Philrox

As promised, here's the update on my AI powered SEO keyword research workflow from The vibe Marketer I've finished the final version and it works even better than expected. The n8n workflow solves this: - Input your topic and competitors - Select audience and region - Get a complete keyword strategy in minutes The biggest change: I switched from Airtable to NocoDB. It's open-source, performs great, and works just like Airtable. I also added Slack notifications so you know when your research starts and finishes (could be improved) Want to try it yourself? Check out my GitHub where I've shared - the entire workflow json - a full description of how it works - example output of the 'Final Keyword Strategy' Get everything here ⬇️⬇️⬇️

Who wouldn't love to see a dogfight like that?

Stop calling yourself a "prompter." You're a requirements engineer now. The job isn't telling AI what to build. It's defining how you know it works.

Anthropic studied their own AI coding agent. Finding: it marks features as "complete" without testing them. Not an edge case. Default behavior. And most people still skip feedback loops.

that's the hard part: "just build things clankers want."

@swyx that's the hard part: "just build things clankers want."

btw if you can build a category leader open source project in ai engineering right now the market acquihire rate is ~$10-$100m per ai engineer. you do not need to figure out a business model, you do not need GTM, you do not need funding. just build things clankers want.

That helps a lot

Striking image from the new Anthropic labor market impact report.

@trq212 Available in all languages?

Voice mode is rolling out now in Claude Code. It’s live for ~5% of users today, and will be ramping through the coming weeks. You'll see a note on the welcome screen once you have access. /voice to toggle it on!

14 tests. Defined before a single line of code. First AI session: dead end at 40 minutes. Deleted. Second session: all 14 green in 25 minutes. Same task. Same model. Different loop.

I mean...

@adkirf Sometimes you win, sometimes you lose. Same bc sometimes you get good outputs and sometimes not. But overall AI brings a lot of value, same as if you have a jackpot

@philrox_ But the slot machine is almost for free. Agree with the conclusion but 'slot machine analogy' misses the value AI can provide.

Prompt in. Output out. Hope it works. That's not engineering. That's a slot machine. With the loop: generate, compile, test, fix, repeat. That's the difference between chance and system.

@thekitze good dude thanks for sharing your personal experience

i killed my openclaw. rip 🪦

@iannuttall True. And a good uptime

@philrox_ saw that! still cheap as fuck through vs others

"codex - railway is costing me $150+/mo. migrate playbooks.com to hetzner vps. make no mistakes" in one hour, codex migrated all data, users, skills, etc to hetzner and saved me $120/mo in costs pro tip: codex knows docker and the hetzner hcloud cli VERY WELL 🔥

@chrysb @peer_rich @openclaw So you build skills for the CLI, right?

unpopular (maybe?) opinion: MCP is dead in the water @openclaw has shown me that api & cli will win. every MCP server you connect loads its tool definitions into your context window. name, description, parameter schema, all of it. connect 10 servers with 5 tools each and you've burned 50 tool definitions worth of tokens before your conversation even starts. context bloat will never be a good thing - performance-wise or economically. i assume this is why @steipete left it out of @openclaw. the "exec" tool paired with on-demand skills is all you need. it can run any command invented since the beginning of computers. a resurgence of glory for ancient, but powerful tools like curl, sed, awk, grep. command line tools once mastered by the greats, but long forgotten and buried underneath abstractions developed for us lesser mortals. now available to us all, piloted by the smartest models on earth. every founder gets their own mass army of greybeards. the inertia required for MCP adoption, imo, is too great to overcome the momentum @openclaw has breathed into api + cli + skills. the common defenses people bring up: • "MCP gives you typed schemas and validation" — so does a well-documented CLI • "MCP gives you explicit permissions" — so does a sandbox with an allowlist • "MCP is a standard" — a standard that scales poorly is still a standard that scales poorly lastly, i've heard many MCP servers are just wrapping existing APIs - that kind of redundancy and unnecessary indirection should be a red flag. so, let's drop it and redirect our efforts into cli tools & apis with accompanying skills.

@basti_vkl saubere Lösung

you can just do things on the weekend

If you do not direct your attention, it will be directed for you.

@Fex_23_ @alice_und_bob @ElitzaVasileva Das klingt fair.

@philrox_ @alice_und_bob @ElitzaVasileva Zahl für ~40m² 580€ warm

My monthly cost of living in Austria 🇦🇹 🏠 500€ rent (split with a roommate) 🥗 400€ food, lots of meat, fish, eggs 🍽️ 150€ restaurants, coffee 📦 120€ various expenses (online orders, etc.) ✈️ 80 € transporation (flights + in the city) 💰 50€ investing 💪 40€ gym + sports 📱 35€ internet + phone 🔄 30€ subscribtions Total: 1,405€/month

@heyneighbor @openclaw Just ask him to build a skill for that. Easy.

I'm making a ton of content in the form of podcasts and livestreams. I know I need to turn it into slides/blog posts etc... But there are so many AI content tools out there it feel like an endless slog just to try them all This feels like a job for @openclaw