tweetindex

Snorkel AI

@SnorkelAI · Redwood City, California · joined 11 Jul 2019

Frontier AI Data Lab advancing AI through better data

17 744Followers
366Following
2 093Posts total
8.2KViews on collected posts

Against accounts of the same size

9 posts from the last 90 days, next to the 10K–100K follower range. reaches fewer people than peers of the same size.

Median views359this account1 637median for 10K–100K
Reach, %2.02%this account5.01%median for 10K–100K
Engagement, %1.54%this account1.81%median for 10K–100K
MetricThis accountMedian for 10K–100KRatio
Median views per post3591 6370.22×
Reach (views ÷ followers)2.02%5.01%0.40×
Engagement rate1.54%1.81%0.86×

Others in this range →   Compare with another account →   How these benchmarks are built →

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

3.6K17 Jul
34
21920 Aug
8921 Aug
1425 Aug
1.4K1 Sep
1.5K2 Sep
9763 Sep
3594 Sep

Last 9 collected posts, oldest on the left. The scale is logarithmic: one post can outrun the rest a hundred times over.

Engagement rate per post

1.55%17 Jul
0.00%
1.37%20 Aug
1.12%21 Aug
0.00%25 Aug
2.15%1 Sep
2.99%2 Sep
2.97%3 Sep
2.79%4 Sep

Reactions — likes, reposts, replies and quotes — divided by views. Median for 10K–100K accounts is 1.81%.

What the audience does

Likes64.9%122 in total
Reposts14.4%27 in total
Replies10.6%20 in total
Quotes2.1%4 in total
Bookmarks8.0%15 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.

Latest posts

How do you turn an open research problem into a company? @vincentsunnchen joins the "Papers to Production" panel at the Science x AI Summit at @Caltech, Sept 11, hosted by @SAIRfoundation. Register here: https://t.co/O1Qx5hXLRU https://t.co/er5L1juSkI 359 views · 8 likes · 2 reposts · 0 replies 04 Sep 2026 Yesterday's virtual Reading Group brought in @yuan_mengq43669 of @XLangNLP to break down OSWorld 2.0: Why Long-Horizon Computer-Use Agents Still Fail Four Out of Five Tasks. Recording/transcript: https://t.co/hSlmCHmKMU The best AI agent evaluated in the paper completes only htt 976 views · 19 likes · 3 reposts · 6 replies 03 Sep 2026 We evaluated Fable 5.1 against Opus 5 on Terminal-Bench+, our proprietary set of frontier coding tasks. Fable 5.1 leads on debugging (87%) and games (88%), using 58% fewer tokens per successful run. Failure analysis shows the errors cluster in execution, not reasoning. Full http 1.5K views · 28 likes · 7 reposts · 8 replies 02 Sep 2026 Claude Fable 5.1 just topped the Senior SWE-Bench leaderboard, matching Fable 5's performance at roughly half the cost. Our team also found it tends to over-engineer at higher effort tiers. Full results: https://t.co/sACFBBALNS 1.4K views · 25 likes · 4 reposts · 1 replies 01 Sep 2026 @SnorkelAI Sounds like a great event, especially with the focus on open benchmarks and practical frontier research. Looking forward to seeing what new evaluation approaches and datasets come out of the summit. 14 views · 0 likes · 0 reposts · 0 replies 25 Aug 2026 @SnorkelAI @_leann_chen! (https://t.co/iI5tgtMmRs ) 89 views · 1 likes · 0 reposts · 0 replies 21 Aug 2026 Pretraining data is often curated by quality scores or semantic similarity, but neither necessarily tells us whether selected docs give a model distinct learning signals. In a small GPT-2 test, we took 10 FineWeb documents spanning different FineWeb-Edu quality scores and asked: 219 views · 1 likes · 0 reposts · 1 replies 20 Aug 2026 @SnorkelAI En iyi zihinler @grik bu paylaşımdan hangilerri 34 views · 0 likes · 0 reposts · 0 replies 17 Jul 2026 We're inviting researchers and practitioners building open benchmarks to present their work in several dedicated poster sessions at our first-ever Frontier Data Summit. Taking place on October 8 in San Francisco, the one-day event will bring together some of the best minds in ht 3.6K views · 40 likes · 11 reposts · 4 replies 17 Jul 2026

Similar accounts