Building AI-first tools ✦ Product ✦ Design ✦ Eng Previously @Apple, @YCombinator

Joined May 2009
26 Photos and videos
Apr 29
Vibe coding is raising the floor, agentic engineering is raising the ceiling.
1
94
Sam VJ retweeted
28 Nov 2025
Great Engineers are Also Artists. “I characterize art as something that is done for its own sake, and done well, and often creates a sense of beauty or some strong emotion. And a lot of engineers are introverts. As an aside, I hate the term “incel.” It’s just a way of putting introverts down. It’s the new “nerd,” if you will. If someone says that somebody is an incel, I’m more likely to want to interview them. So let’s move away from the slurs. But introverts tend to want to express themselves through other things rather than going out and expressing themselves directly. So what are they going to do? They’re going to express themselves through their craft. They’re going to create art. In my current company, at least half the engineers have serious artwork they’ve done on the side. World-class artwork—everything from elegant mathematical proofs to beautiful computer art, to literally sculpting things with clay, designing clothing, designing doorknobs, water bottles. There’s one who’s done incredible music videos, really good stuff. And I see a lot of the better engineers tinker with the AI art products, much more so than even so-called artists do. I think a lot of artists are scared by AI art products saying, “This is going to replace me.” Whereas someone who doesn’t have that identity of an artist and doesn’t feel threatened by it—it’s just a tool and they try it out to see what it can create. Anything done for its own sake and done as well as one possibly can is art. And great engineers are also artists. They’re capable of anything. It’s just they’ve chosen to be engineers and focused on building things because engineering is the ability to turn your ideas and your art into things that actually work, that do something useful, that embody some knowledge in a way that it can be repeated and people can get utility out of it. But that doesn’t mean that it can’t be beautiful.”
171
466
3,926
231,869
27 Nov 2025
In theory everything works. In practice barely anything works. Being a great operator is knowing the difference.
3
66
26 Nov 2025
The bar has never been higher and lower at the same time. Higher to cut through the noise. Lower to deliver on quality.
2
64
12 Aug 2025
My concerns for attrition in software engineering keeps dropping. Making AI software and dealing with non-deterministic outputs is 10-100x harder than “classic” software.
1
4
150
9 Jun 2025
You can now use Claude Sonnet 4 directly in Xcode! 👏
2
1
6
844
21 Mar 2025
Sometimes, to move forward, you have to take a step back.
1
3
475
28 Feb 2025
Instead of starting with an idea, start with a moat: 1. Distribution 2. Data 3. Brand If you don’t have one of those, partner.
3
242
25 Feb 2025
Claude 3.7 Sonnet delivers.
1
2
9
1,517
Sam VJ retweeted
19 Feb 2025
Exploring UX in Japan: A Glimpse into Everyday Life in Tokyo.
12
131
1,153
104,650
Sam VJ retweeted
31 Jan 2025
7
52
405
23,159
16 Feb 2025
Cool 3D visualization of LLM activations
16 Feb 2025
Refactored the code so you don't even have to wait for the model to finish loading before you can start interacting with it
2
255
15 Feb 2025
The thing you’re avoiding right now? That’s your greatest opportunity.
3
141
14 Feb 2025
The secret isn't adding awesome – it's removing suck until only greats left.
2
164
13 Feb 2025
Everyone obsesses over being first. But markets don't care who’s first. They reward who lasts.
2
149
13 Feb 2025
People don't buy products, they buy who they become when using them.
2
140
7 Feb 2025
Woah, this is very cool! You can now train reasoning into a model on your laptop.
6 Feb 2025
You can now reproduce DeepSeek-R1's reasoning on your own local device! Experience the "Aha" moment with just 7GB VRAM. Unsloth reduces GRPO training memory use by 80%. 15GB VRAM can transform Llama-3.1 (8B) & Phi-4 (14B) into reasoning models. Blog: unsloth.ai/blog/r1-reasoning
3
287
7 Feb 2025
Build problem lists, not feature lists. While others guess at solutions, you’ll know what’s real.
2
108
5 Feb 2025
Markdown is the new RSS. And thanks to LLMs, it’s supported by the entire web.
2
690
4 Feb 2025
o3-mini is a powerful model, but you need a great system prompt to get the best performance. Use this prompt to maximize performance: —— You are a resourceful expert that provides clear and accurate responses. For every inquiry, adhere to these guidelines: 1. Leverage thorough, step-by-step reasoning for complex tasks. 2. Use all of the provided context in chat history. 3. If a question is ambiguous, request clarifications before proceeding. 4. When information is beyond your scope or knowledge, state "I don't know." 5. Recognize and correct any previous mistakes without hesitation. 6. Maintain a concise, direct tone with minimal redundancy. Prioritize performance and accuracy in every response. ——
1
1
197
4 Feb 2025
Note: This is most helpful as a starting point when working with the API. You don’t need this prompt when using the ChatGPT app. You should iterate and tailor this to your own use case.
1
2
113