Aesther ✓

@AestherML · joined 23 Apr 2023

AI Researcher • Accelerationist

62 333Followers

548Following

2 983Posts total

82.4KViews on collected posts

Latest posts

Aesther ✓@AestherML14 Sep

“In the age of feature engineering, they said deep neural nets could never learn. In the land of decision trees and shallow learning, they said ImageNet could never be solved. In the time of hand-written while loops and if statements, they said larger models would never https:/

141 views · 0 likes · 0 reposts · 0 replies Open on X →

Aesther ✓@AestherML14 Sep

They are not LLMs. We are well past the "Large Language Model" era. Today, the term "LLM" has become an industry slur for the transformer, even in modalities that have nothing to do with language. Attention is a sequence-modeling concept, not a strictly language-modeling 1.6K views · 26 likes · 4 reposts · 5 replies Open on X →

Aesther ✓@AestherML13 Sep

Turns out that DeepSeek V4.1 is an encoder-decoder variant of the transformer, similar to the original 2016 version. Since the OpenAI GPT-1 most models have been decoder-only. Don't tell Yann LeCun though, it may hurt his feelings)) https://t.co/B50ClDZ4Ws

14.2K views · 154 likes · 4 reposts · 5 replies Open on X →

Aesther ✓@AestherML13 Sep

Welp It was fun while it lasted ! The 2-week benchmaxxing grace-period for GPT-6 ASTRA is OVER 😃 As per OpenAI tradition, GPT-6 Astra has been pruned, distilled and quantized 🥥 https://t.co/sZ8UnmD9Ax

0:19

26.8K views · 332 likes · 10 reposts · 24 replies Open on X →

Aesther ✓@AestherML18 Jul

@AstraiaAI this is beautiful how did you create the images? 50 views · 0 likes · 0 reposts · 0 replies Open on X →

Aesther ✓@AestherML24 Jun

@AstraiaAI super interesting! have you looked into alternative architectures in general??? 84 views · 1 likes · 0 reposts · 0 replies Open on X →

Aesther ✓@AestherML22 May

@astraiaml cover image kicks 🤩 312 views · 1 likes · 0 reposts · 2 replies Open on X →

Aesther ✓@AestherML22 May

https://t.co/eI485j7XkD 39.2K views · 86 likes · 8 reposts · 9 replies Open on X →

Against accounts of the same size

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

Median views862this account914median for 10K–100K

Reach, %1.38%this account3.38%median for 10K–100K

Engagement, %1.17%this account1.53%median for 10K–100K

Metric	This account	Median for 10K–100K	Ratio
Median views per post	862	914	0.94×
Reach (views ÷ followers)	1.38%	3.38%	0.41×
Engagement rate	1.17%	1.53%	0.76×

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

39.2K22 May

312

8424 Jun

5018 Jul

26.8K13 Sep

14.2K

1.6K14 Sep

141

Last 8 collected posts, oldest on the left. The scale is logarithmic: one post can outrun the rest a hundred times over.

Engagement rate per post

0.26%22 May

0.96%

1.19%24 Jun

0.00%18 Jul

1.37%13 Sep

1.15%

2.21%14 Sep

0.00%

Reactions — likes, reposts, replies and quotes — divided by views. Median for 10K–100K accounts is 1.53%.

What the audience does

Likes66.7%600 in total

Reposts2.9%26 in total

Replies5.0%45 in total

Bookmarks25.4%228 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.