Hello Everyone!
I’m back in the free world post-quarantine and enjoying the ability to walk on streets ands stare at trees again. I luckily caught the tail end of the cricket season here in Sydney too, so that’s my excuse for experimenting a bit with the format this week. I’m thinking of alternating between essays like last week and content like today, but let me know what you think and we’ll refine it as we go! With that, let’s hop in:
Reads
The Purpose of Technology | “If the proximate purpose of technology is to reduce scarcity, the ultimate purpose of technology is to eliminate mortality.”
A quick 5-minute read, but one that really zooms out to delineate why we innovate and how we could accelerate it. The crux of it revolves around how we’ve historically defined a breakthrough: making something faster, cheaper, and easier, essentially reducing scarcity. But these are incremental steps - we only care about reducing scarcity because we have limited time. Expand the time available (“eliminate mortality”) and everything becomes cheaper - scarcity is systematically eliminated. The author, Balaji Srinivasan, goes on to discuss how people building the future should prioritize evangelizing their progress to inspire others too - to encourage it, he’s been working on a newsletter that pays you to learn and he pretty much broke Twitter with this podcast he did with Tim Ferriss.
What happens when all the giants stand on your shoulders? | Big Tech is rushing to copy a drop-in audio chatroom startup
Clubhouse has been extremely buzzy these past few months. The one year old startup’s concept is quite simple: groups start a live audio chat room, start talking about a particular topic, and other people enter and leave as they please. Audience members then choose to either listen in or engage with questions. Much of Clubhouse’s success has hinged around the authenticity benefits of going “direct to audience”. By having unedited conversations (versus traditional podcasts or interviews) and encouraging participation, prominent figures suddenly became more accessible (Elon Musk did a clubhouse, for example). But now every tech company under the sun is copying them with their own “live, drop-in, audio social chat” (the latest list stands at Facebook, LinkedIn, Twitter, Spotify, Slack, Discord, and Reddit), which isn’t necessarily a bad thing. What’s peculiar about this is we usually see platforms emerge then unbundle over time, but Clubhouse is being unbundled while it’s being built (Ex. Spotify’s approach will be for sports talk only). Entirely possible this is a lockdown powered trend that’ll fizzle out with too many chefs in the kitchen, but I hope it leads to the emergence of some innovative audio experiences.
Margin Stacking and the cost of AI | Let’s build an AI camera
By running through the various decisions involved in designing and building an “AI camera”, this article does a phenomenal job of painting the picture of today’s AI landscape. It touches on some core trends in the industry such as large companies building huge advantages by internalizing AI chip design, a refocus on chip manufacturing, and the prohibitive hardware costs for smaller companies trying to build AI into their products.
CB Insights announces their annual Top 100 AI Startups list | Healthcare and Transportation dominate again, but Machine Learning Operations is quickly rising
Every year, CB Insights identifies the top 100 most promising AI startups based on factors such as patent data, market size, traction, and sentiment analysis from news/social media. Healthcare and Transportation focused startups dominated again, but in light of the increasing importance around scaling Machine Learning (MLOps), CB Insights included a new category: Computing, AI processors, and AI deployment. The energy around MLOps is reflective of many Machine Learning projects moving out of experimentation and into the murky world of figuring out deployment, scaling, and ongoing monitoring without breaking the bank. To get a better picture of the space, here’s a handy graph that paints the picture around categories and companies pretty vividly.
Watch
The Surprising Habits of Original Thinkers | Adam Grant
Well worth the 15 minutes, Adam Grant breaks down 3 surprising habits he’s observed in original thinkers (He wrote a book about it too). The most striking was the following graph on procrastination - it turns out there’s an optimum level of procrastination to maximize creativity. While rushing too early or waiting too long can impede creativity, “optimal procrastination” engages the subconscious to think non-linearly and more creatively.
Startup of the Week
EdTech | OnDeck Raises $20M to scale its community based education platform
While MOOCs (Massive Open Online Courses) like Coursera are incredible for unlocking information, they're marred with low completion rates because they lack the community-based learning we're used to in school. Naturally, many companies tried variants of this community-based approach to EdTech, but they too hit ceilings quickly. It turns out scaling top-tier education and communities is really hard because as more people join, the quality of interactions decreases.
OnDeck has a pretty cool solution to this problem - they not only create learning cohorts for their programs, but tightly link the cohorts to feed off of each other and build a broader ecosystem. For example, OnDeck founders give exclusive deal flow to OnDeck investors. OnDeck investors then invest in OnDeck founders. And OnDeck creators can write about both, continually accelerating the flywheel. This lets OnDeck scale while still maintaining optimally sized groups, provided they're deliberate on what they're teaching. They call this concept the "Cross community flywheel", allowing graduates to get a head start to hit escape velocity in whatever direction they choose. Here's what the community flywheel looks like:
The combinatorial advantage of linking cohorts is massive. By giving students immediately actionable applications in the course, it unlocks a wide range of serendipitous opportunities they would’ve potentially missed. This also ensures expansion doesn’t degrade the community’s quality - by capping cohort sizes and building tight connections between them, OnDeck gets to have its cake and eat it too. Maintaining this quality will get harder as they grow, but they seem to have the adaptive building blocks in place to quickly adjust and test new ideas.
Thanks again for reading! As always, please feel free to hit reply with any feedback, especially around areas that could be better.
Just like last time, here’s another one of my favorite Pacific Northwest gems, Lake Serene:
See you all next time!
Aqil
Great inside into MOOCs. Technology has always focussed on scaling products but it seems that the traditional element of exclusivity is necessary here.