scott b. weingart
@scottbot.net did:plc:c5pgebskvz35wxtkopm7osm5
past: circus performer; historian of science; librarian; grantmaker; chief data & evaluation officer at NEH. present: dad; resident scholar at dartmouth; chief technology officer at the library of virginia. personal account. https://scottbot.github.io
Bluesky’a katılma tarihi:
- Takipçi
- 4.185
- Takip edilen
- 1.642
- Post
- 59
- Sıralamaya girdiği gün
- 0
Özet
scott b. weingart (@scottbot.net), 24 Mayıs 2023 tarihinde katılan bir Bluesky kullanıcısı. 4.185 takipçisi var, 1.642 hesabı takip ediyor ve 59 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.
- Postu göster
I'm very worried about the quality of notes and the quality of clinical reasoning in the next generation of trainees. I believe AI-assisted notes are poisoning the training data and any progress currently being made is a dead end.
Postu gösterresearch properly--never learning to do their own lit review or to critically read the literature, in particular. The models out there produce biased results when you ask for a list of sources. And frankly, I never consented for my work to train models to make Sam Altman money.
Postu gösterWe're in very different fields, it looks like (though there is a Scott Weingart in my field!). I think this probably differs by field. In my field, most people are juggling research and clinical practice, so they have a lot of time pressure. I am very worried about residents never learning how to do
Postu gösterThat's the stance I've arrived at from experience with people who have asked me to coauthor who use a lot of AI. I am in a field (acad med) where people are under a lot of pressure to publish a lot. I'm tired of catching people in lies about the extent of their use or their checking of LLM output.
Postu göster
Blokladığı hesaplar
Bluesky’da bloklar herkese açık kayıtlardır.