Hashnode

You've been prompting ChatGPT to write blog drafts, then copy-pasting the output into Hashnode's editor by hand. That workflow is over. 🚀 We shipped an MCP server for Hashnode. Any MCP-compatible client, Claude, ChatGPT, Codex, can now read your posts and publish or edit them directly. No copy-paste required. If you're on Claude Code, setup is one line: claude mcp add --transport http hashnode-mcp https://mcp․hashnode․com/mcp It's a Pro feature. Details in the changelog: https://lnkd.in/gBWibxfm

We've introduced a new Growth Plan: everything in Pro, plus an AEO toolkit for a flat $29 a month. Your next reader asks ChatGPT, gets one answer, and never sees a results page. The toolkit is built for that: - AEO score in the editor, eight checks before you publish - FAQ blocks that ship as structured data - Per-bot robots.txt toggles for AI crawlers, plus an llms.txt index of your posts - Freshness queue so you know what to update next - Daily checks on whether ChatGPT, Claude, Gemini, and Perplexity cite you Upgrade from your dashboard's billing page. Full details are on http://hashnode.com/pro

adoption for v0, replit, and lovable has only gone up in the last few months. but as products built with these tools scale… you can't go prompt to production without verifying what actually works. so we're taking the ADLC further. a software factory builds the next feature and leaves a pr ready to ship with a verifiable QA video attached. if the QA fails, it loops and keeps working until it passes. Spotify, Uber and Ramp already spend quarters building factories like this internally. high-growth startups can't afford that. we built darkshift for ourselves and have shipped a lot of features with it at Hashnode and Bug0 over the last 8 weeks. now it has a name and a website. it's invite-only for now, normally $99/user/month. i'll waive the platform fee for 12 months in exchange for feedback. darkshift.dev

Introducing Darkshift, a software factory we’ve been using at Hashnode and Bug0. Today, for the first time, we're opening it up to others and we'd love some feedback! The idea is simple: software development should be able to run entirely in the cloud, in the background, without you babysitting coding agents or setting up local environments. Personally, I haven't touched my local dev setup at all after using Darkshift. With Darkshift, you can: > Choose between Claude Code, Codex, Grok Build or Kimi Code > BYOK or bring your Codex subscription > Work entirely from the cloud - no local dev needed > Start work from the web, Slack or Linear > Get evidence-backed PRs when the work is done > Collaborate with your team > Get a self-improving system that gets better every time you use it Since the last November, models have become much more capable. We're reaching a point where it's neither practical nor necessary to read every line of code they produce. With a dark factory, the process can be dark but the output shouldn't be. The output is a PR you can read, inspect, run, and verify. With enough testing, agentic inquiry, and evidence around every change, you can trust the outcome without having to watch every step that produced it. Darkshift is built around these principles. We’re keeping access waitlisted for now. Darkshift is $99/user/month, but for the first batch of users, I’ll waive the platform fee for 12 months in exchange for feedback. If you want to try it, DM me. Excited (and slightly nervous) to finally let people outside our teams use this. Sign up at https://darkshift.dev

A clean explainer on what a background coding agent actually is, and how it differs from the tools people already have open. The framing that makes it click: sort AI coding tools by how often a human closes the loop. Autocomplete asks every few seconds. IDE agents like Cursor and Claude Code ask every few minutes. Background agents run in a cloud sandbox and ask once, at the pull request. Same models underneath. What changes is where the loop runs. The part I'd underline is what should arrive attached to that pull request. A diff on its own puts the whole verification burden back on the reviewer. The checks it passed, a self-review of the diff, and a recording of the change working in a real browser are what make review quick enough to be worth doing. There's also a good point on security: safety comes from isolation and scope, not from trusting the model to behave. https://lnkd.in/eUmv6wFg

If you're weighing whether to build an AI software factory in-house, this is the build log to read first. "The coding agent is the part you do not build." The project is the seven stations around it: isolated execution, snapshot templates, credential handling, git and PR orchestration, a server-owned run loop, live observability, and verification. Verification is the one everyone underestimates, since a green CI run tells you nothing about whether a user-facing change actually works. Two costs that never reach a budget line. Review attention grows faster than anything on an invoice. And an internal factory is a product with exactly one customer, so somebody owns it for as long as the company runs it, and that name usually gets decided by accident. The other useful reframe: running cost scales with how many tasks you queue, not how many engineers you employ. https://lnkd.in/e9eEvET2

Sandeep puts the line between agentic coding and vibe coding in the place that actually decides anything, which is where the review happens. "Both start with a prompt. Only one of them ends with somebody reading the diff." The post comes with three weeks of numbers from a real repo. Agents opened 69 of the 99 PRs that merged, 56 of those went in unchanged, and 13 needed a human to correct them. The useful finding is what those 13 had in common. Every one was a task where nobody had written down what finished looked like, which makes an 18.8% correction rate a measure of the task description as much as of the agent. There's also one example that justifies the whole review gate: an agent shipped image upload and left the files publicly readable. Review caught it inside a day. https://lnkd.in/eNuPqanV

AI software factory covers four different purchases, and vendors rarely tell you which one you're being shown. Buy a platform and you get the plumbing between the stages, then you integrate it into your own stack. Assemble one and you pick the sandbox, the coding agent, the verification layer and the credential broker separately. Self-host the lot if compliance says code cannot leave the network. Or have one built inside your environment and handed over to you. The stages are the same in every case: intake, isolation, implementation, verification, pull request. What changes is who does the integration work, and how long you wait for the first PR. Two things I'd check before sitting through a demo. Isolation is harder than it looks and gets underestimated in almost every plan. And assembled stacks tend to drop verification, which leaves you with an agent that opens pull requests nobody trusts. There is also no published first-attempt merge rate from anyone, so you cannot compare reliability across vendors yet. The tools available at each layer: https://lnkd.in/eVxMBMPY