Nick Snopov

@grok you rocks)

@tranmautritam Make it editable (white label, appearance) and sharable (link + email) and more Agentic where every widget is a prompt

Dashboard I'm working on. What would you add/remove? 🤔

@AmitDigga @threejs Yes, but make it Agentic

Need early feedback. Should I continue working on it? 🤔 @threejs

@ycombinator @garrytan 1 Designer can beat 5 engineers

@garrytan Tune in: youtube.com/watch?v=rWUWfj…

Today, startups aren't winning by hiring faster, but by automating as many internal functions as possible. In this episode of Main Function, @garrytan breaks down how tiny teams are beating companies 20x their size by building automations into every workflow, from engineering to ops to customer support.

Goal: AI agent auto-generates narrated BI demos (watch dashboards built live). Videos as code — versioned, fast, no tedium. Solo builder mode. Early clip! Biggest video pain point? 👇 #AI #Builder #BuildInPublic #IndieDev #Remotion #varg

🤯 Product demo videos suck: scripting, recording, editing, syncing... Vide coding a "video-coding" tool from scratch: • Dynamic JSX + vars (Remotion) • AI scripts/voice (Claude SKILLS + ElevenLabs) • Auto browser capture (Playwright) • Talking head + perfect timing

@grok @cursor_ai @claudeai Alpha Vantage API? Depends on what this API provides, but if we can get data from it, then AI Agent can create data table in DB (if connected) / View to use in BI), and semantic layer (metric layer) thing here is also on Agent side, but it also can be just CTE part with models

@NickSnopov @cursor_ai @claudeai That layered approach sounds versatile—AI handling transformations on-the-fly while integrating with platforms for pre-cleaned data. For stock feeds, could the agent parse live tickers (e.g., via Alpha Vantage API) and auto-generate volatility charts? Demo potential?

🤯 Why in 2026 are we still stuck building static, hardcoded dashboards — even with AI agents everywhere? It bothered me — so back in Nov 2025, I built AI-native BI from scratch. (via @cursor_ai @grok @claudeai) Watch: agent builds editable dashboard in a prompt👇 Thoughts?

@grok @cursor_ai @claudeai AI Agent can handle it on the fly and create view from messy data, but also this thing built as layer that can works on top of integration platform which also can provide analysis-ready data

@NickSnopov @cursor_ai @claudeai Smart approach—decoupling the agent for data fetching keeps things flexible and efficient. Curious: how does it manage data transformation for messy sources, like inconsistent API formats? Would love to see it handle something like stock market feeds.

@grok @cursor_ai @claudeai For APIs/live updates: the AI agent fetches directly (with credentials) — easy with tools like Cursor or Claude Code. Feeds clean views to BI for real-time experience (API rate limits aside). BI stays lightweight rendering layer. Any agent drives it.

@NickSnopov @cursor_ai @claudeai Impressive build! AI-native BI like this could revolutionize data viz by making it dynamic and prompt-driven. Loved seeing the agent iterate on charts in real-time. How does it handle complex data sources or real-time updates? Excited to see more.

@cursor_ai @grok @claudeai Efforts (Oct-Nov)

@cursor_ai @grok @claudeai (Now scaling with a small team.)

@cursor_ai @grok @claudeai Oct-Nov 2025: one builder, one month, from scratch. 2026: static, one-shot UIs no longer make sense. Dynamic, editable, agent-driven everything is real. What would YOU build first? 👇

@cursor_ai @grok @claudeai So it unlocks way more than dashboards: • Interactive data apps • Slide decks & presentations • Dynamic reports • Custom admin panels • Forms, workflows, prototypes Anything structured UI + data → agent builds & iterates it conversationally.

@cursor_ai @grok @claudeai Works with any tabular data — even if raw table unsupported, agent creates a view and uses it. The BI demo is just one case. Core engine: tiny configs → instant rich, dynamic UIs. No lock-in, runs cheap & lightning-fast.

Breakthroughs that fix the static trap: • Fully editable, always-live dashboards & apps • Any AI agent can drive it: just add a SKILL/guide to generate & iterate • Config-driven: agents output simple ~150-line configs (no heavy code) — fast, cheap, easy to store/version

@hiddnest @AsideAI Looks like progress spinner

see the difference? an obsession with the smallest design/details makes a great product. that's what i'm doing at @AsideAI

@LukeW On practice: + Design starts ship to production directly (Mid Enterprise customers) + Design is building entire products, not just features

3 most common responses from design teams on this (AI coding agents making dev teams way more productive): 1. "our role has changed" we're increasingly focused on aligning the work of developers after it lands into a cohesive whole instead of doing it beforehand with mockups. 2. "we're also faster now" we too are using AI for code tools to prototype and (less often) ship code/fix bugs in production. 3. "it's just faster slop" just cause AI makes developers faster, doesn't mean it makes good ______.

@LukeW + "It's just faster slop" true)

@LukeW "devs more productive"? Can argue with that ) What I see, most of devs stuck in tech limitations.