tweetindex
EN

Georgia Channing ✓

@cgeorgiaw

AI4Science @ 🤗, PhD @OxfordTVG — science is a candle in the dark

7 430Followers
477Following
1 229Posts total
18.2KViews on collected posts

Latest posts

@cgeorgiaw Also science is actually hard and slop seldom translates to reproducible outcomes, communicable protocols, and powerful inference. And by inference I mean the real statistical meaning, not whatever tf ai vibe-ppl think it is 285 views · 3 likes · 0 reposts · 0 replies Open on X →
And we’re seeing progress already. More scientific datasets are being shared on @huggingface. We’ve built a space to track this trend, and it’s growing fast. Leaders include @AIatMeta, @PolymathicAI, @entalpic_ai, and academic groups led by @BoWang87 and @kmjablonka. 645 views · 11 likes · 0 reposts · 1 replies Open on X →
If you want to see this happen in real-time, check these out: Position Paper: https://t.co/OYSt2i1h46 Science Release Heatmap: https://t.co/HJs6Ag3bQk Discord: https://t.co/E4AWDcsF2y huggingscience-org: https://t.co/HJs6Ag3bQk https://t.co/b0LnI6kt3K
630 views · 15 likes · 1 reposts · 0 replies Open on X →
To change that, we built a new community: hugging-science. 🤝 1 Discord server 🤝 5 major projects launched (so far) 🤝 700 contributors 🤝 5 Spaces 🤝 1 position paper In just a month, 4,000+ people have joined to build ML for science together. With crazy diverse expertise. 785 views · 14 likes · 1 reposts · 1 replies Open on X →
So hope isn’t lost. We can learn to communicate better, and that starts with sharing spaces. Machine learners and scientists need to share the same reading lists, the same problems, the same social spaces. 717 views · 10 likes · 0 reposts · 1 replies Open on X →
@evijit and I wrote a whole position on this, but here's the gist: 👩‍💻 Machine learner: "All your problems can be solved with my ML." 🧑‍🔬 Scientist: gives a 30-minute jargon-filled talk MLer zones out → "So...what do I optimize?" Scientist gives up. Nothing happens. https:
5.1K views · 15 likes · 7 reposts · 2 replies Open on X →
The reason AI isn't transforming science faster isn't: ❌ not enough compute ❌ not enough data ❌ not good enough models ❌ not enough AGI The real problem is social. 10K views · 59 likes · 7 reposts · 7 replies Open on X →

Growth & engagement

How the posts we collected actually performed: views and reaction rate post by post, what the audience did with them, and where the follower count goes.

Views per post

10K6 Oct
5.1K
717
785
630
645
285

Last 7 collected posts, oldest on the left. The scale is logarithmic: one post can outrun the rest a hundred times over.

Engagement rate per post

0.75%6 Oct
0.49%
1.53%
2.04%
2.54%
1.86%
1.05%

Reactions — likes, reposts, replies and quotes — divided by views.

What the audience does

Likes64.1%127 in total
Reposts8.1%16 in total
Replies6.1%12 in total
Quotes1.5%3 in total
Bookmarks20.2%40 in total

Share of every reaction we collected for this account. Replies mean argument, reposts mean endorsement, bookmarks mean the post was worth keeping.

The follower curve appears once this account has two daily snapshots — we take one a day, and this one is on its first.

Similar accounts