Tony L.
tlint.bsky.social
Data scientist, cheminformatics, drug discovery, just trying to get by
👤 Joined on 13 January 2025
32
Followers
62
Following
49
Total Posts
📜 Kullanıcı Geçmişi
- 13 Oca 2025, 01:43
🎉 Bluesky'e katıldı - İlk kullanıcı adı:
@tlint.bsky.social
🚫 Blokladığı Kullanıcılar
❤️ Beğendiği Gönderiler
As someone who works all sides of this problem: generating the data for ML, doing the ML, using the predictions of the ML, and paying for all of it (so I look to minimize spend over the cycle), one needs to be very careful about what data one is using, and how one builds the tools.
Show PostI'm not saying this is impossible. But pretending that good data is easy to get and meaningful is a very significant error in judgement. ML accelerates something we can already do: extract patterns from data. That is the EASY problem. Good, clean data? That's the hard problem.
Show PostAvoiding learning/thinking doesn't help you learn/think better in the future.
Show PostThe AI tools essentially undermine the entire basis of education. Although I disagree and think it is wrong, a reasonable case can be made for experts to use AI tools to increase productivity. But no case can be made for students to use it. Writing is hard. The only way to get better is to do it.
Show PostResearch in a university context is teaching. You do it with students, one on one, or at least one with a small number. You mentor how to think deeply about solving problems that have never been solved. It is the PUREST form of teaching. The rest is, comparably, content delivery.
Show Post