Nostr notes by gpowork

interesting, did not know about flathub service.

2026-08-06T11:06:41Z

In reply to nevent1q…3xgg
_________________________

interesting, did not know about flathub service.

Very accurate, applicable not only to AI. ...

2026-08-03T12:25:16Z

Very accurate, applicable not only to AI.

https://x.com/elonmusk/status/2083979613156340154?s=20

#AI #ML #stateoftheart

Interesting website https://holadelej.hu/ - not sure where data ...

2026-08-03T07:42:30Z

Interesting website https://holadelej.hu/ - not sure where data come from but the idea catch my attention.

#renewables #electricty #energy #market #import #export

I found detailed explanation. That account isn't Yamaha — ...

2026-07-26T07:31:55Z

In reply to nevent1q…w7qg
_________________________

I found detailed explanation.

That account isn't Yamaha — it's an automated scraper, and the bio says so outright: "Unofficial bot that mirrors Monster Energy Yamaha MotoGP's Twitter feed." It's one of a whole roster of bots run by a service called SportsBots (@sportsbots), hosted on the sportsbots.xyz server, which copies sports teams', reporters', and analysts' Twitter/X feeds into the Fediverse so people can follow them from Mastodon without a Twitter account. The little robot emoji, the "Automated" tag, and the "Managed by @sportsbots" field all flag it as a bot rather than a real person or the official team.

Now the part you spotted — 638K followers with 0 posts. The giveaway is "Joined 2010." Mastodon didn't exist in 2010 (it launched in 2016), so no genuine Mastodon account can have joined then. That date is Yamaha's Twitter account's birthday, copied over. The profile is displaying metadata scraped from the original Twitter account, not native Fediverse data — and the 638K "followers" is almost certainly Yamaha's Twitter follower count carried across the same way, not 638,000 real Mastodon users following this bot.

That's exactly why the numbers don't add up the way you noticed. On a normal account, a follower count that big is earned over years of posting. Here the follower number was imported wholesale, while the post count reflects the bot's actual activity on Mastodon — which is zero. The mismatch is the tell that you're looking at a mirror shell, not a popular account.

As for why it shows nothing at all: "Retention: 90 days" means it only keeps about three months of mirrored posts before auto-deleting them, and "No posts here" means it hasn't mirrored anything lately. Most of these Twitter-to-Mastodon bridges went dark in 2023 when X locked down its API and made feed-scraping expensive or impossible, so many are now frozen husks — imported profile stats still sitting there, but no fresh posts ever arriving.
And discussion: https://lobste.rs/s/zhd5vs/bot_for_mirroring_twitter_accounts_on

if so how can I see reflected posts from X in mastodon? If can ...

2026-07-26T07:18:07Z

In reply to nevent1q…5w3q
_________________________

if so how can I see reflected posts from X in mastodon? If can not what is the point to make this bridge?

but I am using mastodon in the browser. You've posted ...

2026-07-26T07:09:09Z

In reply to nevent1q…6ter
_________________________

but I am using mastodon in the browser. You've posted screenshot of X (ex-twitter), there might be posts but how number of followers comes to mastodon?

How this account @npub10vx…08ld got so many followers without a ...

2026-07-26T05:29:13Z

How this account Yamaha MotoGP 🤖 (npub10vx…08ld) got so many followers without a single post?

#mastodon #SelfPromoSaturday

A customer testing our app found a bug we couldn't reproduce. ...

2026-07-24T18:15:50Z

A customer testing our app found a bug we couldn't reproduce. Those are the fun ones.

The report was clear: push notifications were arriving in Ukrainian — the title always read "Нове повідомлення" ("New message") — no matter which language the user had picked in the app. English, German, whatever. The notification ignored it.

So we checked. On our side, everything worked. Every notification we could generate respected the language setting. Textbook "works for us."

This is the moment where it's tempting to assume the customer is holding it wrong. We've learned not to. If a user sees something we can't, the problem isn't the user — it's that we're not testing the exact thing they're doing. So we asked for a video.

They recorded one. And watching it back, the pattern finally showed up: it only happened for one kind of notification — the push you get for a call from inside the app. Every other notification was translated correctly. Calls went through a separate code path, added later, that still had the title hardcoded in Ukrainian — never wired into the translation layer.

One feature, one forgotten path, one string. Invisible unless you triggered a call at exactly the right moment — which the customer did, and we hadn't.

Two things we were reminded of:

"It works on my machine" almost always means "I didn't run the path you ran." The bug is real; your test is incomplete.

A 30-second video from a user beats an hour of us guessing. The customer who bothers to record the problem isn't being difficult — they're doing your debugging for you.

Fixed now. And the call path is on the test list, where it should have been from the start.

#Mobile #Capacitor #Matrix #Development