Subscriber Update
Hello! It’s been a long time since my last post. I wanted to give a quick update to you all on what I’ve been up to, what’s next, and what to expect from Machine Yearning moving forward.
What’s Happened
In January, I wrapped up a multi-year stint working with Andrew Ng and a team of talented builders at his venture studio, AI Fund. Together, we built 9 companies across various domains, including Workhelix, a GenAI readiness assessment platform for executives. We developed and shipped AI-powered products into many different verticals before and after ChatGPT’s explosive release.
In that time, I learned that:
My skills and expertise were best utilized collaborating with teams building foundational tech like infrastructure, agentic systems, developer tools, and frontier AI research.
I believe we’re still too early to crown winners in the application layer, as there are many product categories founders are shoehorning AI into that won’t even exist three years from now.
Just as Uber’s success relied on “existential technologies” like mapping systems and cheap mobile bandwidth, AI and its supporting systems are existential technologies that will spawn entirely new categories. “Electricity,” as Andrew puts it. We still have a lot of work to do building out the power grid.
After meeting tons of founders, judging hackathons, and even participating in a few, I co-founded an angel fund called SHACK15 Ventures with my partners Jørn Lyseggen, Bogdan Cristei, and Ashwin Ravichandran. Operating out of the eponymous social club SHACK15 atop San Francisco’s iconic Ferry Building, we’re working with a vibrant community of investors, entrepreneurs, and hackers, including Cerebral Valley, known for their exceptional AI hackathons. We’ve already made our first few investments and will have more to share in the near future.
To stay updated on Cerebral Valley, subscribe to their lu.ma page.
Machine Yearning
I started this blog back in 2021 as an attempt to organize my thoughts on the emerging AI industry. As a product manager at the time, I wanted to ask better questions to my machine learning engineers, and more thoughtfully build great AI products from the ground up. I believed that only a deep, bottoms-up understanding of what we were building could deliver that.
This experiment yielded some promising results:
I wrote about large language models and transformer architectures before they hit the mainstream. The lessons I learned helped position our team at Spiketrap for an acquisition by reddit, two months before ChatGPT's release.
Last year, I wrote about platform risk and “sherlocking”, and we're beginning to see those warnings play out for some of GenAI’s first movers.
That said, there’s still a lot to explore. As we enter what some are calling an LLM-powered “creativity apocalypse”, it’s probably good practice to curate a whitelisted corner of genuine conversation.
Obviously, I'm using LLMs as creative writing assistants, just like everyone else. Just count the number of times you see 'delve' in a sentence.
What’s Next
People who know me know I like to launch trial balloons into conversation – ideas shared with enough conviction to provoke reactions from those with more expertise. It’s also the Bayesian thing to do, publishing priors and updating them in real time.
Machine Yearning will be switching to a bi-weekly format, sharing unfinished thoughts and observations. These pieces may feel raw at times, but that’s actually the point.
If you’d like to get in touch, DM me on Twitter or LinkedIn, and follow what we’re up to at shack15.ventures.