Post

@seyong Hi Se x.com/Shesho3/status/207894658…

@seyong yo lebron james mentioned but i hear ya x.com/LAfitnesshooper/status/2…

@seyong Hopefully your actual engineers can save you from going to prison Seyong.

Agree. I vibe code a lot and even as founder I would never ship anything serious without a technical cofounder or at the very least a CTO. That being said, a creatives brain is wired different than a an engineers brain and they keep us creatives grounded. Even if I could operate without one, I would never

@seyong Everyone thinks AI is a tool. it's the first technology that charges you per second of existence while it thinks about whether you exist. the bill for thinking

@seyong Preach brother. 🙏

@seyong This will be untrue in about 4 months.

I don't think you're wrong but I imagine that the major gap between the two is speed to delivery rather than how elegant or well optimized the code is the #1 engineer will obviously deliver a better output but will do it in half the time or less using AI than a vibecoder just due to existing knowledge and more efficient, targeted prompts

Ofc an engineer will be faster, and get the integrations worked in easier too, and can fix it fast as well. But i dont you are right, as vibecoding allows people to be founders, or marketers Things that a lot of engineers wouldnt be good at. And, if mythos was released, we would be closer. So I don't think that many orders of magnitude is correct, and Sam Altman thinks it won't be long until there are one-person billion-dollar companies. And I know that that's not going to compete with Meta or with Google, of course. But let's say many orders of magnitude here is 5. So, 100,000x as "hard" to comprehend the next Meta or google... When, probably by Jan 2027, the engineering at these places drops 20%, and the first 1-2 man, billion $ company exists.

@seyong Context is everything. @shaunmmaguire articulated a similar understanding of excellence as you described here in this interview (around 8 mins in): youtu.be/62X1vm3QBuE?si=7hn3C-…

@seyong Who tf cares, its all about who can get more users or make more money in the end

yeah, LLMs don’t provide that level of precision, no matter how much data you feed them. a farmer won’t suddenly gain the technical precision, architectural vision, or nuanced understanding that comes from building software. sure, as you said maybe one-shotting meta may become possible. but if everyone can do it, the most creative and technical people will build artifacts orders of magnitude beyond what we’re used to. something something, “bicycle for the mind” ~ Jobs.

@seyong Feels the same for any practical application I think. I like to think about AI maximizing our abilities, not replacing

@seyong You don't know, what you do not know.

@seyong bro just used three different metaphors to say 'you dont know what you dont know' and somehow the basketball one hit different bc i felt it in my soul like a missed layup in 5th grade gym 💀

@seyong Lots of great apps will be built from a vibecoded prototype then a cracked engineer coming in to fix it

AI is definitely raising the floor, but the ceiling is still determined by skill. The best engineers aren’t just writing code faster, they’re making better decisions about architecture, tradeoffs, debugging, and what to build in the first place. That’s the gap most people don’t see.

@seyong $fomo x.com/Terafabsol/status/207908…

@seyong It's like comparing a sketch to a finished painting in terms of depth.

@seyong Hence why education in computer science and software engineering still exist to this day

@seyong Real asf

@seyong people confuse shipped with built lol

@seyong form gap

@seyong @Crypto_McKenna :)

@seyong the gap is probably way bigger than people want to admit lol x.com/0xkyliekim/status/207918…

Good breakdown. I see this a lot with vibecoded consumer apps saying distribution is their moat. It only works when they're copying stuff big tech already made, and then big tech just buys them. Nothing changes. It's the easy road. It's like opening a burger stand across from McDonald's. Best case, McDonald's buys you. You were never gonna beat them selling the same burger. Whoever beats McDonald's is gonna sell something people have never eaten before. I think that's what agents are. Ask anyone what Google is and they'll say a search engine. Ask them what an agent is and you won't get a straight answer. Whoever defines that for people wins. That's the stuff worth building toward imo.

@seyong Just wait until you discover how horrible assembly code 99% of compilers write. And in comparison to an elite assembly programmer it looks indistinguishable for the common dev!

@seyong Very accurate comparison. From the outside it may look like everyone gets the same result, but the quality of decisions and architecture is still defined by the human.

@seyong also experienced engineers probably burn far fewer tokens to get to the same outcome