Stuart Sul ✓

@stuart_sul

ml @Cursor_ai (now @SpaceXAI), cs phd @Stanford, @HazyResearch

4 105Followers

162Following

135Posts total

104.8KViews on collected posts

Against accounts of the same size

9 posts from the last 90 days, next to the under 10K follower range. below its peers on both reach and engagement.

Median views3 081this account2 290median for under 10K

Reach, %75.05%this account159.23%median for under 10K

Engagement, %0.51%this account1.64%median for under 10K

Metric	This account	Median for under 10K	Ratio
Median views per post	3 081	2 290	1.35×
Reach (views ÷ followers)	75.05%	159.23%	0.47×
Engagement rate	0.51%	1.64%	0.31×

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

18.1K5 Aug

3.1K

11.5K

3.7K

59.7K

2.4K

1.2K

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

1.28%5 Aug

1.30%

1.27%

0.51%

0.91%

0.41%

0.17%

0.32%

0.35%

Reactions — likes, reposts, replies and quotes — divided by views. Median for under 10K accounts is 1.64%.

What the audience does

Likes55.6%568 in total

Reposts5.8%59 in total

Replies2.3%23 in total

Quotes0.9%9 in total

Bookmarks35.5%363 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

@stuart_sul @bilaltwovec @HazyResearch @pangram ai? (It's a good post, sir) 2K views · 5 likes · 1 reposts · 1 replies 05 Aug 2026 · Open on X →

@stuart_sul @srush_nlp @HazyResearch Nice, I have a similar one! (https://t.co/PWZSvQkPZm) 1.2K views · 4 likes · 0 reposts · 0 replies 05 Aug 2026 · Open on X →

@stuart_sul @HazyResearch well yes and no, please everybody!! pivot to verifiable and proved PL & TT approaches to making programs and gpu programs, not adding more arbitrary obscurity and interface layers 2.4K views · 4 likes · 0 reposts · 0 replies 05 Aug 2026 · Open on X →

(6/6) Read the full blog post, written with @HazyResearch: https://t.co/oz1LD0qDOg 59.7K views · 215 likes · 17 reposts · 7 replies 05 Aug 2026 · Open on X →

(5/6) We're no longer working with a dumb executor that needs every gap specified, but an intelligent executor that reads underspecified specs and does the right thing. It may be fuzzy, but the ideas travel better. 3.7K views · 33 likes · 0 reposts · 1 replies 05 Aug 2026 · Open on X →

(4/6) Abstraction as a cognitive offloader is going obsolete. By induction, it appears all the beloved CUDA DSLs are next in line to retire. 11.5K views · 52 likes · 4 reposts · 3 replies 05 Aug 2026 · Open on X →

(3/6) For instance, we built a library for writing megakernels last year (https://t.co/PO6Cu37FsP). Writing one from scratch was miserable, so we accepted limits like coarse shared memory granularity to overcome the complexity. Today, that tradeoff has become unnecessary. 3.1K views · 34 likes · 4 reposts · 1 replies 05 Aug 2026 · Open on X →

(2/6) Why? Because the complexity those layers were hiding is now directly manageable, so the intermediate layers turn into pure overhead. 3K views · 37 likes · 1 reposts · 1 replies 05 Aug 2026 · Open on X →

(1/6) If a prompt can generate every layer of the stack, what are programming abstractions worth? GPU programming is at a weird inflection point. For decades we stacked layers of abstraction to ease our cognitive load. This year, we find ourselves constantly deleting them. https 18.1K views · 184 likes · 32 reposts · 9 replies 05 Aug 2026 · Open on X →