Info Diet offers a peek into the personally curated feeds and media habits of the people shaping the future. In each installment, a different builder spends two days chronicling everything they read, follow, listen to, and watch in order to stay ahead of what’s next. This time: Amjad Masad, cofounder and CEO of the online coding platform Replit. A software engineer by training, Amjad was born and raised in Jordan and now lives in Palo Alto, CA. His mission is to make programming more accessible: “I think it’s absurd that while software is eating the world, very few people know how to make it.”
8 a.m.: I wake up, make my usual triple-shot espresso, and check my phone to ensure nothing is burning. (All good.) I open the read-later app Pocket and scroll through my saved articles. There are a few about AlphaCode, the new AI program from DeepMind that writes code at a competitive level. What I’ve read in the past has made it seem like DeepMind could have solved artificial general intelligence — which is impossible. That would mean the end of the world as we know it. I decide I need to go to the original source, so I open DeepMind’s 73-page paper and read the most relevant parts. The results are impressive, but there’s still a brute-force aspect of their architecture that makes me doubt the AI is actually “thinking” through the problem.
3 p.m.: I get home from a gym-and-sauna session and I’m starving. (I’ve been fasting since yesterday.) I air-fry myself a burger and scroll Twitter while I eat. I find a discussion about web3, as one often does these days, and an article someone shared, called “When the Stagnation Goes Virtual.” It’s published on Palladium, which I’ve never heard of before. The article is really well-written; it uses the metaverse-and-NFT wave to set up a more general critique of our generation. The claim is that we’re abandoning the physical world in favor of the virtual, which is no different than prior generations finding escapism through drugs. I’m not sure I agree with the author’s point, but it does make me think.
I click around the rest of the site and I’m impressed, so I subscribe to Palladium’s newsletter. Two other newsletters I’ve been enjoying are Astral Codex Ten, a psychiatrist’s musings on reasoning, philosophy, science, AI, and other topics; and The Diff, which focuses on tech and finance.
9 p.m.: After dinner, I head to our home office to read more. As my morning reading probably makes clear, I’ve been interested in (OK, maybe even a bit obsessed with) AI recently. I find the transformer-based language models fascinating, mainly when applied to code. This is something we’re starting to use at Replit: We integrated an AI that explains a piece of code in plain English.
I’ve learned to trust my instincts when something piques my interest. To that end, I picked up the quintessential AI textbook, Artificial Intelligence: A Modern Approach. But it was too thick. I realized I needed something more practical and to the point. I’ve heard good things about fast.ai, an online video course that focuses on the massively popular machine-learning Python library PyTorch. I begin watching the first lesson. It’s about an hour long. In the middle, I get distracted by a group chat with my friends. Group chats have been a great way to discuss current events in a safe space, away from the social media mobs. Then, back to fast.ai. I finish up the lesson and call it a night.
8 a.m.: Once I’m out of bed and have my espresso in hand, I try to read something on my phone. But my son runs toward me, looking to play. I still manage to fit in some Twitter scrolling before my 9 a.m. meeting. Nothing too interesting today.
I’m a fan of Twitter lists. I named my favorite one “High Signal.” One person on it is Paul Graham, the founder of Y Combinator (and a good friend), who always has new, noteworthy ideas. Another is Shane Parrish, the podcaster and blogger behind Farnam Street.
9 p.m.: I had nonstop meetings and minimal downtime today. No media consumed, unless Zooms and emails count. But now I’m done for the night, so I settle in and read by the fireplace.
Lately, in addition to AI, I’ve also been interested in the mind. How do our brains work? Where does consciousness come from? Right now, I’m reading Surfing Uncertainty by the philosopher Andy Clark, a book that lies right at the intersection of my interests. In it, Clark uses modern machine learning research to explain various aspects of cognition. Clark’s main thesis is that the brain does only one thing: prediction. (Incidentally, this is what artificial neural networks do as well.) So his claim is that consciousness is an emergent property of the brain trying to “predict” the sensory input it’s getting. It’s a cool, persuasive theory. But I can only get through a couple of pages before my mind is fried. I’m ready to relax and not do any deep thinking.
10:30 p.m.: I scroll through Twitter mindlessly while I wind down. I try to go to bed at exactly the same time every night. It’s the best way to get a good night’s sleep, I’ve found. Once again, nothing too interesting, so I head to YouTube. I watch a couple of Joe Rogan clips on Powerful JRE, a channel that curates highlights from the podcast.
When the clock strikes 11, I put my phone down and (try to) turn off my mind.
Get Amjad’s Info Diet
- “Competition-Level Code Generation with AlphaCode,” by DeepMind
- “When the Stagnation Goes Virtual,” by Ginevra Davis (Palladium)
- Astral Codex Ten, by Dr. Scott Alexander (Substack)
- The Diff, by Byrne Hobart and Jack Wiseman (Substack)
- Surfing Uncertainty, by Andy Clark
- Powerful JRE (Youtube)