Spaces:

CompactAI-O
/

Homepage

Running

App Files Files Community

CompactAI commited on Feb 28

Commit

fe569e0

verified ·

1 Parent(s): 4e0af78

Create blog-Training-Models-on-a-Ramen-Budget.html

Browse files

Files changed (1) hide show

blog-Training-Models-on-a-Ramen-Budget.html +213 -0

blog-Training-Models-on-a-Ramen-Budget.html ADDED Viewed

	@@ -0,0 +1,213 @@

+<!DOCTYPE html>
+<html lang="en">
+<head>
+<meta charset="UTF-8">
+<meta name="viewport" content="width=device-width, initial-scale=1.0">
+<title>Training Models on a Ramen Budget | FMN-GPT - CompactAI</title>
+<link href="https://fonts.googleapis.com/css2?family=Inter:wght@300;400;500;600;700&family=JetBrains+Mono:wght@400;500&display=swap" rel="stylesheet">
+<style>
+:root{--color-bg:#faf8f5;--color-bg-alt:#f5f0e8;--color-bg-dark:#1a1815;--color-bg-dark-alt:#252220;--color-accent:#e85d3b;--color-accent-light:#ff8a6b;--color-accent-dark:#c44a2d;--color-secondary:#d4a853;--color-text:#2d2a26;--color-text-light:#6b6560;--color-text-muted:#9a948d;--color-border:#e5e0d8;--shadow-md:0 4px 20px rgba(45,42,38,0.12);--font-sans:'Inter',-apple-system,BlinkMacSystemFont,sans-serif;--font-mono:'JetBrains Mono','Fira Code',monospace;--container-max:1200px;--section-padding:100px}
+*,*::before,*::after{box-sizing:border-box;margin:0;padding:0}
+html{scroll-behavior:smooth;font-size:16px}
+body{font-family:var(--font-sans);background:var(--color-bg);color:var(--color-text);line-height:1.7;-webkit-font-smoothing:antialiased;display:flex;flex-direction:column;min-height:100vh}
+main{flex:1}
+.container{max-width:var(--container-max);margin:0 auto;padding:0 24px}
+h1,h2,h3{font-weight:600;line-height:1.2;color:var(--color-text)}
+a{color:var(--color-accent);text-decoration:none;transition:color .2s}
+a:hover{color:var(--color-accent-dark)}
+code{font-family:var(--font-mono);background:var(--color-bg-alt);padding:.2em .5em;border-radius:4px;font-size:.9em;color:var(--color-accent-dark)}
+pre{font-family:var(--font-mono);background:var(--color-bg-dark);color:#f5f0e8;padding:1.5rem;border-radius:12px;overflow-x:auto;font-size:.875rem;line-height:1.6}
+pre code{background:none;padding:0;color:inherit}
+.main-nav{position:fixed;top:0;left:0;right:0;background:rgba(26,24,21,.95);backdrop-filter:blur(10px);z-index:1000;padding:1rem 0}
+.main-nav .container{display:flex;justify-content:space-between;align-items:center}
+.nav-brand{color:#fff;font-size:1.25rem;font-weight:600}
+.nav-links{display:flex;gap:2rem}
+.nav-links a{color:var(--color-text-muted);font-size:.9375rem;transition:color .2s}
+.nav-links a:hover{color:var(--color-accent)}
+.footer{padding:3rem 0;background:var(--color-bg-dark);text-align:center}
+.footer-text{color:#fff;font-size:1.125rem;margin-bottom:.5rem}
+.footer-subtext{color:var(--color-text-muted);font-size:.875rem;margin:0}
+.blog-post-section{padding:var(--section-padding) 0;background:var(--color-bg);flex:1}
+.blog-post-content{max-width:700px;margin:0 auto}
+.blog-back{display:inline-block;color:var(--color-accent);font-weight:500;margin-bottom:2rem}
+.blog-post-header{margin-bottom:3rem}
+.blog-post-header h1{margin-top:1rem}
+.blog-post-body p{font-size:1.125rem;line-height:1.8;margin-bottom:1.75rem;color:var(--color-text)}
+.blog-post-body p:first-of-type{font-size:1.25rem}
+.blog-post-body h2{font-size:1.6rem;margin:2rem 0 .8rem;color:var(--color-accent)}
+.blog-post-body blockquote{border-left:4px solid var(--color-accent);padding:1rem 1.5rem;margin:2rem 0;background:var(--color-bg-alt);border-radius:0 8px 8px 0;font-style:italic;font-size:1.1rem;color:var(--color-text)}
+.blog-post-body blockquote p{margin:0}
+.blog-post-body ul,.blog-post-body ol{margin:1.5rem 0;padding-left:1.5rem}
+.blog-post-body li{margin-bottom:.75rem;color:var(--color-text);line-height:1.7}
+.blog-post-body ul li{list-style-type:disc}
+.blog-post-body hr{border:none;height:2px;background:linear-gradient(to right,transparent,var(--color-border),transparent);margin:3rem 0}
+.blog-post-body pre{margin:1.5rem 0}
+.blog-post-body a{text-decoration:underline;text-underline-offset:2px}
+.blog-post-body strong{color:var(--color-text);font-weight:600}
+.blog-post-body em{color:var(--color-text)}
+.blog-meta{display:flex;gap:1rem;margin-bottom:1rem}
+.blog-date{color:var(--color-text-muted);font-size:.875rem}
+.blog-tag{background:rgba(232,93,59,.1);color:var(--color-accent);font-size:.75rem;font-weight:600;padding:.25rem .75rem;border-radius:50px;text-transform:uppercase;letter-spacing:.05em}
+@media(max-width:768px){:root{--section-padding:60px}}
+</style>
+</head>
+<body>
+<nav class="main-nav">
+<div class="container">
+<a href="index.html" class="nav-brand">FMN-GPT</a>
+<div class="nav-links">
+<a href="blog.html">Blog</a>
+<a href="status.html">Model Status</a>
+<a href="https://huggingface.co/CompactAI" target="_blank">HuggingFace</a>
+</div>
+</div>
+</nav>
+<main>
+<article class="blog-post-section">
+<div class="container"><!DOCTYPE html>
+<html lang="en">
+<head>
+<meta charset="UTF-8">
+<meta name="viewport" content="width=device-width, initial-scale=1.0">
+<title>The Goalpost Has Legs: Why AGI Keeps Running Away | FMN-GPT - CompactAI</title>
+<link href="https://fonts.googleapis.com/css2?family=Inter:wght@300;400;500;600;700&family=JetBrains+Mono:wght@400;500&display=swap" rel="stylesheet">
+<style>
+:root{--color-bg:#faf8f5;--color-bg-alt:#f5f0e8;--color-bg-dark:#1a1815;--color-bg-dark-alt:#252220;--color-accent:#e85d3b;--color-accent-light:#ff8a6b;--color-accent-dark:#c44a2d;--color-secondary:#d4a853;--color-text:#2d2a26;--color-text-light:#6b6560;--color-text-muted:#9a948d;--color-border:#e5e0d8;--shadow-md:0 4px 20px rgba(45,42,38,0.12);--font-sans:'Inter',-apple-system,BlinkMacSystemFont,sans-serif;--font-mono:'JetBrains Mono','Fira Code',monospace;--container-max:1200px;--section-padding:100px}
+*,*::before,*::after{box-sizing:border-box;margin:0;padding:0}
+html{scroll-behavior:smooth;font-size:16px}
+body{font-family:var(--font-sans);background:var(--color-bg);color:var(--color-text);line-height:1.7;-webkit-font-smoothing:antialiased;display:flex;flex-direction:column;min-height:100vh}
+main{flex:1}
+.container{max-width:var(--container-max);margin:0 auto;padding:0 24px}
+h1,h2,h3{font-weight:600;line-height:1.2;color:var(--color-text)}
+a{color:var(--color-accent);text-decoration:none;transition:color .2s}
+a:hover{color:var(--color-accent-dark)}
+code{font-family:var(--font-mono);background:var(--color-bg-alt);padding:.2em .5em;border-radius:4px;font-size:.9em;color:var(--color-accent-dark)}
+pre{font-family:var(--font-mono);background:var(--color-bg-dark);color:#f5f0e8;padding:1.5rem;border-radius:12px;overflow-x:auto;font-size:.875rem;line-height:1.6}
+pre code{background:none;padding:0;color:inherit}
+.main-nav{position:fixed;top:0;left:0;right:0;background:rgba(26,24,21,.95);backdrop-filter:blur(10px);z-index:1000;padding:1rem 0}
+.main-nav .container{display:flex;justify-content:space-between;align-items:center}
+.nav-brand{color:#fff;font-size:1.25rem;font-weight:600}
+.nav-links{display:flex;gap:2rem}
+.nav-links a{color:var(--color-text-muted);font-size:.9375rem;transition:color .2s}
+.nav-links a:hover{color:var(--color-accent)}
+.footer{padding:3rem 0;background:var(--color-bg-dark);text-align:center}
+.footer-text{color:#fff;font-size:1.125rem;margin-bottom:.5rem}
+.footer-subtext{color:var(--color-text-muted);font-size:.875rem;margin:0}
+.blog-post-section{padding:var(--section-padding) 0;background:var(--color-bg);flex:1}
+.blog-post-content{max-width:700px;margin:0 auto}
+.blog-back{display:inline-block;color:var(--color-accent);font-weight:500;margin-bottom:2rem}
+.blog-post-header{margin-bottom:3rem}
+.blog-post-header h1{margin-top:1rem}
+.blog-post-body p{font-size:1.125rem;line-height:1.8;margin-bottom:1.75rem;color:var(--color-text)}
+.blog-post-body p:first-of-type{font-size:1.25rem}
+.blog-post-body h2{font-size:1.6rem;margin:2rem 0 .8rem;color:var(--color-accent)}
+.blog-post-body blockquote{border-left:4px solid var(--color-accent);padding:1rem 1.5rem;margin:2rem 0;background:var(--color-bg-alt);border-radius:0 8px 8px 0;font-style:italic;font-size:1.1rem;color:var(--color-text)}
+.blog-post-body blockquote p{margin:0}
+.blog-post-body ul,.blog-post-body ol{margin:1.5rem 0;padding-left:1.5rem}
+.blog-post-body li{margin-bottom:.75rem;color:var(--color-text);line-height:1.7}
+.blog-post-body ul li{list-style-type:disc}
+.blog-post-body hr{border:none;height:2px;background:linear-gradient(to right,transparent,var(--color-border),transparent);margin:3rem 0}
+.blog-post-body pre{margin:1.5rem 0}
+.blog-post-body a{text-decoration:underline;text-underline-offset:2px}
+.blog-post-body strong{color:var(--color-text);font-weight:600}
+.blog-post-body em{color:var(--color-text)}
+.blog-meta{display:flex;gap:1rem;margin-bottom:1rem}
+.blog-date{color:var(--color-text-muted);font-size:.875rem}
+.blog-tag{background:rgba(232,93,59,.1);color:var(--color-accent);font-size:.75rem;font-weight:600;padding:.25rem .75rem;border-radius:50px;text-transform:uppercase;letter-spacing:.05em}
+@media(max-width:768px){:root{--section-padding:60px}}
+</style>
+</head>
+<body>
+<nav class="main-nav">
+<div class="container">
+<a href="index.html" class="nav-brand">FMN-GPT</a>
+<div class="nav-links">
+<a href="blog.html">Blog</a>
+<a href="status.html">Model Status</a>
+<a href="https://huggingface.co/CompactAI" target="_blank">HuggingFace</a>
+</div>
+</div>
+</nav>
+<main>
+<article class="blog-post-section">
+<div class="container">
+<div class="blog-post-content">
+<a href="blog.html" class="blog-back">← Back to Blog</a>
+<header class="blog-post-header">
+<div class="blog-meta">
+<span class="blog-date">2026-03-22</span>
+<span class="blog-tag">Hot Takes</span>
+</div>
+<h1>The Goalpost Has Legs: Why AGI Keeps Running Away</h1>
+</header>
+<div class="blog-post-body">
+<p>Imagine handing Claude Opus 4.6 to someone from 2004. They would think you summoned a minor deity. You ask it to write a sonnet about quantum entanglement while debugging a Python script and it just... does it. While making a joke about the halting problem.</p>
+<p>And our collective response? A polite nod. A slight shrug. "Cool. But can it do original scientific discovery? Can it understand why my cat judges me? Can it fold a fitted sheet?"</p>
+<p>Welcome to AGI, the finish line that sprints away every time we get close.</p>
+<p>Twenty years ago we thought chess mastery was the peak. Deep Blue arrived. We scratched our heads. "Okay but can it hold a conversation?" GPT-3 arrived. "Great, but can it reason through novel problems?" New models arrive. "Fantastic, but can it do all of that while being energy efficient and ethically aligned and also making my morning coffee?"</p>
+<p>The pattern is so consistent it feels like a law of physics. Every breakthrough becomes the new baseline. Every "wow" becomes "yeah, but".</p>
+<h2>Intelligence Is a Mirror We Keep Repolishing</h2>
+<p>Maybe the issue is not the models. Maybe the issue is us. We are spectacularly hard to impress. We are the toddler who receives a shiny new toy and immediately asks "what else you got".</p>
+<p>AGI is less a technical milestone and more a collective mood. It is the horizon we walk toward. The closer we get, the farther it seems.</p>
+<p>Here is my unpopular take. We might never "achieve" AGI because we keep redefining what achievement looks like. The moment a model can do something impressive, we quietly update the spec sheet. The goalpost does not just move. It has legs. It is doing cardio.</p>
+<blockquote>
+<p>Maybe general intelligence is not a destination. Maybe it is a direction. Or maybe it is just a really good marketing term. Either way, the models keep getting wilder.</p>
+</blockquote>
+<p>This is not a complaint. It is an observation. And honestly? It is kind of fun. The chase is entertaining. The models are getting better at writing, reasoning, coding, joking. That is wild regardless of what label we slap on it.</p>
+<p>If we ever do hit whatever we decide AGI means today, I predict we will immediately ask "cool but can it do my laundry". And when it can, we will ask about the fitted sheet. And so on, forever, because humans are built to want the next thing.</p>
+<p>So let us enjoy the ride. Let us marvel at what these systems can do. Let us keep pushing. And let us maybe, just maybe, give our past selves a little credit. The person from 2004 would be losing their mind right now. And that is worth something.</p>
+<hr>
+<p><em>Current status: Asked a state of the art model to explain why we keep moving the AGI goalpost. It wrote a surprisingly poignant haiku. Then I asked it to debug my snake game. It did. The goalpost jogged another lap. I am not even mad.</em></p>
+</div>
+</div>
+</div>
+</article>
+</main>
+<footer class="footer">
+<div class="container">
+<p class="footer-text">Built with curiosity over compute.</p>
+<p class="footer-subtext">FMN-GPT by <a href="https://huggingface.co/CompactAI" target="_blank">CompactAI</a> - 2026</p>
+</div>
+</footer>
+</body>
+</html>
+<div class="blog-post-content">
+<a href="blog.html" class="blog-back">← Back to Blog</a>
+<header class="blog-post-header">
+<div class="blog-meta">
+<span class="blog-date">2026-03-29</span>
+<span class="blog-tag">Dev Log</span>
+</div>
+<h1>Training Models on a Ramen Budget: Or, How We Learned to Stop Worrying and Love the Free Tier</h1>
+</header>
+<div class="blog-post-body">
+<p>Picture this: you, me, a dream, and a GPU budget that currently consists of three expired coffee shop gift cards and a hopeful glance at my laptop's cooling fan. This is the reality of trying to train AI models as a regular person with regular finances and an irregular relationship with electricity bills.</p>
+<p>We have all seen the headlines. "Training state of the art models costs more than a small house." "Researchers spend millions on compute." "Please do not ask us to open source the weights, the electricity company is still calling." Meanwhile, I am over here trying to train a tiny model that can finish my sentences and my credit card is sending me concerned text messages.</p>
+<p>So what if training did not require a venture capital round? What if you could tinker, experiment, and maybe accidentally create something useful without selling your future earnings to a cloud provider? That is the question we are asking. That is the problem we are clumsily, hopefully, stubbornly trying to solve.</p>
+<h2>The Comedy of Errors (Mostly Mine)</h2>
+<p>Let me share a brief, humiliating timeline of my personal journey into cost effective training. Week one: I tried to train a model on my laptop. The fans sounded like a jet engine preparing for takeoff. The model learned to predict the next character in "hello world" and then gave up. Week two: I signed up for a free cloud tier. I accidentally left a job running overnight. I received an email that used the words "unusual activity" and "account review." Week three: I discovered that "free" often means "free until you blink wrong."</p>
+<p>Through this parade of mistakes, a pattern emerged. The barriers are real. The costs are real. But so is the creativity of people who really, really want to build things without going broke. We started talking to other folks in the same boat. Students, hobbyists, researchers with tiny grants, curious coders with big ideas. We all had the same wish: what if we could share the load?</p>
+<blockquote>
+<p>The goal is not to train the biggest model. The goal is to make training possible for the person who has the idea but not the infrastructure.</p>
+</blockquote>
+<p>That is why a few other very patient, possibly confused people and I are building a website. It is a place where you can pool resources, share compute time, and maybe, just maybe, train a model without needing a finance degree to understand the bill. We are stitching together free tiers, optimizing every byte, and laughing nervously as we try to make the math work. Is it elegant? Not yet. Is it ambitious? Absolutely. Are we the right people for this? That is the funniest question of all.</p>
+<p>And before you ask, because I can hear the collective internet inhaling: FMN-GPT is not abandoned. It is not lost. It is not quietly weeping in a corner. It is training. Slowly. Carefully. Like a sourdough starter that occasionally bubbles and makes us all very excited for no reason. We are iterating, learning, and yes, sometimes breaking things. That is how this works.</p>
+<h2>Free Is a Feature, Not a Bug</h2>
+<p>Building for cost effectiveness means asking uncomfortable questions. Do we really need that extra layer? Can we quantize this? Is there a free tier we have not yet annoyed? It means embracing constraints as a creative force. It means celebrating a 0.1 percent efficiency gain like it is a national holiday. It means sometimes the most advanced technique is just turning things off and on again.</p>
+<p>We are not promising magic. We are promising effort. We are promising a space where you can try, fail, learn, and try again without the fear of a surprise invoice. We are promising that the journey of building AI can belong to more than just the well funded. And we are promising to document our stumbles along the way, because if my errors can save you five dollars and an hour of confusion, that is a win.</p>
+<p>So if you have ever looked at a training script and thought "I wish I could run this" and then looked at your bank account and thought "I wish I could eat this week," you are who we are building for. Come hang out. Bring your ideas. Bring your weird little datasets. Bring your skepticism. We will bring the optimism and the increasingly well documented list of what not to do.</p>
+<hr>
+<p><em>Current status: The website is in progress. The models are training. The coffee is strong. The GPU fans are, for now, silent. FMN-GPT is alive, learning, and occasionally outputting something that looks almost like a sentence. Progress?</em></p>
+</div>
+</div>
+</div>
+</article>
+</main>
+<footer class="footer">
+<div class="container">
+<p class="footer-text">Built with curiosity over compute.</p>
+<p class="footer-subtext">FMN-GPT by <a href="https://huggingface.co/CompactAI" target="_blank">CompactAI</a> - 2026</p>
+</div>
+</footer>
+</body>
+</html>