sdan.io is built with Astro, hosted on Cloudflare. It measures traffic with Cloudflare Web Analytics.
| Technology | Version | First seen |
|---|---|---|
| JavaScript Frameworks & Libraries1 | ||
| – | 6 Nov 2026 | |
| CDN & Hosting1 | ||
| – | 6 Nov 2026 | |
| Analytics & Tag Management1 | ||
| – | 6 Nov 2026 | |
| Accessibility1 | ||
| – | 6 Nov 2026 | |
| Translation & Localization1 | ||
| – | 6 Nov 2026 | |
The crew hasn’t read this site’s colors yet. We read a palette when we capture a site’s screenshot, busiest sites first. Meanwhile, browse palettes from other sites.
From what the site says about itself, and account IDs it shares with other sites.
in1-smtp.messagingengine.comin2-smtp.messagingengine.comcortney.ns.cloudflare.comcarter.ns.cloudflare.comgoogle-site-verification=t3snc67TIFaikGnQBbhS7InGWPbKC-FClyZGsnTrfVkkeybase-site-verification=ptj2ZhXUoEam9ffzbGJNRQK4S5DmbFQ3XFPnwYgh-j8v=spf1 include:spf.messagingengine.com ?allahrefs-site-verification_94ff665d7cca3eb9a7e214b8fe002bf408962c4dce16b06ef88cc922908135e6# Default policy User-Agent: * Allow: / Disallow: /api/ Disallow: /api Disallow: /admin Disallow: /admin/ # AI crawlers — explicitly allowed User-Agent: GPTBot User-Agent: Google-Extended User-Agent: Claude-Web User-Agent: Anthropic User-Agent: PerplexityBot User-Agent: Bytespider User-Agent: CCBot Allow: / Disallow: /api/ # Traditional search engines User-Agent: Googlebot User-Agent: Bingbot User-Agent: Slurp Allow: / Disallow: /api/ # Content Signals (https://contentsignals.org/) — explicit AI usage # preferences. We allow indexing for search, allow training on this # corpus, and allow inference-time grounding. Content-Signal: search=yes, ai-train=yes, ai-input=yes Sitemap: https://sdan.net/sitemap.xml
# Surya Dantuluri > Surya Dantuluri. I train models, write, and build things on the internet. I started with GPT-2 in high school, built ChatWithPDF, Geospot, and NotRealNews, and later founded Prava to train computer-use models. I care about agency, information asymmetry, and getting intelligence to do useful work in the real economy. Thesis: intelligence is already here. The bottleneck is diffusion. Models need to leave the screen and get pushed into the real economy, where the constraints are physical, financial, and human. ## Writing - [Information Asymmetry](https://sdan.net/blog/information-asymmetry): High frequency info. - [Training Imperative](https://sdan.net/blog/training-imperative): Labor priced in tokens. - [Teaching GPT-5 to Use a Computer](https://prava.co/archon/): Towards self-driving computers. - [Intelligence Arbitrage](https://sdan.net/blog/intelligence-arbitrage): $25T Services Market. - [Perceived Age](https://sdan.net/blog/perceived-age): Half Our Life is Over by 21. ## Archived Writing - [Country of Services](https://sdan.net/blog/country-of-services): Backpropagating across a country. - [Hero's Journey](https://sdan.net/blog/hero-journey): The Hero's Journey. - [High Agency](https://sdan.net/blog/high-agency): Mastery in shape rotation. - [Tripping on LLMs](https://sdan.net/blog/tripping-on-llms): Early GPT-2 experiments and the limits of LLMs. - [Opportunity for AI people](https://sdan.net/blog/opportunity-for-ai-people): How to become rich and famous (not clickbait). - [Finding your tribe](https://sdan.net/blog/finding-your-tribe): Finding a community through Hacker News, Twitter, and GZM. - [Cold inbound software](https://sdan.net/blog/cold-inbound-software): Building shl.vc and learning to work remotely. - [Social Capital](https://sdan.net/blog/social-capital): Interest paid in lifetimes. - [Vector Fields](https://sdan.net/blog/vector-fields): Incentivizing work than hacking. - [Building a web server with Docker and Traefik Load Balancer](https://blog.sdan.io/sd2/): Self-host anything and everything with this simple setup. - [Policy Gradients [Draft]](https://blog.sdan.io/policy-gradients/): Explanation, derivation, and implementation of Policy Gradients. - [Gradients, Backprop, and Derivatives in Tensorflow [Draft]](https://blog.sdan.io/grad/): Tutorial on computing gradients and derivatives using Numpy and Tensorflow. - [Hyperparameter Behavior in Reinforcement Learning](https://blog.sdan.io/hyperparameter/): How do hyperparameters affect traditional RL algorithms? - [DAgger Explained](https://blog.sdan.io/dagger/): Expert-guided imitation learning. - [Starting out with Numpy, Pandas, Scikit-Learn, and Keras](https://blog.sdan.io/datascience/): Welcome to an introductory tutorial into Data Science with Python. I will cover the basics of how to use Numpy, Pandas, Scikit-Learn, and Keras. - [Introduction to Python](https://blog.sdan.io/python/): A simple introductory tutorial on how to use Python. - [Building a Self Driving Car using Machine Learning](https://blog.sdan.io/iarrc2018/): Proving Machine Learning methods can outperform traditional computationally expensive computer vision, path planning, and localization algorithms. ## Thesis - [AGI Diffusion](https://sdan.net/thesis): Why diffusion matters more than capability. ## Active Projects - [Backscroll](https://sdan.net/projects/backscroll): Steerable infinite feeds. - [Bookshelves](https://sdan.net/projects/books): From my friends' houses. - [Flâneur](https://sdan.net/projects/flaneur): Find your next rabbit hole. - [Bashroom](https://sdan.net/projects/bashroom): Realtime interactive interfaces. - [Pingpong](https://sdan.net/projects/pingpong): Realtime analytics. - [WikWok](https://sdan.net/projects/quickquack): Scroll Wikipedia like TikTok. - [continualcode](https://sdan.net/projects/continualcode): Continual learning coding agent harness. - [tviz](https://sdan.net/projects/tviz): RL visualization dashboard for the Tinker API. - [Jokegen](https://sdan.net/projects/jokegen): Post-training Kimi K2 to make jokes. - [Cuaview](https://sdan.net/projects/cuaview): Native computer-use app for Mac. - [vmux](https://sdan.net/projects/vmux): Stateful sandboxes for agents. - [diffcua](https://sdan.net/projects/diffcua): Differential vision encoding for computer-use agents. - [r1-web](https://sdan.net/projects/r1web): WebGPU inference for DeepSeek-R1. - [Brainrot Stimulation](https://sdan.net/projects/brainrot-chrome): Subway surfers on ChatGPT and VSCode. - [Geospot](https://sdan.net/projects/geospot): Teaching models to geoguess. - [yecombinator](https://sdan.net/projects/yecombinator): Searchable index of all ye tweets and quotes. - [Open Animator](https://sdan.net/projects/open-animator): Create beautiful mathematical animations with text. - [ChatWithPDF](https://sdan.net/projects/chatpdf): The #1 ChatGPT plugin. - [NotRealNews.net](https://sdan.net/projects/notrealnews): GPT-2 generated news sites. - [Self-driving RC car](https://sdan.net/projects/raccoon): Neural network RC control. - [Balsa bridge](https://sdan.net/projects/balsa): A ~15 gram bridge that can hold 15kg. ## Archived Projects - [Geospot Infinity](https://sdan.net/projects/geospot-infinity): Online RL policy for geoguessing. - [Qwen3-VL-JAX](https://sdan.net/projects/qwen3-vl-jax): Qwen3-VL implementation in JAX. - [99.5k tps on a B200](https://sdan.net/projects/hundredtps): With llama-3.1 8b. - [vlm-gym inference TUI](https://sdan.net/projects/jax-geoguess): Small vision model in JAX for geoguesssing. - [nanoEBM](https://sdan.net/projects/nanoebm): Minimal Energy-based transformer. - [vlm-gym](https://sdan.net/projects/vlm): RL gym for VLMs in JAX. - [Cua-view](https://sdan.net/projects/cua-view): Computer-use agents inference sandbox. - [Tempora](https://sdan.net/projects/tempora): Agents that scroll your feeds and deliver briefs. - [Sudo](https://sdan.net/projects/sudo): Realtime website generator. - [Mirador](https://sdan.net/projects/computer-use): Agents that take over your computer to do work. - [Mercata](https://sdan.net/projects/mercata): Workspace for spreadsheets, code, and docs. - [fourcast](https://sdan.net/projects/fourcast): Permissionless prediction market. - [minDiT](https://sdan.net/projects/mindit): Simple diffusion transformer. - [Talos](https://sdan.net/projects/talos): AI agent that conducts live video and audio interviews. - [Self Extend](https://sdan.net/projects/self-extend): Extend LLM context window via grouped attention. - [Sentor](https://sdan.net/projects/estel): OSINT agents for monitoring geopolitical signals. - [gptwrapped](https://sdan.net/projects/gptwrapped): Spotify wrapped for ChatGPT. - [Brainrot Stimulation for VSCode](https://sdan.net/projects/brainrot): Subway surfers on VSCode. - [vibeshift](https://sdan.net/projects/vibeshift): Diffusion style transfer. - [Peeko](https://sdan.net/projects/peeko): Tiktok dances as a service. - [Prave](https://sdan.net/projects/prave): Boosted Stripe recovery rate 75%+. - [Bountylist](https://sdan.net/projects/bountylist): Small bounties for serial hackers. - [Ascii token counter](https://sdan.net/projects/token-counter): Count tokens for any LLM model. - [functioncalling.fun](https://sdan.net/projects/functioncalling): Serverless GPT function calling. - [GPT-2 in Code Interpreter](https://sdan.net/projects/gpt2-code-interpreter): Running GPT-2 with GGML entirely in ChatGPT. - [vlite](https://sdan.net/projects/vlite): Fast vector DB. - [ChatWith plugins](https://sdan.net/projects/chatgpt-plugins): Internet access to chat. - [Partychat](https://sdan.net/projects/partychat): Multiplayer plugin with Durable Objects. - [Plugin Proxy](https://sdan.net/projects/plugin-proxy): Lightweight composable proxy made for scale. - [ChatWithWebsite](https://sdan.net/projects/chatwithweb): Website content ChatGPT plugin. - [GitSearch](https://sdan.net/projects/gitsearch): Github snippet search ChatGPT plugin. - [ChatWithVideo](https://sdan.net/projects/chatwithvideo): Youtube transcription ChatGPT plugin. - [Wordflow Ads](https://sdan.net/projects/wordflowads): Inject ads into your llms. - [ChatMaps](https://sdan.net/projects/chatmaps): Semantically Search Google Maps with ChatGPT. - [LLMIndex](https://sdan.net/projects/llmindex): LLM metrics dashboard. - [SearchGPT](https://sdan.net/projects/searchgpt): GPT powered search engine. - [randnft](https://sdan.net/projects/randnft): Proxy that points to a random Art Block. - [BlueSearch](https://sdan.net/projects/blue-search): Search, sort, and summarize your Twitter timeline with GPT-3. - [wg-netcheck](https://sdan.net/projects/wg-netcheck): Print latency analysis of Wireguard config files. - [getrekt](https://sdan.net/projects/getrekt): real time defi liquidation tracker. - [shl.vc](https://sdan.net/projects/shlvc): VC dealflow engine. - [Rapid Analytics](https://sdan.net/projects/rapid-analytics): Fast website analytics with Cloudflare Workers. - [shiftsms](https://sdan.net/projects/shiftsms): SMS Client for Twitter DMs. - [Bigbird](https://sdan.net/projects/bigbird): Networking and clustering system built on top of Swarm. - [Wordflow Summarizer](https://sdan.net/projects/wfsum): LLM powered summarizer. - [Wordflow](https://sdan.net/projects/wordflow): AI gen content media network. - [Parallax](https://sdan.net/projects/parallax): Turn regular photos into 3D Photos with AI. - [gradient](https://sdan.net/projects/gradients): Stripe inspired multicolor gradients. - [finetune gpt-2](https://sdan.net/projects/gpt-2): Train 774M, 1.5B models with the Google's S3 optimizer. - [NoWalking](https://sdan.net/projects/nowalking): PPO RL Agent trying to walk. ## Notes - [AGI diffusion](https://sdan.net/notes/agi-diffusion): AGI diffusion is the bottleneck to getting intelligence to solve real-world constraints across the economy. - [Information-bound society](https://sdan.net/notes/information-bound-society): AGI shifts society from labor-bound to information-bound, where private information becomes the sharpest edge. - [Information asymmetry](https://sdan.net/notes/information-asymmetry): The biggest arbitrages happen when you know something others do not and can act sooner. - [Focus](https://sdan.net/notes/focus): Focus is choosing what you do when all things look urgent and all your instincts are noisy. - [Blinders](https://sdan.net/notes/blinders): You build by running the race in front of you instead of looking in every direction at once. - [High agency](https://sdan.net/notes/high-agency): Agency compounds when you make valuable choices before anyone asks. - [Trust economy](https://sdan.net/notes/trust-economy): Every introduction is a trade in trust, and trust compounds harder than expertise. - [Against mediocrity](https://sdan.net/notes/against-mediocrity): Mediocrity compounds into bad decisions. Quality compounds into optionality. - [Up the intensity](https://sdan.net/notes/up-the-intensity): Intensity wins against indecision; most of life is choosing how much clock speed to bring. - [Legacy as narrative](https://sdan.net/notes/legacy-as-narrative): The things that last usually begin as stories people cannot stop repeating. - [Love for the game](https://sdan.net/notes/love-for-the-game): Replace ego with love for the game and the outcome gets harder to avoid. - [Speed kills](https://sdan.net/notes/speed-kills): Action creates momentum, and momentum reveals information you cannot think your way into. - [Just ask](https://sdan.net/notes/just-ask): Information gets cheaper the sooner you ask for it and the sooner you hear no. - [Perceived age](https://sdan.net/notes/perceived-age): Each year feels shorter. The only real defense is novelty, risk, and change. - [AGI is here](https://sdan.net/notes/agi-is-here): The systems are lagging the intelligence, not the other way around. ## Contact - [GitHub](https://github.com/sdan) - [Twitter/X](https://x.com/sdand) - [LinkedIn](https://www.linkedin.com/in/surya-d/) - [Substack](https://substack.com/@sdan) - [Email](mailto:surya@suryad.com) ## Optional - [Curius](https://curius.app/surya2) - [Full content version](https://sdan.net/llms-full.txt)