I still see SEO plans that could have been written in 2018.
"Here are the keywords. Let's rank."
Ranking helps.
Being the actual answer, and making the visit count, is the job now.
Founding engineer vs. experienced engineer as hire #1?
The title matters less than:
– can they say no to a feature
– can they triage a vibe-coded prototype (keep / rebuild / delete)
– will they own life after launch
Titles buy commitment. They don't buy judgment.
First founder question usually isn't skills.
It's: what hours are they actually online?
AU → most of the day overlaps
US East → 4–5 shared morning hours
US West → standup + strong async
Ideas. Hours. Trust.
Comment HOURS if remote overlap is your bottleneck.
AI changed what a good engineer looks like.
Not "faster with a Copilot licence."
One person. Full judgment stack:
→ coding agents day to day
→ APIs, RAG, evals, guardrails
→ owns the work end to end
Not an ML researcher. A product generalist.
Comment JUDGMENT.
FOR your team ≠ WITH your team.
Same standups. Your codebase. Your tools.
You interview. You decide.
One senior AI-fluent generalist inside the workflow —
not a bench you manage through tickets.
Comment WITH if you've felt that gap.
The offer gets the spotlight.
The first 45 days decide if it works.
Most remote hires don't fail on skill.
They fail on small things nobody said early.
Check-ins. Honest feedback. Replace free if the fit is wrong.
Comment FIT if you've lived that first-month mess.
Small SaaS founders don't need a bigger org chart.
One senior owner who ships with AI agents:
→ coding agents day to day
→ AI product features when needed
→ product-minded, end to end
Speed got cheaper. Ownership didn't.
Comment OWNER if that's who you're hunting.
@luije87@marclou@denk_tweets The saga ends when someone has to maintain theirs for a year. Vibe-coding the copy is the cheap part. Owning billing, deliverability and support is what the money was paying for.
@DevaBuilds Depends what's broken. If you need someone to sell the vision and recruit, the exited founder. If you need the product to actually hold up in production, the experienced engineer. A lot of seed teams need the second and write the job post for the first.
@ImKingGinger I vibe-code a lot of internal tools too, and it's great for exactly this. The line I watch for is when other people start depending on it: permissions, data you can't lose, someone to fix it at 2am. That's where 15 minutes turns into a real project.
@marclou The leaderboard dying after a month and the marketplace coming out of what users asked for is the most underrated part of this. The first build is a question, not the answer. Congrats on 200.
Before vibe coding, I built prototypes in Balsamiq and walked real workflows with end users.
I still start there when I can. Now I also vibe-code first versions — fastest way to learn if an idea's worth it.
But the prototype shows what the product should do, not how it survives real users.
AI made the first version cheap. It didn't make engineering judgment cheap.
@Campaign_Create This split is underrated. A working prototype answers “should we build this?” Production answers “can we sleep when customers depend on it?” Looking forward to the risk list — especially the ones that only show up after the demo.
@nadeemk Congrats on the open roles. Curious what you’re optimizing for in the first few hires — raw product velocity, or people who already know how to own production when things break at 2am?
@majilesh Respect the “agents in production” framing. Curious what broke first when Pivra left the demo path — auth, tool permissions, or the edge cases users invent on day two?
@erikgoinsHQ Fair take on the hobby end. Where I’ve seen the opposite: once there’s a real customer and real data, “ship another vibe project” stops being an option. Then judgment and ownership suddenly matter more than the next prompt.
@yairszn This is the honest version. Vibe coding gets you a clickable first cut. The company still gets built in the boring loop — bugs, customer replies, and deciding what to throw away.
@foundrceo Agree on the catch. Speed used to be the moat — now it’s table stakes. The hard part is still knowing which 10% of the prototype is worth keeping when real users show up.
@SofiaTachExpert The scary part isn't the agent, it's that it had prod credentials and a 3-month-old backup. Boring stuff like scoped access and tested restores matters more now, not less.
@MRehan_5 The founder, every time. I vibe-code early versions myself and love it, but before anything touches customer data a real engineer reviews it. Speed for the prototype, care for production.