Think ⇄ Experiment

It's taken me awhile to work out the mechanics of my creative process, and I learn more about it each day.

There's three elements, (1) the broader system of "Think ⇄ Experiment", (2) "Think" and (3) "Experiment".

"Think ⇄ Experiment"

When I am doing my best work, there's a natural rhythm between Think ⇄ Experiment. There's defined space for both and I have a clear sense of what mode I am in at each stage.

When I don't have space to effectively Think ⇄ Experiment, I get anxious and very stressed out, even existential, because being able to Think ⇄ Experiment is how I survive, it's how I provide for my family, and contribute to my community.

"Think"

"Think" mode is always triggered by the same event, I get this feeling of fullness in my body and mind that feels like absolute certainty that I have learned so much, but I don't know exactly what I've learned.

This feeling continues to build and build along with its correlated stress and anxiety because it feels like there is so much information inside me that I cannot think straight.

I worry that I've disrupted by natural Think ⇄ Experiment rhythm. Sometimes I worry I've lost my creative energy.

This continues until I cannot take it anymore and decide I'm willing to do anything to stop feeling this way, and that's when the thinking begins. I ask myself, how do I get from the state I'm in, to a new state where I don't feel this way?

I start writing things down, putting thoughts into clusters and themes, making visuals, collecting references, asking people questions, carrying a notebook.

Now, the puzzle starts to come together.

Each day I see things more clearly, until I know exactly what I know, and exactly what I don't know.
And then, a new question pops into my mind.

"Experiment"

"Experiment" mode is always triggered by a question. Sometimes I get interested in questions because everyone around me believes something is true, but I get this feeling that something is "off".

"Is that actually true? How can I validate it?"

"What happens if I do this?"

"How does this actually work?"

More recently, I've been interested in the question: "How malleable is Toronto and what does that mean?"

These questions lead to experiments and prototypes that I create to gather more information. I do these rapidly and across all mediums and facets of life. Over and over again. My little fury of creative activity.

These experiments spill into all areas of my life, they might influence the way I talk, who I hang out with, the clothes and aesthetics I'm interested in, the way I write. I try to push my limits. Do anything to get to the bottom of my questions. Anything to get to the truth.

By now, all of my closest relationships know these patterns. They can tell when a new "Experiment" mode is beginning, "oh god, what's he into now?"

These experiments always have tight feedback loops. I do something and then follow the feedback, then do something again. Sometimes these cycles last months, but most times they last 1-3 years.

The experiments continue until I become confused, and hit that fog again. Until I feel like I have learned so much but I don't know what I have learned.

And then it's time to think again.

my natural think⇄experiment rhythm

Futureland 2 has a really good input mechanism, when I have time I'll wire it up to this website because it'll make it far easier for me to capture things I'm finding interesting when I'm around in my day to day life, for now the stuff I write here are things I think about while researching and working on stuff using my MacBook Pro

I wonder what the ultimate purpose of this page will end up being, there's definitely things I would write here, that I wouldn't write on X. The context is so different. It's more working through my own mind, this is kind of like how it would be if you were sitting beside me at my office while I was working. The thoughts are much more contextual and inspired by what I'm currently working on. I'm not entirely sure who I am writing to either with these posts which is interesting, the audience is extremely small, there's no way to even read all this stuff unless you visit this page, I might consider adding comments to this section, maybe even some kind summarization element, maybe.

hey just testing this out this new input mechanism I'm working on

if you create something and it is mediocre, instead of trying to just make it better, first write about why it is mediocre to accelerate your progress

if something interests you in any way, then try to form an understanding of why

watching my son use tinkerCAD with an instance of chatgpt voice on the screen, and he’s asking the LLM questions as he makes stuff.

Unreal engine learning sprint

If you make it easier for people to do things then it can literally change how they view the world.

- how do I write to an MP?
- what happens when I do?
- are there defined feedback channels for the canadian government?
- is social media a valid and active feedback channel? why or why not?
-which government in the world has the best feedback loops?
- what can be learned from them?

this will be very revealing, but on a fundamental level, what is an MP?

- how do I write to an MP?
- what happens when I do?
- are there defined feedback channels for the canadian government?
- is social media a valid and active feedback channel? why or why not?
-which government in the world has the best feedback loops?
- what can be learned from them?

my current workflow for thinking through all problem sets:

1. Prompt

via audio I ask Claude to "simulate a brain trust of individuals that would give me a unique perspective on [insert long riff on problem set], the individuals must be real individuals, they can be dead or alive, as long as they can give me a unique perspective on what I am working through, I just want to see them riff, I want to observe the discussion, you can be the moderator, but I just want to observe the discussion take place, see the back and forth, the questions, the debates, see them building on each others ideas"

2. Parsing

then I highlight specific things I read and drop it into Arena (
@aredotna
), just like fragments of sentences,

3. Summarizing

then when Claude asks me what should happen next, which it eventually does, I download all of the text excerpts from Arena, and I have a python script that goes through all of it combining everything into a single body of text that I send to the Anthropic API with a prompt that asks it to categorize everything and turn each sentence fragment into its own point, all in markdown.

4. Feedback loop into brain trust

then I take parts of that summarizing that I like and feed it back into the brain trust in Claude and ask them to continue the discussion based on that and repeat the process over again.

In simple terms, what are the primary issues in Canada that people are frustrated with? Is there a way to group them into extremely simple categories?

Is it possible to improve a country by continuously posting about how it is shit?

toronto clothing as an interpretation and expression of SPEED

If a space did exist and it was specifically designed to improve the city around it for its citizens and on all levels, how would it function, and why?

it's extremely useful watching streams of how people work, streams of how they use specific tools, there's little abstractions you can pick up on that can really accelerate your own workflow.

What should and should not be explained?

Robot Wars in Toronto

If the speed of building continues to increase and become more accessible, then architectural thinking becomes more and more important as well. Extremely important. Hm. I need to understand what this means fundamentally and for myself.

A new kind of school

maybe what's emerging here is a net new level of human agency

with these new technologies, feels like we are entering the era of ambitious personal projects.

everyone will push to discover the new state of the art for what's possible by an individual or a group of friends.

environments will need to be redesigned around these findings.

all ways of working and living will need to reorganize themselves against a new status quo for what's possible by someone with a computer and an internet connection.

what is the mental framing of an architect? what are their tools and processes? what do they read and how do they evolve? how does this apply across mediums? what does it mean to me?

that whole thing where people say you are the 5 people you spend the most time with is interesting because I wonder what that means now with LLMs, it’s weird, to think about, is it kind of like saying you are how effectively you use LLMs now?

On a fundamental level, what exactly is Twitter?

if my default is that I won't be saying anything, then it allows me to put even more energy into listening.

Program 01: Agents Notes

Location: 535, east end of Toronto

The general format of Program seems to be, you show up to learn a new skill, and then talk about it.

This is definitely a step in the right direction. Sarim and Tarun did an excellent job with the first one.

It hits on a question I think about a lot:

What are the optimal environments for accelerated learning and why?

A suggestion for something to try:

Instead of running a step by step workshop with space for open questions, try giving the group an objective and a time limit, and let them use LLMs + their own tooling to figure it out with support from leaders who walk around the room answering questions or helping people get unstuck.

I really like the idea of integrating "expert perspectives" at the end to facilitate a broader discussion about what all of this means.

Through that discussion, the participants can take what they have learned and extrapolate on what it might mean for their work, their daily lives, and humanity more broadly.

my obsession with side effects continues, I love the term, I love everything about it

what did I collect today?

Been going on a lot of walks lately, sometimes like 6 hours of walking a day. Adjusting my schedule cause this doesn’t necessarily make sense, but it’s definitely interesting for now.

Is it possible to contribute to the evolution of a large scale system using only a leaf as a reference.

prompt engineering is the process of structuring and instruction that can be interpreted and understood by a generative artificial intelligence model. A prompt is natural language text describing the task that an AI should perform.

A prompt for a text to text language model can be a query such as "what is Fermat's little theorum?", a command such as "write a poem in the style of Edgar Allen Poe about leaves falling", or a longer statement including context, instructions, and conversation history.

Prompt engineering may involve phrasing a query, specifying a style, providing relevant context, or assigning a role to the AI such as "act as a native French speaker"

with LLMs making it easier to write code, personal websites are going to become much more elaborate and complex, giving viewers unique experiences to help them understand and dive into the world of the individual, these websites will have more custom tooling and in some instances become even personal operating systems for those who want to put in more time in creating them.

FAQs have an entirely diff role right now because instead of just providing questions and answers you can provide key context or information and let the user ask questions themselves.

What is the optimal environment for making things given current technologies and planetary needs

New forms don't emerge fully formed - they start as confusion in the old order.

Often the correct question isn't, 'what should I think about this?' but rather 'what is trying to happen here?'

Current theories on Toronto:

can shift the culture of the city by scanning for what’s impossible, then doing it, documenting it, and broadcasting it over and over again.

if a physical space ever changes the fundamental trajectory of the city, it would focus on compressing time, going from idea to physical reality faster than shenzhen, china, what takes 3 months somewhere else, takes 3 days in the space.

too difficult for a bunch of friends to quickly rearchitect a whole city, but might be possible to create the most futuristic 5,000 sq ft in the world.

people in Toronto love asking about the main effects of projects, but very few ask about the side effects or the regenerative effects, it's very rarely investigated in most parts of the Greater Toronto Area except maybe Mississauga.

true wealth is in the regenerative patterns

Thinking through some new tools that I can use with Kai at home, the most obvious new addition is a 3D Printer.

Plan is to intensely focus on some independent projects in new mediums over the holidays with Kai.

I'm so interested by the idea of making a new kind of personal computer for kids, where the centre of the experience is an LLM

the defining theme of the future will be something like SPEED or TIME COMPRESSION

man, the range what's possible with just LLMs and function calling is so crazy to think about, still processing, but feel like it changes a lot of things, especially from a design perspective, yeah, processing..

still need to understand software architecture principles

There's learning building and production building.

Learning Building

- You build to understand
- You build to learn
- You build to feel problems
- You build to grasp possibilities
- You build to speak the language

Production Building

- You build for output
- You build for speed
- You build for scale
- You build for results
- You build for impact

Core: fast
Support: enables speed
Community: at human pace

everything should have some relationship to speed, but not everything needs to be about speed

in the west end of toronto, there's a difference between The Speed Engine and The Ecosystem.

"the 'crazy' is where the leverage is"

i want to form a deeper understanding of formula 1 pit crews

prototype climate solutions in weeks
test new urban systems in days

The end game isn't speed itself - it's what becomes possible when you dramatically compress time scales for innovation

can we solve this in 7 days?

what's the difference between a formula 1 pit crew and a research lab?

speed as a means of
proving what's possible, change people's assumptions, create new capabilities, attract resources, build momentum

You prove it works by creating outputs that are unmistakably different from what people think is possible right now.

And crucially - outputs that are immediately valuable to others, while building toward your bigger vision.

Not easy.

Shenzhen, MIT Media Lab, SpaceX facilities, Toyota City, and the west end of Toronto?

some people want to pop off.
some people have questions.

Toronto’s blank slate quality as a key element in creating a planetary futures laboratory?

architectural sensibility in clothing means treating garments like built structures.

a microcosmic proving ground: small enough to control variables, but complex enough to generate meaningful data

how does rate of creative and technical output increase if you increase levels of integration through a 'minimum viable integration space' to test all layers of integration?

layers of integration =

physical
social
technical
knowledge

Types of integration:

Physical integration
Social integration
Technical integration
Knowledge integration

what is the minimum viable integration

what is your personal operating system? can you define and outline its parameters? do you see all of its components clearly?

what are your systems to acquire sufficient resources?

what are your systems to reduce energy depletion?

experiment with simulating brain trusts in LLMs to navigate specific problem sets

a note from a late night conversation with daniel in manhattan

key ring on end of jacket sleeve, keys dangle from sleeve instead of from belt loop

gentle monster is interesting

published a post today titled, "Why I Love Questions"

internetvin.tv/questions

when a culture is transient it’s puts the core ideas of the place over the people in a lot of ways, people come in and people leave, some of them become the expression of the cultures ideas and some don’t, the culture continues on, 10 people leave and 10 more arrive.

it's so dumb to listen to someone's perspective about something instead of just running an to understand the thing for yourself

quick riff coffee shop notes

i don't like that in a coffee shop everyone is a stranger, would be cool if the people there were somewhat related to my network, with a bit more curation in terms of who's coming in and out, doesn't have to be a lot but some level of filtering where it makes it so there's more trust and less friction for serendipity to occur, a blend of a workspace and a coffee shop, but with some kind of shared value system.

there should still be a clear separation between like deep focused work without disturbances, and casually pulling up and just sitting down for a few hours to get a few things done.

as you walk deeper into the space, there should be more segmentation, maybe 3d printers and CNC machines

there should be an open space that is easily reconfigurable, where you can pull up a bunch of chairs and a lecture, demo or Q&A starts, this should happen in a way that adjusts the energy in the coffee shop while still not making it weird to put on your headphones and continue working, if you're not part of that event, every now and then you can take off your headphones to hear what's being said, and then that might give you quick inspiration or perspective on something

there should be many things around the space, many many things, so many things that it creates the affordance of just browsing, someone might walk in to grab a coffee to just walk around and look at things, someone might take a break from their work to walk around and look at things

I don't like working at coffee shops, but I like having continuous access to coffee, and being around other people working on stuff.

every city is an OS, and it makes sense to have proficiency in every OS relevant to your objectives.

toronto might be the most under-appreciated city in the modern world, but that's changing because a group of people are running experiments, and more accurately discovering its functionality and potential relative to other cities.

"the only luxury is time, the time you spend with your family" - ye

On a fundamental level why exactly is the food in SF not that good?

Sitting at Noisebridge in San Francisco, trying to understand what this city is, coming here was useful

"the vibe was welcoming and smart"

Lunch is an important element of linking up, it's really special in how it can effect a space if everyone is grabbing lunch in the space, I noticed this at Shopify a lot.

Link up for lunch, and there's a natural connection between that and other things like a lecture, or Q and A. It's interesting. hm.

Grabbing lunch is an event.

a space in Toronto that focuses on working, talking and making, would not be complete without a focus on eating.

Things to experiment with here

- integrate a narrator at the top of the page, the blocks are me as a character, some interface component at the top is the narrator.

- would love for things I track on FL2 to be displayed here, would be cool if that was the primary input mechanism to this website

- I want to integrate some kind of LLM based response to what’s entered here, it’s for me to deepen my understanding of myself and also the things that are catching my interest but also open to displaying the back and forth publicly, not sure what the effects of that will be but it’s def publishing in public but doing it with a collaborative partner I guess

- one of the things even subconsciously that I like about Twitter is that I can post something on that site and a response can occur, it’s interesting, but depending on how nuanced and specific my thoughts are it can hit for someone or it can not it depends on the context they are in when they read it and how it resonates or doesn’t resonate with them but if I integrate an LLM here I’m curious what the effects would be on my creative process, I could basically get a consistent response from everything I post if I wanted one and that could be observable. Hm.

need to update how time is handled on this site, will update that soon have a bunch of notes on what to experiment with next should write that down

it’s crazy how fast an insight or idea pops into my mind and then how fast it’s gone, kind of sad

So much of what makes life work for me is just laughing consistently

Would be interesting if there was a mixture of like LLM generated text in line with my own generated text here if there was some kind of running adjacent summary or something, a summary of my day or something, maybe even a back and forth conversation of things to think about. I post something and then there’s some kind of reply that happens that I can view and interact with.

Some labels here might be useful. Maybe depending on project or something not sure yet. I like the idea of that but I don’t actually ever label things myself well, I never stick to actually noting the label. Maybe it’s something an LLM could do automatically if that’s useful in some way to me but hard to say rn

I have a lot of thoughts that are contextual to my process or projects but maybe don’t make sense to post on Twitter (I like using Twitter as a kind of running notes thing) but maybe those notes would make a lot more sense here I don’t know. There’s something really useful to me about writing things and then putting them somewhere public like this. Something about the public dynamic of it improves my process, it helps create more momentum or something.

So I think it would be cool if I could add artifacts from other tools I use here, some that come to mind are:

(1) things I bookmark on Twitter. I could use this as a way to catalog specific things, but also to add my own tweets here if I wanted to.

(2) and maybe documents and other media I save in a specific google drive.

(3) maybe there's other input mechanisms I want to add here, def would love to include some of the things I track on Futureland.

I think it would be really interesting, especially given the dynamics of the current times, if someone created a hardware company that specifically focused on kids. Had an interesting conversation about it with Abtin tonight in Bernal Heights, San Francisco.

I’m experimenting with the are.na api right now. Wondering if this is a way to put real time notes and stuff on my website.

I’m not sure if the input mechanisms of are.na are fast enough for this to work but it’s interesting to think about.