Melv

@mal_shaik great breakdown. i read it from a different angle, what it teaches about building AI products rather than using them the "500k lines gap between demo and product" hit me differently as someone currently building

i read through the entire claude code source code so u dont have to 11 layers of architecture. 60+ tools. 5 compaction strategies. subagents that share prompt cache. most people are using maybe 10% of what this thing can do. heres everything i found:

@nichxbt @Fried_rice nice. the deferred MCP loading is clever, connects 5 servers without slowing every request the architecture is basically designed to never get in your way until you actually need a tool

@Fried_rice Created MCP Server for the leaked Claude Code source code Claude Code Explorer - MCP Server npm i claude-code-explorer-mcp npmjs.com/package/claude…

Claude code source code has been leaked via a map file in their npm registry! Code: …a8527898604c1bbb12468b1581d95e.r2.dev/src.zip

the architecture is wild. as a founder building an AI app, the thing that stood out most: 512k lines and there's still a "buddy" module, a companion sprite, quietly shipping in there. anthropic humanized their CLI. that's not an accident. wrote what this teaches about building AI products: x.com/devmelv/status…

@mark_k read through the architecture, what got me as a founder: 512k lines just to handle the edge cases. the demo of claude code is ~500 lines. the product is everything else. wrote up what this teaches about building AI products: x.com/devmelv/status…

Claude Code source code leaked:

i read the leaked claude code source code, not as a user. but as a founder building an AI app. here's what 512,000 lines of code teaches you about building AI products: x.com/devmelv/status…

@tibo_maker Curious how you think about the first 10 customers, cold outreach, community posts, or do you try to pre-sell before writing a line of code?

if i had to start over in 2026 i'd build a micro SaaS in 30 days here's the exact playbook: find a problem people are already paying to solve don't innovate on the problem innovate on the execution build the MVP in 2 weeks max launch on product hunt get your first 10 customers manually then optimize distribution: SEO for longtail keywords reddit for niche communities twitter for building in public scale what works ignore what doesn't the goal isn't to go viral the goal is to find 100 people who will pay you $50/month that's $5k MRR that's freedom most people overcomplicate this they spend 6 months building then wonder why nobody cares ship fast iterate faster full breakdown here: tmaker.io/blog/if-i-had-…...

@swyx Curious what the organizational blocker actually is, is it trust in the bot's judgment, the cost, or just inertia? Because the security case alone should be enough to mandate it everywhere.

this customer got notified 45 mins after attack, 1.5 hours before announcement of the attack. Generalist coding agent is also a better security reviewer than you! there is basically 0 reason not to have a Review bot enabled for all the things. it’s not just a security thing, it’s for paying attention to all the little stuff (9x% of the time it doesn’t matter, 10-x% of the time it matters a lot, do expected value calc of your time/money saved*) *yes this is a bit of a Pascal’s wager/insurance argument, doesn’t mean it’s wrong

@lennysan @clairevo Does Sage push back when you're behind, or mostly surface status? The difference between a real PM and a dashboard is whether it creates pressure.

One of my fav OpenClaw use-uses from @clairevo: "We created Sage the course bot. We're about a month off from launching our Maven course and Sage is project managing us to make sure that we are prepped for this course on time. It knows when the course is launching, it knows that Zach and I are (believe it or not) very introverted engineers that don't wanna market, and don't really wanna talk to humans. And so every Monday, it's like 'Claire and Zach, have you remembered to post on LinkedIn about your course? Here's a nice little post for you to copy and paste.' We would never be able to afford hiring an ops person, or hiring a content manager, or hiring a software engineer. This has allowed me to spin up a business with a quote unquote 'employee' that will eventually maybe be big enough that we can hire other people."

@gregisenberg Biggest mistake I see: founders treat distribution as a phase after building, not a constraint that shapes what you build. The projects that get customers usually baked the channel in from day one.

200,000+ new vibe coding projects get created every day yet almost NONE of them get customers 7 distribution strategies that actually work right now for your startup: 1. build an MCP server. when someone asks claude or chatgpt the question your product answers, your tool shows up. the AI becomes your sales team. 2. programmatic SEO. pick a keyword pattern (best X for Y). use firecrawl to pull real structured data so pages have actual value. one next.js template, AI generated content, human editing loop so it doesn't read like AI. 10,000 pages × 30 visits × 2% CVR × $10 = $60k/month from pages you built once. 3. vibe code a free tool (calculator, software etc). one problem, one tool, ship it today. it ranks, lives in people's workflows, markets your brand for years. ahrefs' free backlink checker has sent them more customers than most paid ads ever will. 4. answer engine optimization. people are getting answers from chatgpt and perplexity now, not just google. find the top questions your customer is asking AI. publish structured, definitive answers. one founder went from 4% to 20% AI referrals in a month just by doing this. 5. make the output of your product shareable. think spotify wrapped. think github graphs. think stripe atlas. what does your user want to screenshot and send? build that moment. add a pre-filled share button. every share is free impressions to your exact audience. 6. buy a niche newsletter. 10k subscribers for $5k to $20k. most owners are making $0 to $500 a month. DM them "ever thought about selling?" you inherit trust and a direct channel to your exact customer on day one. underrated. 7. 30 minute voice memo into claude: five tweet threads, three linkedin posts, one newsletter, short form clips. do this weekly. in 3 months you have more content than competitors who aren't doing this. obviously, your project needs to be optimized so it isnt ai slop, but you'll get there. code is commoditized. time to focus on distribution. pick 2 of these ideas and start this week to get customers. this episode was designed to get your creative juices flowing. maybe it'll give you more ideas on growth tactics you'll use this week. full breakdown on the @startupideaspod watch.

@claudeai How does auto mode handle cases where the right move is to stop and ask rather than push forward? That judgment call seems like the hardest thing to get right in fully autonomous coding agents.

Auto mode for Claude Code is now available on the Enterprise plan and for API users. To try it out, update your install and run claude --enable-auto-mode.

@dkundel sounds great! Ill give it a try

@devmelv A lot of people love Codex' code review capabilities, using it for bugs that CC can't figure out or just to have a second opinion. It's also a good way to just test out Codex' capabilities.

I built a new plugin! You can now trigger Codex from Claude Code! Use the Codex plugin for Claude Code to delegate tasks to Codex or have Codex review your changes using your ChatGPT subscription. Start by installing the plugin: github.com/openai/codex-p…

@levelsio The „just works“ principle is massively underrated in product design. DOOM understood that the best UX is no UX, you literally just click, you play. Most modern games (and SaaS projects) have forgotten this completely imo

Playing DOOM in 2026 is honestly boring but at least it lets you instantly play without an hour intro of cut scenes and tutorials and then when you load it a second time downloading a 50GB update for half an hour Modern games really just suck and it's no wonder the industry is suffering: - Revenue is down after inflation - Player counts are down - Studios are doing massive layoffs They've ignored what people want for the last decade and now they're seeing it in their numbers

@claudeai the manual test loop is the slowest part of building an ios app write code -> simulator -> click through -> notice what breaks -> describe it back to claude having claude close that loop itself is the unlock

Computer use is now in Claude Code. Claude can open your apps, click through your UI, and test what it built, right from the CLI. Now in research preview on Pro and Max plans.

building in public means admitting: i have been postponing my dentist appointment for months because the only way to book is to call heyo is literally my therapy

building heyo because i genuinely cannot explain to my doctor why i haven't booked a follow-up (the reason is the phone call)

@fwdaniels youre just trynna make us relax so you can build your b2b saas without any competition

you’re just in your 20s, relax.

@KITKAT god I thought they meant 12 trillion

Regarding recent press coverage

i just told my parents that im building an app that makes phone calls for people my dad: „why would anyone pay for that“ my mom: „can it call the doctor for me too?“ we have a product market fit with at least one person