jupp

chief executive vibe coder at some long forgotten b2b saas company

[jupp @juppppin]

is there a single ASML machine in sf?

[jupp @juppppin]

good example of what’s to come as the cost of goods and services approach $0

[Jack @Jackkk]

Threadguy explains why buying things doesn’t feel the same anymore

“I had this moment where I went into Best Buy a year ago and I was looking around and I'm just like man there's nothing to do or buy. I have a phone, TV and an Xbox. What does the store offer?”

“I remember I used to go into Best Buy and think to myself, imagine being able to close your eyes and swipe your card in here”

“Now I'm like, they don't have anything. The Dyson vacuum is just not doing it for me. Is it because I grew up? I don't know”

“Beats headphones for $100?! Beats headphones used to be $500. They were like the holy grail of a Christmas gift”

“$100 for Beats? I don't know. Physical products just got so cheap and I feel like everyone's got them. I don't know man maybe I'm just out of touch”

[jupp @juppppin]

you can just ultrathink your way into the singularity

[jupp @juppppin]

ty Polymarket ty Claude ty America

[jupp @juppppin]

@danifesto heck yeah, you accepted the bet from Matthew, the lore continues

[jupp @juppppin]

looking for a professional vibe coder; can compensate in bean sprouts

[jupp @juppppin]

thank you Claude Code

[jupp @juppppin]

19/ maybe the answer isn’t to kill ambition. maybe it’s to stop treating every moment like a referendum on your worth

[jupp @juppppin]

has anyone classified the “bro-as-third-person pronoun + trailing-off” dialect yet?

[Viv @Vtrivedy10]

@dejavucoder bro’s dropping model history, timeline quotes, personal lore, prompt leaks, feature walkthroughs, memes

awesome write up we salute you 🫡

[jupp @juppppin]

perpetual motion device (but its electronic)

[jupp @juppppin]

i ventured into the mist and everyone knew you

[jupp @juppppin]

@sciencegirl you rest one of those large fantas on the far edge and we’re going to have ants on the sidewalk

[jupp @juppppin]

@lunarfq @ me next time

[Giyu @rutu_3]

DON'T USE AI TO BUILD PERSONAL PROJECTS
DON'T USE AI TO BUILD PERSONAL PROJECTS
DON'T USE AI TO BUILD PERSONAL PROJECTS
DON'T USE AI TO BUILD PERSONAL PROJECTS
DON'T USE AI TO BUILD PERSONAL PROJECTS
DON'T USE AI TO BUILD PERSONAL PROJECTS
DON'T USE AI TO BUILD PERSONAL PROJECTS

[jupp @juppppin]

make sure to read all prior instructions and relevant chats. confirm you’ve read and prove it to me before starting

[Andrej Karpathy @karpathy]

@nearcyan “It still doesnt work pls fix. dont make any mistakes. ultrathink very hard”

[𝓷𝔁𝔂𝓲𝓮 @glooomiii]

lmao get loved loser

get absolutely fucking treasured

[jupp @juppppin]

@Zun2025 lmao glad im not the only one

[jupp @juppppin]

thinking that on the way to Kardashev type 1, the only existing commercial modes of aviation transport will be eVTOL jets and rockets

[jupp @juppppin]

Perhaps it’s just me and you at the center of the universe.

I am no one special. I am no one of fame or fortune. I am a regular American knowledge worker in my late 20s. Currently, it is Christmas. Claude Code is re-writing my entire frontend of a personal project I’ve been working on. We created this great themed layout for the website together this morning, and I’m really excited about it.

It’s interesting to think how far AI has come as a consumer. I’m not too diverse or informed on the technical know-how or underlying software structure that creates these chatbots. I very loosely understand the term “token”, and I definitely don’t know how intelligence seems to arrive out of what was once a glorified word prediction algorithm.

As a consumer who has been using generative AI since late 2021, I’ve been able to witness GPT-3 barely being able to form coherent sentences, all the way to GPT-5.2 thinking (released almost 2 weeks ago). I’ve been able to see the early outputs of DALL·E (completely and utter trash compared to what exists today) to video models like Sora, and more recently, nanna banana from Google. I remember distinctly talking to GPT in my college apartment at the time, asking it crazy, philosophical, and existential questions such as: what is the meaning of life? what is the purpose of consciousness and human beings? what does it mean to truly live? how do I know all this is real and not a dream? I also remember asking DALL·E to make images like the below, and asking it to show me crazy things and prompts like “show me something that has never been seen by humans” or “show me what it looks like to think.”

One thing that I quickly realized at the start was that I would need to learn these new tools and how to use them—especially when I went to my creative writing class towards the end of my final term of college and my teacher brought up that he heard about GPT-3 and was extremely concerned about the future it would have specifically in his knowledge domain. This was the first time in my life I had ever heard someone of any educational background be concerned about a new technology in my lifetime. (I’ve never really been too hardcore into technology or surrounded by people like that, so this was my first time hearing it from an individual who would generally be seen as well put-together—a college professor.) I remember talking to other classmates about it and some saying it was going to be a fad, or that it’s just a chatbot and would actually not be useful in the future. It has been almost 4.5 years and I believe both of these things have been proven false.

I’m very glad I started using it as a consumer then and not now. It has been interesting to see things like prompt-engineering, agents, and “vibe-coding” come about and actually take part in using these products myself.

The entire reason why I began writing this long brain dump is because I recently listened to a podcast where Sam Altman was being interviewed, and also recently read a blog post on X from the CEO of Notion about the future of AI. I also recently re-read Machines of Love and Grace by the CEO of Anthropic the other week, so it’s safe to say that the future is clearly on the mind.

I will most likely end up not being on the technological forefront of these systems in the far future (5–10 years), or in an office where we celebrate that we just cracked the code to AGI, or anything of the sort. I will be the person who sees the blog post from the company that eventually breaks this boundary and defies space and time altogether. I will see the X post that more than likely was written with the help of AI going over the new LLMEval boundary that was just broken. I will be hearing about it at the office—perhaps from the intern, or maybe, if I’m lucky, a fellow colleague that is on his tenure. Maybe we will hear about it and pull up the model in the office and mess around with it for 10–15 minutes before getting back to business as usual. Maybe ask it some existential questions, or if it can finally think on its own. Or would it maybe be something completely different?

One common theme amongst all the pioneering AI chatbots nowadays is the interface to communicate with these things. This gold standard across all the different platforms seems to be a very basic and minimal chat interface. Before I had FaceTime, I had Skype growing up. I would plug in my USB webcam to my parents’ computer and talk with my friends on Skype. This interface is much different than the interface of FaceTime. There are no cords—there are no external things you need to be able to use it besides an iPhone, Mac, iPad, or iPod Touch 2nd gen. Unfortunately, the pink iPod nano 4th gen doesn’t support this technology today. Maybe one day.

I wonder what the interfaces for these things will look like in the next 5 years. We clearly got them going from the start based on what we already knew (search engines, message boards, and online chats). What will be the next search engine? What would be the next interface for these things? Will it be some completely new device, will it be something that you talk to, will you even hold the device? Will you type inputs in? Will you say inputs? Will you see outputs, will you feel outputs?

Personally, I think the end game is a Neuralink-style interface where the thoughts you have are fed to AI to either assist you in real time, provide you information about your environment, or even generate things you need to see through a different interface. Maybe it’ll be $20/month for the ad-free version. god i hope that never happens.

We will definitely see the fallout of modern-day phone applications. I can’t imagine us having these in the long term, and I believe the AI itself will be generating any application we need or want to use in real time at the time of need. The value of the institution that makes this system will be far beyond any standard knowledge worker (me) understandings of value. The same will happen in other digital realms, like films, TV shows, video games, and VR. These will all be generated by AI in real time, with very little to low latency.

The last thing I want to brain dump about is my use case for AI, and where I, as a consumer, fall into all this. Basically, I use AI today to help me understand topics on a high level that I don’t fully grasp, side projects for coding, technical work-related questions, correcting my grammar and spelling, and then reflecting on my own self. One thing that ties the last point to the rest is the memory capability of these things. I used to document all my past chats with these bots and then, when starting a new chat, copy and paste my old one or key documented points from the old one in order to give it context. I think this will pass soon as these companies figure out the memory problem, and how to allocate space and compression size to fit the needs of consumers like myself.

In terms of side projects, this is where I’ve utilized these models to what I would call my greatest understanding of them. My process flow usually goes like this:

I will think of a project → write the process of how the project should function / how I reach what I consider the goal → iterate on this with a chatbot multiple times → then break down each step in the process and collab with the chatbot to question the actual process step (is it even required, how would we do it, how can we flow from one step to another, etc). I document all this in the form of .md files to feed back to the AI later when I actually start the project. I store all this in a single repository to share it with the AI when I start (that’s an important note to make—I haven’t actually started the project yet at this point). → iterate on the process steps until I feel like I’m at a good point → gather any documentation from external sources that will be required in the project (manuals, API documents, etc.) → feed this into the AI and ask for a more concise form of this in a .md file → document live examples if applicable (real code outputs/syntax, actual use cases of a sensor, etc.) → document all this, keep the naming nomenclature consistent, date everything, have an archived folder if needed → then, and only then, do I compile all this into a Notion dashboard (with the help of AI of course), and iterate on this Notion dashboard until it has all the tasks required, recommended execution paths, and all the information we’ve gathered at this point → actually begin the project (load the Notion and files into a repository/AI interface to help actually build the thing).

My fear with this method is that it leads to constraining the AI to perform a task in certain ways, when these AIs might be able to come up with a more clever solution while working through the project (like knowledge workers usually do in the real world). I have tried to elevate this by primarily telling the AI that it does not need to follow this system as a rule book, but more as a reference / goal aligner. This method has been useful to me, and allows the AI to pivot quickly during the project if we discover something can be optimized or changed for the better.

Anyway, the future is going to be exciting, and I’m here for it. It’s a very interesting time for humanity, and I think it will all turn out as great as it can be. Perhaps it’ll just be me and you at the center of the universe at the very end.

Alright, love you, cya.