drafts of this!)

I continually see group identity both asserted and embraced in incontrovertible terms. It looks like this: (using myself as an example)

  • “Scott is male”
  • “Scott is American”
  • “Scott is Black”

In my experience, few things are so easily defined. And too often, I see conversations derail over a disagreement on what is what.

I see groups as an emergent construct from a gestalt of opinions. For each group, every person has their own opinion about both (a) every other person's inclusion and (b) their own affinity to it. I find that, upon interrogating these opinions, three structures appear:

  1. Internalised affinity: a group someone includes themselves within
  2. In-group embrace: anyone who includes someone in “their” group
  3. Out-group assertion: anyone who includes someone in a group they exclude themselves from

Examples

“Scott is male”

  1. I see myself as male, though I feel no particular affinity to many context's perspectives on masculinity.
  2. Most men agree I am male, though I am often ostricised from male contexts.
  3. Most non-men agree I'm male, and as such I agreeably self-select myself out of male-exclusionary contexts.

The bounds of each structure is tight. “How” male I am has never been questioned. I consider myself lucky, in this regard.

“Scott is American”

  1. I see myself as American.
  2. Most other Americans agree.
  3. Most non-Americans agree.

The bounds of each structure are less tight. “How” American I am is a rare topic, reserved for detailed-oriented conversations like: “will you still be American after you get an Australian citizenship,” “don't say American because it's offensive to people from the rest of the Americas,” and American white nationalists.

Funny story: once at a comedy gig a comedian asked if there were any Americans in the audience. I raised my hand. The comedian takes one look at me and replies, “wrong kind of American.”

“Scott is Black”

  1. I see myself as Black, if “Black” means a descendant of the African-American diaspora.
  2. Blackness is highly contextual; in many contexts people who see themselves as Black would not see me as so, or at best conditionally embrace me.
  3. Blackness is highly contextual; in many contexts, I'm not considered by non-Blacks as Black.

The bounds of each structure are less tight. “How” Black I am is a common topic, especially as I don't live within the United States. “Blackness” in some contexts is a patrimonial legal construct, where instead I am “white.”

Funny story: a close friend of mine in Taiwan asked me if I thought I was Black, and upon hearing my affirmative response, replies “but you're coffee coloured!”

Another funny story: two men in Ethiopia once stressed that I was Ethiopian— not Black— and that my family had obviously either forgotten or lost our heritage.

How real is a belief?

I've focused on social identity in the above examples. But I find this model useful to gauge the “realness” of things. The tighter the bounds, the more real; the looser the bounds, the less real.

How real are the following items? (taken from 2022-12-19's nytimes.com frontpage)

  • Blackness, and Black pain
  • Catholicism, and how intrinsic is its opposition to abortion
  • Childhood, and if a 16 year old is an adult in the context of crime
  • China, considering Taiwan and Xinjiang
  • Family, and how Christmas gifts build closer ties
  • Israel, and Democracy in a Jewish state

In Practice

This is a tech note, so let's get technical.

The above is an argument that incontrovertible identity can be both convenient and an inaccurate representation of reality.

Consider these two choices:

  1. An Account with:
    • An Email Address field
  2. An Account with:
    • A has-one relationship named “preferred address” to an Email Address
    • A has-many relationship named “addresses” to an Email Addresses

We should prefer option two, despite it being more up-front work to do.

  • Use CalVer for applications
  • Use SemVer for libraries

Why?

  • CalVer promises maintenance; final products (like applications) age (see: app stores)
  • SemVer promises composition; components (like libraries) are parts of a greater whole (see: package managers)

Caveats

  • Absent tooling to enforce it, SemVer is an optimistic promise

ChatGPT is the best chatbot ever created, knowledgeable and intelligent about everything except what you have expertise in.

Gell-Mann Amnesia:

Briefly stated, the Gell-Mann Amnesia effect is as follows. You open the newspaper to an article on some subject you know well. […] You read the article and see the journalist has absolutely no understanding of either the facts or the issues. […] You read with exasperation or amusement the multiple errors in a story, and then turn the page to national or international affairs, and read as if the rest of the newspaper was somehow more accurate about Palestine than the baloney you just read. You turn the page, and forget what you know.

  1. Use a noun, don't use a verb e.g. “installation” instead of “install”

  2. Put it in context e.g “app installation” instead of “installation”

  3. Use a plain word that explains itself e.g. “keyset” instead of “material”

(All credit to Clay Garrett and Undine Rubeze.)