Vijay

@vitaliidodonov Alpha

I’m launching Stanley (AI Head of Content) on June 1, 2026. Over the next 6-12 months, I'll grow it from $0 → $10M ARR. In public. (While running my $30M ARR business full time) I will also be using only AI employees to do it, so you can copy the playbook. I'll update this thread as I go. Bookmark it and follow along to hold me accountable 🤝 Comment "alpha" to get early access. My social media stats as of today:

@herohalldon @btaylor Double , triple agents …. Lol rogue agents

@btaylor an agent that builds agents is cool until you realize the second agent will also want to build agents and now you have a pyramid scheme with API keys

Today, Sierra is releasing Ghostwriter, our agent for building agents. With Ghostwriter, you can create an AI agent for your customer experience — one that can chat, pick up the phone, speak dozens of languages, take action on your systems of record, and be protected with industry-leading guardrails — simply by having a conversation. No clicking, no forms, no menus. Codex and Claude Code have transformed how we build software, making it possible for software engineers to orchestrate and review the work rather than doing all the work themselves. We think the same transformation will happen for all software. Rather than every enterprise app having a web app for humans and an API for automation, every software platform’s UI will be an agent that can do the work on your behalf. I recorded a demo of my building and optimizing an agent with Ghostwriter so you can see how powerful and easy it is to use. It’s completely changed the way our early adopters build agents, and it’s changed the way I think about the software industry. Let me know what you think, and, if you’re interested in trying it out at your business, please reach out directly.

#origin=https%3A%2F%2Fwww.google.com%2F&cap=swipe,education&webview=1&dialog=1&viewport=natural&visibilityState=prerender&prerenderSize=1&viewerUrl=https%3A%2F%2Fwww.google.com%2Famp%2Fs%2Fwww-marktechpost-com.cdn.ampproject.org%2Fc%2Fs%2Fwww.marktechpost.com%2F2026%2F03%2F25%2Fhow-to-build-a-vision-guided-web-ai-agent-with-molmoweb-4b-using-multimodal-reasoning-and-action-prediction%3Fusqp=mq331AQIUAKwASCAAgM%25253D&_kit=1" target="_blank" rel="nofollow noopener">marktechpost.com/2026/03/25/how…

Skyroot Raises ₹100Cr From BlackRock To Scale Rockets youtu.be/GH73eehLOEc?si… via @YouTube

Stop scrolling! @karpathy just turned 100 dollars of GPU time into your own ChatGPT: Train a small chat model end to end in about 4 hours on a single 8×H100 box, then chat with it in a web UI. Clean code. One script. Full pipeline. ✅ One command speedrun from empty box to talking model ✅ Full stack: tokenizer, pretrain, midtrain, SFT, optional RL, evals, web UI ✅ Minimal codebase you can read and fork, about 8k lines ✅ Runs on one machine, scales up by just changing depth ✅ Auto report card with CORE, MMLU, GSM8K, HumanEval ✅ Perfect for learning, teaching, and fast hacks If you want a real, hackable ChatGPT clone instead of a black box… This is the one to bookmark. 📌 Repo: github.com/karpathy/nanoc… 📌 Walkthrough: github.com/karpathy/nanoc…

I'm super bullish on WebGPU and wanted to share why I think it's a big deal. - WebGPU isn't just a Web API. It runs in C and C++ through Dawn, in Rust with wgpu and in Deno. It's shaping up to be the first true cross platform solution for efficient GPU access. A nice outcome is being able to run the same WGSL and TypeScript code in the browser and in Deno for games or simulations. - It adds compute shaders which let you run functions in parallel on the GPU instead of sequentially on the CPU. This opens the door to more than graphics. Local inference with libs like ONNX or transformers.js becomes practical and basically free in the browser. - WebGPU puts full AAA games in the browser within reach. With fewer CPU round trips (compared to WebGL), large engines get close enough to native performance. A lot of games bundle Chromium today just for menus and overlays (CEF). With WebGPU the entire experience can run in the browser itself. WebGPU is a major step toward browsers behaving like operating systems. Roughly how this might play out: - professional motion graphics and 2D video tools (2026+) - AAA games and advanced 3D tools (2027+) - Consumer inference running locally in the browser (2028+) Don't sleep on it.

Got a simple implementation of the voxel-based Nubis cloud system working with WebGPU/TSL (references in comments). Only issue is it runs at about 5 FPS 😂 lots of optimizations to work through.

Introducing TurboQuant: Our new compression algorithm that reduces LLM key-value cache memory by at least 6x and delivers up to 8x speedup, all with zero accuracy loss, redefining AI efficiency. Read the blog to learn how it achieves these results: goo.gle/4bsq2qI

Software horror: litellm PyPI supply chain attack. Simple `pip install litellm` was enough to exfiltrate SSH keys, AWS/GCP/Azure creds, Kubernetes configs, git credentials, env vars (all your API keys), shell history, crypto wallets, SSL private keys, CI/CD secrets, database passwords. LiteLLM itself has 97 million downloads per month which is already terrible, but much worse, the contagion spreads to any project that depends on litellm. For example, if you did `pip install dspy` (which depended on litellm>=1.64.0), you'd also be pwnd. Same for any other large project that depended on litellm. Afaict the poisoned version was up for only less than ~1 hour. The attack had a bug which led to its discovery - Callum McMahon was using an MCP plugin inside Cursor that pulled in litellm as a transitive dependency. When litellm 1.82.8 installed, their machine ran out of RAM and crashed. So if the attacker didn't vibe code this attack it could have been undetected for many days or weeks. Supply chain attacks like this are basically the scariest thing imaginable in modern software. Every time you install any depedency you could be pulling in a poisoned package anywhere deep inside its entire depedency tree. This is especially risky with large projects that might have lots and lots of dependencies. The credentials that do get stolen in each attack can then be used to take over more accounts and compromise more packages. Classical software engineering would have you believe that dependencies are good (we're building pyramids from bricks), but imo this has to be re-evaluated, and it's why I've been so growingly averse to them, preferring to use LLMs to "yoink" functionality when it's simple enough and possible.

How Language Shapes the Way We Think | Lera Boroditsky | TED youtu.be/RKK7wGAYP6k?si… via @YouTube

@Aazarshad Mods

Native ads are crushing it on ad accounts. I figured out how to make them with prompts. Instead of 50+ hours analyzing winning hooks, copy them in a few clicks. I made a deck of all of them with Moda. (Usually charge $1000 for this) Like and comment "Moda" and I'll send it.

This guy built a distributed web crawler on Hetzner for $200/mo and wrote about exactly he did it. Banger.

@HenryCrochemore Funnels

$180k+/month brands don’t build funnel pages randomly they follow a clear structure strong hierarchy clean layouts instant readability so the page converts before people overthink look at these listicle-style pages review-driven layouts simple product storytelling different formats same backbone problem → proof → offer most funnels fail because they look the same too much clutter no flow no clear direction top brands simplify everything guide the eye stack proof make the next step obvious no theory no fluff just repeatable page frameworks that convert i broke down the exact structure behind these rt + comment “funnels” and i’ll send it (follow for dm)

@akoratana 100x

We're also giving away a curated collection of 200+ Claude Code Skills our team uses daily — the workflows that made us faster engineers while building PlayerZero. Repost and comment "100X" to get access.

Introducing: PlayerZero The world's first Engineering World Model that puts debugging, fixing, and testing your code on autopilot. We've raised $20M from Foundation Capital, @matei_zaharia (Databricks), @pbailis (Workday), @rauchg (Vercel), @zoink (Figma), @drewhouston (Dropbox), and more PlayerZero frees up 30% of your engineering bandwidth by: 1.⁠ ⁠Finding the root cause for bugs & incidents in minutes that engineering teams take days to identify. 2.⁠ ⁠Predicting in minutes, edge case issues that a 300-person QA team would take weeks to find. ------ Here's why this matters: No one in your org has a complete picture of how your production software actually behaves. Support sees tickets. SRE sees infra. Dev sees code. Each team builds their own fragmented view - and none of these systems talk to each other. When something breaks, everyone scrambles to stitch the picture together by hand. PlayerZero connects all of it into a single context graph - → The Slack thread where your lead said "we went with X because Y fell apart in prod last time" → The PR review where an engineer explained the tradeoff → The lifetime history of your CI/CD pipeline, observability stack, incidents, and support tickets So you can trace any problem to its root cause across every silo. And it compounds. Every incident diagnosed teaches the model something new. The longer it runs, the deeper it understands - which code paths are high-risk, which configurations are fragile, which changes tend to break which customer flows. So when you sit down to debug a live issue, you have your entire org's collective reasoning and production memory behind you - instantly. ------ Zuora, Georgia-Pacific, and Nylas have reduced resolution time by 90% and caught 95% of breaking changes and freeing an average of $30M in engineering bandwidth. ------ Our guarantee: If we can't increase your engineering bandwidth by at least 20% within one week, we'll donate $10,000 to an open-source project of your choice. Book a demo - bit.ly/3NlLMeN

@QueenFlex_AI_ Build

Most people scroll. Few people build. That’s why most stay broke. In 2026, you can start with: • Free AI tools • Zero experience • Just your phone No excuses left. Want beginner steps? Comment “BUILD”

@Mohiniuni @rosemoni18 Kit

Most people think Claude is for writing content. I have created a kit to help you navigate what you can actually do with the tool. → build investor-grade presentations → structure your entire pitch → write hooks, slides, and talking points → even guide visuals and storytelling And this? Just a glimpse. The real power is in the full prompt system behind it. Comment “KIT” and follow me, I’ll send you the complete setup. FREE for 24 hours!!!

@vitddnv @Finance_Nerd_ Send

REVEALED: How I went from $0 to $10K MRR in 14 days. "The 14 Day Revenue Launch System" Comment “send” below and I’ll DM you my full 19-page playbook. (must follow to receive the DM)