OSAMA RADI

Brief one today because midterms are sucking the soul out of me. Woke up early to cram for my econ midterm at 11:35, went really well. Econ is such a cool subject because you are trying to use math to model human psychology. The crazy part is that it works; economics is scarily good at making predictions. Initially, I thought that this was slightly dehumanizing. Are we really that special if a couple equations can predict what we do and when? But now, I realize that it's just a collection of cool little shared quirks between us all; that on aggregate we all share similar values, and that is very human.

After my econ midterm, I hopped on a call with a Computer Architecture engineer at NVIDIA. We nerded out about chips, hardware, and optimizing hardware to support machine learning architecture. Super fun chat. What surprised me most is the striking amount of inefficiency left in NVIDIA's computer architecture. For such a huge and significant organization, there is still tons of room for optimization.

The coolest part of my day was having dinner with Nicholas Chirls, founder of Asylum VC, and super cool and down to earth guy. Favorite VC chat so far. I'll summarize what we spoke about here. First, he said to build where nobody cares right now. He was one of the first investors in Coinbase, back in 2013 when nobody knew what crypto even was. He told me that the long-term successful companies in any wave of hype were built 3 years before anyone knew about it. He spoke about looking for markets which are adjacent to what is big right now, but aren't in the spotlight. He gave two examples, organoid development (adjacent to computer cell simulations), and AI specific hardware (adjacent to huge AI boom). I then asked him where that leaves application based companies, like Lovable or Cursor, to which he replied that they had 0 enterprise value. I paused, thinking he would laugh, or that he was being sarcastic, but he was dead serious. I challenged this idea, how can something have 0 value if it's bringing in 100m ARR and is growing faster than anything we've ever seen? He replied that the fastest way to grow was to sell a dollar for 80 cents, but that doesn't mean that it's a sustainable or valuable business. Cursor and Lovable are both unprofitable, because all of their profits are going to OpenAI. Super eye opening take, and refreshing to hear in a landscape where every VC is asking how your startup is AI oriented. He also gave me a very nice hat, which other VCs have not done.

I then did two PSets in two hours. Not the move.

Alright, that's it for today. The lesson of the day is to work where no one cares right now, but you believe that in 2 or 3 years will be the space to be.