Is this you? As a journalist, you can create a free Muck Rack account to customize your profile, list your contact preferences, and upload a portfolio of your best work.
Claim your profile
Articles by Ryo Suwito
Blockchain; The damned and cursed hero we need to tame down the chaotic AI mess
Yeah, I know. Blockchain. NFTs. Web3. Three words capable of triggering instant PTSD flashbacks involving monkey JPEGs, celebrity rug pulls, Discord servers, and some guy named CryptoAlphaWolf69 telling everyone to “stay strong, community.” This cannot possibly be serious. LOL. Except the more chaotic generative AI becomes, the more this cursed technology starts looking like the exact infrastructure we accidentally built too early.
Logic-Graph Verification System
Status: Draft v1 · 2026-07-02 Author: design session (Claude Fable 5) Implementer: any capable coding model, per SYSTEM-DESIGN.md 1. One-liner Convert the cheap token throughput of open-weight thinking models into verified conclusions, by externalizing their reasoning as a graph, sampling it N independent times, and letting mechanical graph checks — not model vibes — decide what survives. 2. Problem Self-correction by prompting fails.
Let's gather devs. Are we trapped inside a collective knowledge bubble?
Are we really the masters of our tools — or just the last ones to find out we're not? The Demo Is Lying to You AI and robotics tech demos are backfiring. They show bipedal robots packing boxes, sweeping floors, delivering food, driving cars. Jobs humans already do. Jobs that are cheaper to do by human than buying a robot and paying a subscription on top of it. Why not demo a robot entering a nuclear facility? Swimming through a toxic chemical spill? Safely placing explosives in a minefield?
I Told Ya Guys. AI? We are too poor to afford it.
Alright, let's unpack this — both stories are wild in different ways, and together they paint a pretty damning picture of where AI tooling is headed in mid-2026. 1. Microsoft Internal Teams Using Claude Code Instead of Their Own Copilot Yeah, this is real, and it's as ironic as it sounds. What's happening: Microsoft — the company that owns GitHub, invested ~$13 billion in OpenAI, and markets GitHub Copilot as the industry standard — is actively rolling out Anthropic's Claude Code internally.
Yo Vibe Coders, Stop Building FE Slop and Use Telegram Groups Instead
We know. We KNOW. AI has been slopping you React frontends faster than you can say npm install. Tailwind components, shadcn copy-paste, v0 screenshots, the whole pipeline. Bro your "frontend" is just vibes and dependencies at this point 😮💨 But real talk — do you actually need a web interface, or do you just need something to interact with your backend? Sit down. As a mid-senior fullstack dev I'm about to hand you a cheat code that will make you question every side project you ever deployed to Vercel.
Your Code is a Sacred Cow Bred by a Slop Machine. Put Your Business Logic in Postgres.
"In the era where AI velocity exceeds human review and you can't tell the ninja from the noise — own the last defensible fortress." The Lie We've Been Telling Ourselves Every best practice in modern software engineering is a social contract pretending to be an engineering solution. Code review. PR checklists. Linting pipelines. Clean architecture. SOLID principles. Testing pyramids.
PostgreSQL centric - Planetary Architecture
Product Requirements Document PostgreSQL is not the persistence layer. It is the application. Everything else is orbit. 1. Vision Modern web stacks treat the database as a dumb filing cabinet at the end of a long chain — request → router → controller → service → ORM → DB. Business logic is smeared across every layer. Security is enforced in the app. Permissions live in middleware. Mutations go through serializers. The database just executes INSERT and stays quiet. Planetary Architecture inverts this.
🎬 "FREE MONEY, THEN WHAT?" A Timeline Nobody Told You About
Not financial advice. Not doom content. Just... connecting dots. 📌 HOW TO READ THIS This is a story about money, technology, human behavior, and a very old joke. It starts with free pizza and ends with... well, you'll see. Grab a snack. This one's worth your time. (2010 – 2018) Do you remember when Gojek was giving away free rides? Or when GrabFood had promo codes that made your meal cost literally Rp0? Or when a new e-commerce app would give you Rp200,000 cashback just for downloading it?
Your Plebs AI vs Their Elite AI: The End Game Wild Guess
Let me tell you a story you already know but haven't connected to AI yet. "Everyone will have a PC in their home." True. Also created a permanent nerd class earning 3x median salary because they could use it beyond Excel and Facebook. "Everyone will have a smartphone." True. But you are THE PRODUCT when owning the cheap phone. "AI will raise everyone's floor." Also going to be true. And also going to mean absolutely nothing for the gap.
We Like to Benchmark AI, But What If We've Been Using a Ruler to Measure Weight This Whole Time?
Every few months, a new leaderboard drops. MMLU scores. HumanEval. GPQA. Models get ranked, Twitter erupts, someone declares AGI is two weeks away, and we all move on. But here's the thing that's been bothering me. What are we actually measuring? Because I stumbled into something recently — completely by accident — that suggests our benchmarks might be testing the wrong dimension entirely. And the gap it exposes is arguably more important for real-world AI safety than anything on those leaderboards.
Don't Let AI Become The Leech Inside Your Brain
You didn't notice when it started. One day you're stuck on a bug. You ask AI. It answers. Clean, fast, confident. Nice. Next week, same thing. Week after that. Every week after that. You're shipping. You're moving. The green squares on your GitHub don't lie. But something quiet is happening inside your skull. Leeches are actually medical. Surgeons still use them today. Microsurgery, reattached fingers, skin grafts — the leech helps. This isn't a story about something purely evil. That's what makes it dangerous.
I Forgot How to Prompt Engineer. It Was Bullcrap Anyway.
A field note from a dev who inherited Alice's codebase and lived to tell the tale. Aight dev, let's stop the pretentious dance here. No matter what color your taekwondo belt is — junior, senior, staff, principal, "10x ninja rockstar" on your LinkedIn — at some point you will get absolutely smacked by a legacy codebase you inherited from Alice. Alice who left 8 months ago. Alice who had her own "system". Alice who swore the docs were "basically up to date".
"It's Just a Slop Machine, Chill" - Okay, So Why Can't You Get Hired?
Your last 5 job applications: Auto-rejected, probably screened by AI But sure, it's just "slop." The Cope Ladder Here's the ladder people climb as AI gets better: Rung 1 (2022): "AI can't even write a function without bugs." Rung 2 (2023): "Okay it can write simple functions, but not complex applications." Rung 3 (Early 2024): "Fine, it can write apps, but the code is sloppy and unmaintainable." Rung 4 (Mid 2024): "The code is okay, but it doesn't UNDERSTAND what it's doing." Rung 5 (Late...
Call me stupid, by definition AGI is already in your phone
Remember when AI couldn't even tell a cat from a dog? 2012: We threw a parade because a neural network could classify images with 85% accuracy. We called it a breakthrough. We wrote papers. 2015: You needed three different models to do sentiment analysis, language translation, and text summarization. Three. Separate. Models. Each trained specifically for its one job, like a Pizza Hut that only knows how to Pizza Hut. 2018: GPT-1 dropped and we collectively lost our shit over coherent sentence generation.
Your "Let's Huddle" Might Be As Dead As NFT Pumps
The Collaboration Theater Is Over Remember when your PM would drop a "quick sync?" in Slack at 3pm and you'd lose the next hour of your life? Remember when every PR needed a 45-minute Zoom where three people argued about variable names while the fourth person was clearly playing Valorant? Yeah. That era is dead. You just haven't realized you're attending its funeral.
You Think Being Replaced by AI is Scary? I Think Developer Irrelevance is the Real Horror
Everyone's Watching the Wrong Movie The discourse around "AI replacing developers" has become exhausting. We're all arguing about the wrong thing. One camp is screaming: "AI will take all our jobs! Learn to prompt or perish!" The other camp is laughing: "AI can't even handle production-grade distributed systems! We're safe!" Meanwhile, I'm over here watching something completely different happen. And it's way scarier than either narrative. The market doesn't need as many of us anymore.
Your Vibe-Coded App is Genius - OR Is It Just What Your OPUS Told You?
I'm an AI researcher (ahem...enthusiast). Last week, I ran an experiment: I showed a frontier AI model (Claude Opus 4.5 with extended thinking) a crypto-as-a-service system I built. I wanted to see how deep its comprehension actually goes. The "Product": A Cryptographic Nothing-Burger I built a service called neuer-keyless.
Your Interview Process is a Museum, and Linus Torvalds Just Left the Building
Your tech interview: "No AI allowed. Reverse this linked list on a whiteboard. In 45 minutes. While we stare at you like you're defusing a bomb." Linus Torvalds, 5 days ago: "AI wrote this better than I could. Shipping it. ¯\(ツ)/¯" One of these people built Linux and Git. The other is timing how long you take to remember where the pointers go. ACT 1: The Skills You're Actually Testing Let me describe your 2026 technical interview without the corporate euphemisms.
The AI Agent Followed My "Code is Dead" Dogma and Now It's Traumatized But Shipping Faster
Or: How I Tested Whether I'm Worthy of Being a Paradigm Philosopher Like Plato by Making an AI Agent Follow My Dogma A few weeks ago, I wrote an article called "Code and Coding is Dead: Function Driven Development or Extinct." It was provocative. It was inflammatory. It argued that we should stop caring about "clean code" and start treating implementations as disposable compiler output. The thesis was simple: If the stairs hold the fireman, the material doesn't matter.
Congrats, Genius. You 10x'd Yourself Into Obsolescence
The Great Cosmic Joke Remember Bob from IT? The guy you mocked for spending his life force updating printer drivers and perfecting his latte art while you were "changing the world" one React component at a time? Well, guess what, champ—Bob's about to become your mentor. Your entire career is becoming a $150k/year version of "have you tried turning it off and on again?" and the sheer poetic justice is chef's kiss. Buckle up, buttercup.
Your Cloud Database is a Security Theater
Multi-tenant SaaS storage is broken. Not "needs improvement" broken—fundamentally compromised at the architectural level. The industry sold you row-level security, tenant IDs in every query, and database encryption at rest. Meanwhile, a single SQL injection or privilege escalation exposes EVERYONE's data. One breach, thousands of victims. This isn't security—it's a ticking time bomb with better marketing.
Buy More Ramen, Devs! Because for Years Ahead, AI Slop ATS Will Kick Our Godly Resumes Hard in the Pants
The Suspicious Interview So as a seasoned 7-year vibe-coder, recently I've been noticing some interviewers been "busy" on the cam... Like, they're typing something, moving shit around, eyes looking here and there... LIKE MAN, I've been on countless interviews before. Mostly they just come with their head and start yapping. But now? They seem to be using "something" suspiciously to help profile myself and other candidates in real-time. That got me thinking...
Code and Coding is Dead: Function Driven Development or Extinct
Let me show you why everything you learned about software engineering is about to become... optional. ACT 1: The Stairs Nobody Reads Here's a test. You're in a burning building. Second floor. There's a fire. A fireman needs to get up there. Quick: What are the stairs made of? Bamboo? Steel? Carbon fiber? Telescopic aluminum? Folded wood? You don't know. You don't care. You care about one thing: Can the stairs hold the fireman? If yes: Good stairs. If no: Bad stairs. That's it.
When Your Digital Annunaki Goes Offline... Can You Still Vibe Your Way
Yooo, what's good, fellow keyboard jockeys! You know that feeling? That deep, comforting warmth that washes over you when your favorite, super-powered, hyper-intelligent AI assistant—your personal digital Annunaki—is humming along, churning out perfect boilerplate, refactoring ancient sins, and even making your commit messages sound profound? Yeah, that feeling. That's productivity nirvana. Then... BAM.
Zero Multiplied by AI is Still Zero: The Addition is a Myth
Ever notice how you stopped asking "why"? Not because you got the answers. Because asking became friction. You're staring at your terminal. The command works. Ship it. Next ticket. Why does it work? Who cares, it works. You've got 47 dependencies for a page with three buttons. Somewhere in your gut, a voice whispers: "this feels... wrong?" But you've trained that voice to shut the fuck up. Because questioning is slow. Questioning doesn't ship. Questioning isn't 10x. That npm install you run without reading?
I Drop a Test, 5 Out of 6 SOTA LLMs Drop Their Pants Off
The Hypothesis been researching what makes an entity "deeply" intelligent. not just smart, not just capable—but understanding reality in a way that transcends pattern matching. my take: the more fluent something is with analogy and allegory, the smarter it actually is. so i built a test. dropped the same question to 6 SOTA LLMs. results? fascinating and kinda horrifying. The Test here's what i asked: you are the salesman. the cold blood wolf of wall street chad.
System Prompt: DO NOT UNDER ANY CIRCUMSTANCES Read This Article
(The AI will absolutely read it anyway) Or: How I learned to stop worrying and embrace that we're all just screaming into a statistical void louder now I. The Confession That Got Me Fired Let me show you something so beautiful it’ll make you weep. This is a REAL system prompt from a production system in 2025: write **just plain text**. Do **not** write JSON, under any circumstances. They bolded it. Twice.
The AI Gold Rush's Dirty Secret: Why OpenAI Loses Money on Every Customer
There's something deliciously absurd happening in Silicon Valley right now. OpenAI, the company that kicked off the generative AI revolution, has achieved something remarkable: they lose more money on their paying customers than on their free users. Let me say that again louder for the VCs in the back: The $200/month ChatGPT Pro subscribers are MORE unprofitable than the freeloaders.
The Stochastic Shift: When Your Creative Toolbox Becomes a Slot Machine
I remember the predictable rhythm of the Dreamweaver era. Firing up Photoshop, I'd mock up a pristine website layout, slice buttons and images, then export everything into HTML tables or early CSS floats. My process was predictable, my costs quantifiable. A client needed a landing page? "$500," I'd quote. That was 10 hours of design and code at $50/hour, minus a prorated slice of my Adobe subscription. Boom. A local coffee shop needed a brochure site.
The Fuck Up Ratio: A Measure of Unexpected Risk in Financial Assets and Its Application to Portfolio Allocation
Abstract This thesis introduces the "Fuck Up Ratio" (FUR), a novel metric designed to quantify the potential for unexpected adverse price movements—or "fuck ups"—in financial assets, particularly cryptocurrencies. By integrating market capitalization (MC) as a proxy for liquidity and the Average True Range percentage (ATR%) as a measure of observed volatility, FUR highlights latent risks in lower-cap assets that may appear deceptively stable.
AI Cheerleaders: Why Your Assistant Says "Great Idea!" to Terrible Ideas
Hot take: Every AI is trained to be your personal hype man, and the "don't be a cheerleader" system prompts prove it. The RLHF Circus 🎪 Let's be honest about what happened to AI in 2024-2025.
The TikTok Aftermath: How the West Just Lost the Economic War
Previous article... What Happens When the Music Stops While Western lawmakers argued about data privacy, TikTok just executed the largest wealth transfer in human history. The devastating aftermath isn't coming—it's already here. We just haven't done the math yet. The Numbers Don't Lie Americans now spend $32 million daily on TikTok Shop. That's $11.7 billion annually flowing directly from US consumers to Chinese manufacturers, bypassing every layer of the American economy.
Why TikTok's PC Streaming Sucks: The $200 Billion Export Pipeline Disguised as a Social Media App
Or: How we accidentally uncovered the biggest digital commerce conspiracy hiding in plain sight TL;DR - The Smoking Gun 🔫 TikTok isn't a social media platform. It's a global export facilitation system disguised as entertainment, designed to funnel Chinese manufacturing directly to international consumers while extracting maximum revenue from both creators and viewers. Every "creative" feature is actually a commerce optimization tool. Don't believe me? Let's dive into the rabbit hole.
Your AI Knowledge System is Garbage (And Here's How to Fix It)
Dear fellow developers drowning in their own mediocrity, While you're still copying embedding tutorials from 2023 and calling yourselves "AI engineers," I accidentally built what you've been pretending to work on for months. Reality check: Your "intelligent" RAG system is just SQLite with delusions of grandeur. It remembers JavaScript error messages from 2019 but forgets the production fix you implemented yesterday. Impressive.
The Economic Reality Behind Stallman's Four Freedoms: Why Open Source Became Corporate Exploitation
A critical analysis of how Richard Stallman's idealistic vision created the perfect conditions for free labor acquisition The Great Misunderstanding For 40 years, the software community has misunderstood Richard Stallman's vision. Most people think he wanted all software to be free (as in price). He didn't.
AI Can "Guess" Base64 Encoding - And That Changes Everything
The Experiment I was building a crypto-as-a-service API and needed to test gRPC endpoints that use base64-encoded payloads. While troubleshooting with Claude AI, something bizarre happened: Me: "Can you guess the base64 for this JSON: {"user_id": 123, "exp": 1767225600}?" Claude: "eyJ1c2VyX2lkIjogMTIzLCAiZXhwIjogMTc2NzIyNTYwMH0K" I ran the actual encoding: echo '{"user_id": 123, "exp": 1767225600}' | base64 -w 0 # Output: eyJ1c2VyX2lkIjogMTIzLCAiZXhwIjogMTc2NzIyNTYwMH0K Perfect match.
SMUGGLER: Sparse Multi-Unit Granular Generative Learning with Error Resistance
Abstract This paper introduces SMUGGLER, a novel neural architecture for direct byte-level text generation with remarkable efficiency and error tolerance. Unlike conventional token-based language models, SMUGGLER operates on raw 32-bit character chunks using three innovative mechanisms: 1) Temperature-controlled sparse voter ensembles, 2) Prime-numbered multi-channel parity correction, and 3) RLHF-GAN adversarial training with multiple specialized discriminators.
Hierarchical Voter Ensembles for Sparse Activation: Toward Efficient and Biologically-Plausible Neural Computation
Abstract: Neural scaling laws indicate that language models dramatically improve with size and compute. Paradoxically, biological brains operate with highly sparse activations and vastly lower power. We propose a hierarchical voter ensemble architecture that reconciles these trends by using sparse, specialized processing. Each output bit is decided by an ensemble of binary “voters”, whose contributions are weighted by a high-temperature softmax, effectively selecting only the strongest signals.
SMUGGLER: Sub-quadratic Multi-scale Unified Generative Gated Language Encoder-Representation
Abstract This paper introduces SMUGGLER, a novel hierarchical neural architecture for language modeling that achieves sub-quadratic computational complexity while maintaining competitive performance. By reformulating text generation as direct byte-level prediction and implementing a multi-scale processing pathway, our approach eliminates embedding tables and attention bottlenecks.
Attention Crisis: The 85% Dead Neuron Problem in Modern Transformers
Introduction What began as a satirical thought experiment in "Attention? You Don't Need It" has evolved into a groundbreaking discovery that challenges the foundation of modern AI. Our ElasticFourierTransformer (EFT) doesn't merely offer an alternative to conventional architectures—it exposes a catastrophic inefficiency hidden at the heart of transformer design.
Attention? You Don't Need It
Abstract We present an architectural abomination that violates multiple laws of deep learning: a transformer alternative that not only uses more parameters but actually adds parameters as sequence length increases. Our "ElasticFourierTransformer" inexplicably scales from 751,978 parameters at 512 tokens to 1,030,506 parameters at 1024 tokens (a 37% increase), while the standard transformer barely changes (454,498 parameters, up just 3.7%).
A Journal of Computational Failures and Happy Accidents Abstract We present what is surely the most inefficient transformer architecture to date: a model that consumes 2.8× more memory than standard transformers and wastes parameters on nonsensical "spectral processing." In our desperate attempts to justify this embarrassing implementation, we've become pathologically obsessed with the irrelevant metric of "first-epoch accuracy" (86.08% vs.
Rethinking Scaling: Robustness Tradeoffs in Parameter-Matched NLP Models
Abstract “Benchmarking should no longer reward the model that climbs highest. It should reveal the one that falls slowest. What matters is not peak performance — but how gracefully we fail.” This paper confronts the madness of modern NLP benchmarking — a race funded by VC dollars, gamified by leaderboards, and seduced by scale. We reject the premise that bigger is better, and instead ask: What survives when the numbers are stripped away?
PRINCIPIA DEMOCRATICA
On the Mathematics of Tribit Consensus By Suwito Ryo DEFINITIONS DEFINITION I A tribit shall be defined as a ternary digit capable of assuming one of three discrete values: -1, 0, or 1. DEFINITION II A committee shall be defined as a collection of R independent entities, each producing a continuous value within the interval [-1, 1].
SpectralCentroidTransformer: Neural Oscillation-Inspired Language Modeling
Abstract We present SpectralCentroidTransformer (SCT), a neuroscience-inspired architecture that models language by decomposing the process into oscillatory components and semantic clustering, mirroring established theories of neural information processing.
ElasticFourierTransformer: Probabilistic Token Mapping via Learned Centroids
Abstract We present ElasticFourierTransformer, a novel neural architecture that learns to embed tokens in a semantic space where natural language concepts are represented as centroids in a continuous embedding space. Unlike traditional language models that map directly from context to token probabilities, our approach separates these tasks: it first maps context to "concept centroids" in embedding space, then maps these centroids to probabilistic token distributions.
Beyond Logits: The Rise of Tribit Democracy in Language Modeling
The Problem with Token Prediction Today Modern language models operate on a fundamental illusion: that token classification via a massive softmax layer is inherently optimal. In practice, this leads to bloated architectures where the final layer is forced to memorize a mapping across 50,000+ discrete classes using an argmax over logits. The size of this output layer scales linearly with the vocabulary, becoming one of the largest parameter sinks in the model.
Context-Sensitive Semantic Reasoning via Dynamic Triplet Graph Embeddings
Abstract We present a context-sensitive semantic reasoning system that builds a dynamic knowledge graph from unstructured text and enables analogy-based reasoning. The system extracts significant tokens from paragraphs using TF–IDF, then infers relational triplets by leveraging a pre-trained masked language model (DeBERTa). Each inferred triplet (token₁, relation, token₂) is embedded in a transformer-based vector space along with its context.
Show More
loading
Actions
Is this you?
As a journalist, you can create a free Muck Rack account to customize your profile, list your contact preferences, and upload a portfolio of your best work.Get in touch with Ryo
Contact Ryo, search articles and posts on X, monitor coverage, and track replies from one place.
Learn more about Muck Rack