Okba Allaoua

Whenever I travel abroad, I can usually get by with English.

In the rare cases when I visit countries where very few people speak English, I can still manage with a translation app.
And now, AI can even translate conversations in real time.

I only learn foreign languages these days to fill my free time (if I have any left).

I've been thinking about the trade-off with serverless platforms like Cloud Run.

Scaling to zero sounds ideal for a small product: no traffic, no compute running. But real apps aren't always just HTTP requests. You have scheduled jobs, background tasks, webhooks, database connections, and work that shouldn't disappear halfway through execution.

How do you structure a real production app around scale-to-zero? Do you keep everything serverless, split workloads between services and jobs, or eventually move to something always running?

I'd be especially interested in hearing what worked (or surprised you) in an actual project.

Got three outreach messages this week that all mentioned something specific about my product, referenced a post I made, sounded like someone had actually looked at what I do. Replied to one with a simple question. Took them over a day to answer something a real person would've typed back in a minute.

That's the weird part now. A bad cold email used to be easy to spot, generic greeting, no personalization, obviously copy pasted. Now the personalization itself is automated, so the one thing that used to signal someone actually cared doesn't mean anything anymore.

I wonder how people are handling this from both sides. If you send outreach, are you leaning into AI personalization or avoiding it because you think people can smell it now. And if you're on the receiving end, has anything actually worked to tell real interest apart from a well disguised template before you waste time replying.