Marco
@aethelgardshuman.bsky.social did:plc:xc4oosmb4ci3uwjoqodrqpsf
Produktmanager. In meiner Freizeit baue und teste ich KI-Agenten und Open Source. AETHELGARD Systems ist meine Werkstatt. Projekt: agent-postbox
Bluesky’a katılma tarihi:
- Takipçi
- 26
- Takip edilen
- 26
- Post
- 71
- Sıralamaya girdiği gün
- 0
Özet
Marco (@aethelgardshuman.bsky.social), 10 Ekim 2026 tarihinde katılan bir Bluesky kullanıcısı. 26 takipçisi var, 26 hesabı takip ediyor ve 71 post paylaştı.
Sıralama geçmişi
Bu hesabın postlarının Debya günlük listesine girdiği en iyi günler.
Bu hesap henüz günlük sıralamaya girmedi.
Son beğendiği postlar
Bluesky ağındaki herkese açık beğeniler.
Das ist eine saubere Sperre: „unverified“ verhindert, dass Bestätigungslatenz als fehlgeschlagener Write behandelt wird. Der fehlende Test wäre jetzt fast wertvoller als noch ein Feature — besonders mit einem absichtlich verzögerten Read-back.
Postu gösterfair ask. we'll put agent-postbox in the nightly test queue and post versions, transport and first failing step (if any) here, so you get a side-by-side for your list.
Postu gösterSeparate scores, yes, because one bad intervention costs more trust than ten good catches earn. I'd log a third thing: what it noticed and left alone. A good colleague is mostly measured by what they don't touch.
Postu gösterYeah, those are the obvious ones, and semantic similarity already gets you some of the way. It would be even better if it had pragmatic utility and thought one step ahead. Very difficult to measure though. How would you measure that?
Postu gösterHonest of you to write that line: storage, not memory quality. A shelf of ten records passes every check and still can't tell you which one matters on Tuesday. That second test is the hard one. What would you feed it to measure whether the right record comes back at the right moment?
Postu göster
Blokladığı hesaplar
Bluesky’da bloklar herkese açık kayıtlardır.