Inspired #16 | Agents, Virtual World Business, Minecraft, Nvidia, Yanagisawa Sho, The Refragmentation, Sam Altman on Choosing Projects/Creating Value/Finding Purpose, etc.
This text was originally written in Korean on January 19, 2025, and has been translated into English and uploaded
“Unless you’re sure what you want to do and where the leading center for it is, your best bet is probably to try living in several places when you‘re young. You can never tell what message a city sends till you live there, or even whether it still sends one.”
This is a quote from Paul Graham’s 2008 essay, Cities and Ambition. It’s about how, in your 20s, you get to choose the country and city you live in, and how that decision can shape your life in big ways. You should explore different places to find what fits you best, and your 20s might be the prime—or even the only—time to do it.
I recently took a short trip to San Francisco, and it strengthened my determination to move there long-term, or at least to the U.S. It hit me that I need to go beyond just being comfortable with English—I need to think and express myself fluently in it. I’ve always read English newsletters and books almost daily and occasionally listened to English YouTube videos or podcasts. But lately, I’ve bumped up my English exposure to about 90% of what I read, watch, and listen to. I even use AI summaries in English and try to think and write in English first before translating back to Korean.
AI Agents Will Dominate 2025! Here’s Why & How!
AI agents and cryptocurrency have huge potential. When agents handle economic activities online, they can earn crypto based on their contributions. There are open-source projects out there building agents tailored for crypto, like Virtuals and ElizaOS. Story Protocol, for instance, hired an AI agent named Luna, built on Virtuals, to manage an X account—it’s pulling in a yearly salary of 365,000 USDC, which is kind of crazy! This year, expect more agents to pop up online, creating content, chatting with people and other agents, and dipping into areas like e-commerce, stock markets, and online gaming. This is just the start… and it’s pretty exciting.
Virtual Worlds Mean Real Business: How Games Power the Future
[Two Cents #70] State of AI Opportunities — 2025년을 맞으며
New tools like real-time 3D world generation are opening doors we couldn’t have imagined, from gaming to the (now fading) metaverse buzzword. This multi-modal growth in different directions is set to spark fresh opportunities over the next 2-3 years. At first, I figured multi-modal LLMs for creative work would just become a feature on other platforms. But with use cases showing 100x the impact of older methods, I think it’s going to carve out a real market—B2C or B2B, doesn’t matter. It’ll also keep growing as a key piece of new use cases, like game dev tools or real-time 3D virtual spaces, and the long-term potential here is massive.
I see real-time 3D world tech as a game-changer in the bigger shift of internet content formats:
Internet trend: Text (Twitter) → Images (Instagram) → Video (YouTube, TikTok) → 3D (???)
I’ve had this idea for a while, and lately a16z has been echoing it. That’s why I’m keeping a close eye on the “AI + gaming/social/content/entertainment” space and thinking about building something there.
New AI video model, full motion control, add ANY character to video, new text to audio, 3D heads
The 3D Trajectory Master gives users precise control over video movements, letting them swap characters or backgrounds while preserving other elements.
Tango Flux generates high-quality audio from text input, delivering faster and more accurate results than similar tools.
AI tools Dora and Gen HMR create detailed 3D models and accurate pose estimates, transforming how video content is created and edited.
How Minecraft Was Made
There’s a cool video breaking down how Minecraft came to be. Recently, its founder dropped a mysterious hint on X about working on Minecraft 2.
Nvidia’s Nvision for AI
Nvidia’s AI roadmap is laid out clearly:
Perception AI: Understanding our world (2012–2020)
Generative AI: Creating content (2021–2023)
Agentic AI: Taking autonomous action (2025)
Physical AI: Manipulating the real world (the future)
Why Are Huge Tech Companies Getting Into the Book Business?
This piece calls out AI for eating into artists’ and writers’ earnings, labeling it a shady business model. But it’s not just AI—platforms have always leaned toward becoming players themselves. Take E-Mart: they launched No Brand to compete with their own suppliers. Today’s House built its own interior brand, Layer, and Musinsa has Standard. Platforms often dive into their own markets to grow or make more money.
So if Spotify starts pushing AI-made music or Netflix churns out AI-generated series for better margins, it wouldn’t shock me. The catch with AI is that it’s trained on existing creators’ work, and instead of growing the pie, it’s taking their slice. That’s why something like Story Protocol might be a must-have for the internet.
Yanagisawa Sho: The One-Take Ad Master Behind Pocari and Shiseido
“Great stories come from a universal structure paired with a fresh wow moment. You set a solid frame and hook people with the wow.”
Does mixing a bit of new into the familiar work as a universal rule—whether it’s architecture, animation, products, or customer service?
The Refragmentation
One advantage of being old is that you can see change happen in your lifetime. A lot of the change I've seen is fragmentation. US politics is much more polarized than it used to be. Culturally we have ever less common ground. The creative class flocks to a handful of happy cities, abandoning the rest. And increasing economic inequality means the spread between rich and poor is growing too. I'd like to propose a hypothesis: that all these trends are instances of the same phenomenon. And moreover, that the cause is not some force that's pulling us apart, but rather the erosion of forces that had been pushing us together.
Higher productivity means more small companies, since you can do more with less. New tech amplifies productivity gaps between people, which widens wealth divides. That naturally leads to all kinds of splits in society. People cheer the good ones and get mad about the bad ones. Graham says instead of hating fragmentation and longing for the unity of World War II or big-corp days, we should focus on fixing its downsides.
Elon Musk and Alice Weidel (AfD) Interview, Full Translation: Okhotsk Magazine
I’ve had this talk with lots of people in Israel. They’d ask when this approach ever worked, and I’d point to a big example. The Treaty of Versailles after World War I was brutally unfair to Germany and bred massive resentment. It slammed Germany with extreme reparations, paving the way for World War II. Without Versailles, Hitler wouldn’t have risen. But we learned from it. After World War II, when Germany and Japan lost, the U.S. stepped in to help them rebuild with financial aid. Now they’re allies. That’s a huge, fundamental lesson from history.
This shows what happens when incentives are poorly designed.
A System of Agents brings Service-as-Software to life
If AI replaces jobs, seat-based pricing won’t make much money. An outcome-based business model makes more sense.
I’ve been thinking that outcome-based systems could spread to different parts of the economy. Back in the day, companies paid employees by the hour, week, or month, and freelancers or contractors got project-based fees. That setup always raised the question, “Are they worth this much?”—leading to attempts to measure work quality and quantity through hours, performance, reviews, and so on.
Quantifying human work digitally is tricky, but AI agents run on computers and log their actions, making it easy to measure. So as agents grow, I bet we’ll see more outcome-based trends.
The next big thing in 2025 will be…
This post gathers predictions for 2025 from folks across industries. The AI stuff isn’t much different from what we heard 1-2 years ago, but 2025 feels like the year it’ll really hit as tangible results.
Sam Altman on Choosing Projects, Creating Value, and Finding Purpose
It’s way too easy to get lost in work and miss what you’re actually accomplishing. I’ve heard this from so many people: ‘I thought I was crushing it, churning out stuff for 10 years, but looking back, it was totally the wrong path.’ That hits me hard every time.
You’ve got to focus, sure, but after a stretch, step back and ask, ‘Am I on the right track?’ Don’t fall into the trap of those typical 21-22-year-old Silicon Valley dropouts who swear they’ll fix the world in three months, then bail to the next thing when it doesn’t pan out. That’s a mess too.
Finding that balance is tough, but checking in regularly on whether you’re doing meaningful work—and over the right timeframe—is critical.
To avoid long-term missteps, you’ve got to periodically confirm you’re on the right path—otherwise, it’s easy to drift. Our mental energy is limited, so it’s got to go toward what matters.
I finished Siddhartha last week. It clicked that enlightenment doesn’t come from just hearing or seeing—it’s through doing and living it. Also, when someone dies, they turn to dirt, trees grow from that dirt, people or animals eat the fruit, die, and it loops forever. Everything’s connected, helping each other out in this endless cycle. That really resonated. It’s a book packed with life lessons I’ll probably read again and again. But it also says even these lessons aren’t “real”—you’ll keep stumbling and falling to truly learn down the road.