Bits and Bobs 10/30/23

Real systems are invariably messy.

That implies you need at the very last escape hatches and flexibility, even if you have well-paved cowpaths.

It means "delightfully tangy". A delightful, interesting, and stimulating surprise.

Noun form: piquancy. The verb form is more familiar: "pique" as in "pique my interest"

Piquancies are little sparks of potential for curiosity, and thus for helping the receiver develop a new insight.

It felt almost offensive to me to have my entire workflow was upended without warning.

It took me a frustrating hour or two to figure out how to clear the backlog of notifications.

Now that I'm feeling a bit more calm, let me try to abduct some insights out of it.

There's a difference between a tool's interface and (data) quality.

Interface could mean your API or your UX.

UX = API for a computer to human interface

If you change your interface customers might get frustrated ("why did this change? It worked perfectly fine before and now I have to spend unexpected time figuring it out").

If you improve or degrade quality, e.g. fewer errors or better ranking, then the interface doesn't really change and you don't need to alert anyone.

If you add a feature to the interface users might not notice or be bothered by it.

For APIs this is likely mostly true.

For UX though it's much less likely to be true.

Because we have peripheral vision, anything on screen that changes might catch a user's attention.

So most additions in UX count as interface changes.

Interface changes are harder to hill climb and experiment with because you need to coordinate with your users who likely would prefer things to just not change.

With quality changes you can hill climb easier, which gives you a better self-steering north star.

Of course, the quality can change so much that it's effectively an interface change ("this workflow that used to work now fundamentally doesn't") but generally small quality deltas, or even large quality improvements, are fine.

Of course, you need to have a solid core (e.g. a viable business model that generates net monetizable value).

If you also have some kind of self-hoisting product quality (e.g. a network effect) then you have not only viability but also compounding returns and you might not need to distill much more of a vision.

Knowhow is a kind of intelligence gained from doing.

Knowledge is about things you've learned, e.g. by reading.

Knowhow is gained from the experience on the dance floor.

Knowledge is gained from the balcony.

An insight that someone develops themselves will have way more resonance than an insight that is handed to them by someone else.

Your own situation is easier to feel intuitively... but also harder to reason about holistically.

The view from the dance floor is very visceral, but more obscured than the view from the balcony.

A mental trick: Think about what advice you'd give to a mentee in your situation. Then do that!

Entropy emerges from the fundamental fact there are more ways to move away from a point than toward it.

The coordination headwind shows up because there are way more ways to be misaligned than aligned.

The possible misalignments scales approximately with the square of number of participants, while the possible alignments rise sub-linearly.

The overall quality of a thing is due to its intrinsic qualities, but also external factors (e.g. network effect, its broader ecosystem)

The primary use case of a luxury brand is significantly higher intrinsic quality.

The secondary use case of a luxury brand is signaling to others.

If the brand is well known and coveted, the network effect of that signaling quality can become huge and dominate the primary use case of intrinsic quality.

A common pattern:

A brand starts off being very high quality and develops a reputation for it.

The secondary use case of signaling grows significantly and comes to dominate.

The product creator cuts a few corners in the core product to enable more scale.

This process continues piecemeal over time.

The original valuable proposition (the primary use case) withers away and leaves only superficial quality.

As information about this percolates through the ecosystem (which takes time), the signaling quality also erodes, perhaps explosively.

Once the signaling value has eroded it's impossible to bring it back.

A useful narrative helps make sense of an uncertain, foggy reality.

When a narrative starts to distill, things that strengthen the narrative will be no-brainers to glom onto.

Things that threaten the narrative will be subtly or actively ignored.

The narrative becomes more and more self-coherent... but potentially decoheres from the actual ground truth.

If others only hear the narrative (and don't have any relevant priors), they will likely find it convincing due to this satisfying self-coherence.

But if the counter-narrative evidence is later discovered, it can shatter the narrative.

In some cases your receivers will already have disconfirming background knowledge or priors "Wait a second, aren't you ignoring..."

In that case, the narrative might not ever be adopted in the first place.

1) An openness to the idea that the world is not single-dimensional and black and white, but multidimensional and shades of gray.

Surprising information is a signal that there might be a shade of gray or a whole dimension you were previously missing.

2) When you encounter surprising information, you don't lean back and say, with disappointment, "that's interesting..." but rather lean in and with active curiosity say "Oh, that's interesting!"

The combination gives you self-accelerating momentum to have a nuanced understanding of any problem domain (wisdom), without a ceiling.

The more wisdom you've earned in the past, the more likely you have a lens or tool from previous problems to apply to new problems, meaning the value of this power compounds.

Things like knowledge management tools and workflows can significantly accelerate this compounding loop.

These two skills are very, very hard to teach or get someone to develop from outside. It has to come from intrinsic motivation.

These two skills are also somewhat contextual; you might be "burned out" in one context but still have energy in another.

One way to cope is to pretend you can't see it, to put on blinders and heroically tilt yourself at windmills.

Another bad way to cope is to give up.

But there's another approach: see if you can hold the problem at arm's length, to see it as an intellectual puzzle.

If you focus on it "it's such a deep bummer this is such a systemic problem, it should be different" then it will crush you.

But if you let go of the moral "this should be different" and instead grapple with "like it or not, this is what it is. What am I going to do about it? How am I going to play with this particular puzzle?" you can often find judo moves.

Everything is hard for mundane reasons: for systemic reasons.

Once you see that, it's easy to give up. But if you can get excited about the intellectual potential you can bootstrap.

Every system is nestled within another system.

It's easy to forget this, because your horizon will be constrained by the system you're inside of, and it will feel like the edge of the universe.

If within the system someone goes "aha! I have you trapped!" the other agent can often pop up to the surrounding system and do a move "outside".

An example: In a game of chess, yelling "Checkmate!"... and your opponent kicks the table over and shoves you in frustration.

A more positive example is the heroes in the Matrix being able to do magic because they know they're in the Matrix.

These outside moves will feel completely shocking to those inside the system, because they were totally inconceivable within the system.

It's important to never forget that a move could come from the outer system, scrambling the internal system.

Some of these moves feel inherently unfair or like "cheating", but sometimes they can be extremely powerful judo moves.

It takes time and effort to abduct these out of your experience into little packages that can be lobbed to others.

From your perspective they feel like sharing lovingly-crafted packets of potential.

But if the receiver isn't open to receiving them ("What do I have to learn from this person?" / "I don't have time for this right now") it might feel like getting pelted with pebbles.

Seeds can stay dormant for a long time, and withstand harsh conditions before ultimately blooming.

These little insights will help you be more effective at what you're doing, accelerating your output.

But it's hard to find time to do it.

One way to do it is to mentor someone--you're helping distill out seeds of insight for the other person, but those insights can help you, too!

All of this is turbo-charged when you use a knowledge management system; each little seed of insight you create is captured for the long-term.

If you have a regular process to abduct insights and put them in your knowledge management system, it's like you're mentoring yourself.

You can have shower thoughts without needing the shower, just go on a walk!

The force of gravity is a strong but unchanging force that influences absolutely everything around you.

Because it never changes, you don't get to see what happens without it; it intuitively feels like some background thing you can safely ignore.

But that force of gravity might have huge determinant power if it ever changes!

You can only say a thing doesn't appear to matter if you have seen how the system changes when that input varies.

Forces of gravity can change. And when they do, watch out! All kinds of completely bonkers things will happen that you literally can't understand.

This is a powerful meta-strategy at the core of many strategic judo moves.

Your counterparties might be your users, your peers, or your partners.

The default approach is to identify a set of counterparties and then invest time and effort to land them.

This has some upsides: a coherent vision that the whole team agrees with reduces coordination costs by huge amounts.

But it also has many downsides:

You have to do the work to identify the counterparties.

You might compel their superficial cooperation but not get actual alignment.

You might miss that there are other much more willing participants nearby.

Using the self-selection meta-strategy has many benefits:

The counterparties do the work themselves.

You get intrinsically motivated counterparties.

You can find good counterparties even among entities you didn't expect.

Think of this approach as nurturing seeds of intrinsic motivation that you find, not trying to force it.

Alternatively, think of it as a big butterfly net to catch butterflies.

A few downsides of this approach:

It doesn't look like anything, so (in any organization bigger than some small size) it's hard to get recognized for it.

You can't rush it or control it much.

It might not work! For example, if there aren't enough counterparties in your vicinity, or there isn't any motivation.

When you use this playbook you want to maximize your surface area (a big butterfly net) to counterparties so you're more likely to find one.

This playbook is the fundamental idea behind doorbells in the jungle.

If you're powerful enough, or you push hard enough, you might get what looks like agreement.

But it might be superficial agreement with underlying misalignment.

This can especially happen if you convince someone's boss but the person who has to do the work is not motivated to do it.

It's much better to rely on intrinsic motivation where it exists.

Toil is the tasks that are just barely complex enough that they can't be fully automated, but are also soul-crushingly banal for the humans who have to execute them.

The amount of human ingenuity that is poured into this crank-turning is huge--a force that is everywhere and mundane and hidden.

But what happens when everyone gets this dirt-cheap spackle? The bar of what can be automated will shift.

It's totally conceivable that people everywhere will be empowered to make not even situated "software" but situated solutions: hacky and ugly, but perfectly situated to their particular bespoke niche, empowering their creator to spend their ingenuity on more levered things.

But even if everyone did a ton of this spackle for toil, it's the kind of thing that you might conceivably miss at the society level, because there would be no concentration of a big, obvious, prominent thing... it would just be a tsunami of little mundane hidden improvements everywhere.