Notes on human development (Innovate Program)

I wrote this essay for a TKS director. TKS is a global accelerator that trains students to solve the world's biggest problems. I did both years of the program, Innovate (learned about emerging tech + build projects) and Activate (learned about global problems + build projects). I love helping the new directors (in this case I'm giving advice about how to maximize the exp. for students) so I wrote this essay.

TKS is analogous to a high powered NFL offense. It has a lethal playbook, an intelligent yet swaggerous set of coaches, and young weapons. While I think there are several powerful components of TKS, there are a few, less obvious parts I’ll walk through too.

First, the powerful.

TKS is a program that teaches students about the world’s technologies, mental models, companies, networks in a “give you the resources, you use them to explore” way. It’s structured enough so students have deadlines and something to focus on but the program thrives on curiosity. Getting a student to salivate for more knowledge at the end of a session is the goal.

How do you get kids to want more learning as much as 5 y/o want candy on Halloween?

The first way is to novelize something they’re familiar with. In other words, bring newness to a concept that’s become habitual or mundane to them. Writing is an example. Many students produce english essays in school but rarely are they a) good, or b) an activity they want to pursue outside of class. Showing them an essay by Paul Graham or Balaji Srinivasan and explaining the writing style and process of figuring out ideas through words is a new way of thinking about something old.

Showing students applications of the technology you’re presenting in a session is the second way to get them salivating. CRISPR is cool, but using gene editing to create new plants that provide vitamins that humans didn’t have access to is cooler. If you lead students down a path that makes them curious enough, their imagination will join in and the combination of those two will create ideas you nor the director team considered. Curious students will exhaustively ask what if… and they might find one paper from 1910 that answers their question but this brings them a step closer to producing something valuable.

The third way is an environmental change that, when done correctly, creates a knowledge flywheel where new knowledge is the only substance that feeds the curiosity. When students begin the program they have social media, iMessage group chats, twitter friends who retweet Taylor Switft’s latest album. Those aren’t bad things but they distract curiosity. And, deep curiosity, the one that leads to questions that produce breakthroughs in science and technology is something that should always be on. One can’t have mode on when they’re switching between a 90 minute video on solar panel deployment in Ghana and their friends homecoming post. Sheer will is too difficult and doesn’t counter distraction. Deletion is default. After the distractions are deleted, add Kindle, iBooks, Apple Podcasts. Ask students to download any book and swipe through it when they have 30-60 minute blocks in the day. I’ve found that reading about the world has made me more curious about what I don’t know, creating a positive cycle that keeps me salivating.

Also, keep in mind what apps I didn’t mention: Spotify, Wall Street Journal, Safari. Apple podcasts contain podcasts only, no music. What do you do if you’re on a diet, no one is home, and you see candy in the guest bedroom? Do you enter the room or walk upstairs? Availability is adverse. WSJ, as Balaji describes, is like digesting 30 cookies. News cycles are fast, (some) books are timeless. Reading books is a better investment than daily news checking. TLDR newsletter is a curated daily news source I recommend instead. Safari gives you too many naughty options. Twitter, random questions about the pop star that your friend mentioned, what the espn baskebtall schedule is, etc.

Do everything you can to get students curious and once they’re curious get them building. That’s why focuses are powerful component #2. Focuses are a) an opportunity to train high standards by creating something useful and replicable b) an opportunity to learn how to write well c) a test of understanding.

Before I disaggregate this list of letters there are ways focuses fail. These are critical to know so you can put students back on a learning path. Focuses fail when current students use past students as a model [of desire]. The first way students use past students is to learn what they build, how they wrote their learn, or what their replicates were. We can call this inspirative discovery. I want to get inspired by other AI builders so let me explore their portfolio. In this discovery period, a surprising amount of students will build the same project as the student they explored. The dino game is the most popular BCI project. Students copy previous students. That’s one pattern you’d like to avoid. Have specific examples of what good looks like but do everything you can to steer students away from replicating something that existed in the past. It’s unoriginal and too many new things have been created in emerging tech since those focuses were done.

The second way focuses fail is when they are done at a low standard. Some students treat this as a school project that’s more fun. Some students don’t know what high standards looks like. Some students are inexperienced in writing code so they’ll find a no-code option for their project and it will be terrible.

For the students who don’t know how to code, let them know that they should focus on the fundamentals of [insert language] and when they’re 50% ready, a project should be built. You can do a lot in ML with functions, loops, classes without knowing about search algorithms or object oriented programming. But, code is essential. And, since you’re interacting with students in 1-1s, you can play matchmaker between a student whose programmed for 5 years and a beginner.

The third way focuses fail is through engineering desire (mimesis). Either a student wants a speaking opportunity/the prestige that Ananya Chadha, Izzy Grandic, and other early TKS students have or they want to make money/say they’re working on x project, etc. In the former, they’re replicating a model and think, falsly, that because they did a focus in cellular agriculture they’re going to be Izzy. This is incorrect but popular logic. In the latter, they’ll query Google to see what emerging tech field makes the most money or sounds the coolest. Students will think about what will impress others instead of what they’re curious about. When a student tells you their focus area, ask why 5 times as you would during a root cause analysis. Play a thought experiment with yourself, the director, to discern the bluff from the truth.

Back to the list of letters, I’ll go from A-C. Focus projects, I think, are the most powerful way to train high standards. The model is set up so you increases the difficulty of the project you’re working on but subtly, the bar you have for the quality of project should increase too. The project requirements include a video, article, and project repo or paper. Encourage students to make a video that was 10x better than their last and walk them through what good looks like for a video. For example, does good have deep technical understanding with analogous explanations? Does it have good lighting and setup? My early focus videos were camera-less loom videos. Low standards and because I wasn’t getting feedback the bar remained low until I reached out to Michael.

Writing well (b) is something I care about. My focus on writing well started because Navid tore apart my first learn article like 5 y/o tear apart wrapping paper on Christmas day. My title was gross, the content of my article made little sense, there wasn’t enough technical depth. I used many words to tell the reader very little. I’ve read many books on writing since I wrote my first article two years ago but the best advice I’d give students on their learn or reflection articles is to send it to their director (you) and get feedback. Send the article to alumni who are good writers and ask them for feedback. When you send the message be brief, tell them what you’d like feedback on and your estimate of how long the feedback will take. And after a student recieves the feedback, see if they improve their next article without needing the same piece of feedback. The feedback I recieved was stored in a notion document that I reviewed before I wrote a piece and after I finished the rough draft. I had a checklist of items to review before shipping.

The projects students create in TKS are interesting but if they’re packaged in poor writing, people won’t tear open the box to see what’s inside.

Focuses, done correctly and at a high standard, can give students PhD level expertise. When I tell professors this, they get offended having spent 8 years and $400k on their doctorate degree. They test a students understanding because if they do their create correctly struggle will be one of the words they use to describe it. Fulfilled will be another, but they’ll struggle because it’s hard. They’ll struggle because they have to implement backpropogation from scratch and learn the calculus behind a loss function calculation in order to write the code for it. One way to see if students understand their focus topic is to send them articles/papers that are being released as they complete a focus. Do AI-focused students understand why transformer models are a big deal? Do blockchain-focused students understand crypto oracles and smart contract systems, do they disagree/agree with the concept of a network state? Tests of understanding come after substantial technical knowledge is built, but these questions develop thought and bring new ideas or questions they’ll want to explore. Technical-first, thought leadership second.

Exposure to tools, people, and projects is the third powerful component in TKS. Some students in the engineering design program at Olin College didn’t learn about Figma until they were in their junior year of college. TKS students design projects in Figma when they’re a sophomore in high school.

Most people, students and professionals, couldn’t name the people TKS reviews in a velocity POTW session. How many people know Balaji? How many know Patrick Collison? How many know Alex Karp? Or, Shane Parrish, Tim Ferris, Chamath Palihapitya, etc. TKS students get exposed to the smartest people and their ideas by searching different words on youtube than their friends. Crazy how powerful thumbs are. Shane Parrish and funny tiktok videos are within an optimal letter range of each other.

When you get exposed to people like Patrick Collison, you learn about Stripe. Alex Karp, you learn about Palantir. If you move up a layer of abstraction, you learn about Financial/Legal Services and data/software technologies. I recommend you spend 1-2 hours in session prep rabbitholing through wikipedia pages, playing hopscotch with hyperlinks until you find something interesting. Go to page 8 of Google to find companies outside the SV popularity scope or query for ideas in [insert topic] that failed. If the technical topic of the week is BCI’s, many students will complete the explore, query BCI companies and click on kernel and neuralink. These companies provide valuable insight on the current developments of BCI tech but what students don’t see or query for is also valuable. In some ways, your extra research removes blindspots they have because their searches weren’t as deep as yours. At meta’s keynote, Zuck introduced new research about Meta’s neural interfaces and augmented reality. Quest users can control what they see by moving their thumb to the right/left instead of using their hand and eyes to change what they see. My hypothesis is students view meta as an AR company, not a BCI company. This is true but the research in big tech’s labs are as interesting as the bci companies (kernel, neuralink, etc).

A culture of feedback is the fourth powerful component in TKS. I’ve mentioned the importance of feedback from directors/students in this essay so I won’t repeat too much.

Two more powerful components of TKS, executable ambition and community.

I’ve tried to leave the TKS community as an experiment and try out other places where people gather to talk about the world, technology, philosophy. I haven’t found anything like it and I think there are two reasons why.

a) TKS attracts 18-wheeler type ambition. Many people are Toyota Corolla sedans. Could you imagine being a driver of the Toyota and seeing a large truck about to collide with you? TKS students want to accomplish many big things, when they talk to others it feels like a collision. There are particular regions of the world where ambition is the norm. TKS has created a digital community where ambition is the norm. Anyone in the community has ambition in their pocket, so long as the slack mobile app is downloaded. Getting students to think ambitiously and map out a plan to obtain their ambitious goal is one of the underrated skills of Innovate. How do I get from A to B when the map doesn’t exist? What do I need to do to create the map, what information do I need to know?

b) You can publish half-cooked ideas and have other popcorn makers help you to produce tasty popcorn. I think alumni are willing to publish their ideas in slack or online and be torn apart by the world. I don’t know if this exists in the innovate community. It’s helpful when you or another director publishes half-cooked ideas in the slack channel, people will model this behavior and be more willing to publish raw ideas.

c) Vulnerability in community was high in activate. While our slack channel was as try as Nevada sometimes, the community events we held were disco party like. My recommendation for you, is to post interesting, atypical questions and give your answer first. Then activate others by tagging them. The ones who were activated will tag others and they’ll want to comment. Questions like: What’s the most fulfilling thing you’ve ever done? Whom have you deeply loved? Where do you go when you want to numb the pain?

In the slack community, get people to move beyond the “oh my gosh, i like this too” or “this is xxx”

This is an example of what good doesn’t look like. (Image redacted because of private slack channel. A student posted an article they wrote and two students responded with "thank you [name]."

Why are students saying thank you? Did they learn something from the article that warrants the phrase thank you?

Here are some alternate options for responding:

a) here’s xxx insight i gained from article and i have yyy question going into this weekend’s session because of your writing. [insert potential thank you]

first is sharing an insight and a quesiton which creates a positive feedback loop where shirley responds and gives her perspective. because other students see the length of the thread getting longer, they want to enter the action. they’ll read (or skim) the article and respond in a similar way to the first responder.

i found xxx part on invasive bci’s curious, neuralink recently published yyy resource on their first round of clinical trials and the results are zzz. do you have a perspective on zzz?

b) second brings new research or breakthroughs to the root writing. it a) exposes shirley to a new resource or project she didn’t know about before and is probably something she will mention in the session. b) is something she knew but attached to her current mental model of bci’s gives her something to add to the article/thread. c) something she knew but didn’t add b/c of x reason.

you mentioned bci’s have xxx potential, I think the opposite. yyy resource is why.

c) third is (lightly) contrarian and these are the comments you want students to post. there’s a line between absolute contrarianism (when someone disagrees with everything) and conditional contrarianism (when they disagree with specific ideas or perspective). you don’t want absolute contrarianism, but you do want conditional. i wonder if students in tks have a bias towards contrarianism or want to publically disagree with someone; if they do, double click on the comments they make in a slack channel. ask them questions to poke at their contarian logic. if there’s not enough bias towards contarianism or disagreement, you play devil’s advocate. you don’t have to believe what you say but offer the other side, it will get kids thinking in two dimensions, the for and against case. some students might find out they agree with your take and disagree with shirley’s. the didn’t know how to express their disagreement before but you’ve made it visible.

Rich discussions can happen on slack when the quality of posts and threads reinforce each other.

I’ve been explicit about the powerful components of TKS: curiosity, focuses, exposure, feedback, executable ambition, and community. I embedded mimetic desire and low standards in the section about where focuses fail. I think these are two anti-powerful components.

I hope this essay helped.