Haven Feng ✓

@HavenFeng

Postdoc @berkeley_ai, building Impossible. Prev. @MPI_IS @AdobeResearch @uni_tue Interested in recreating the physical world and the intelligence to do so.

5 228Followers

1 119Following

259Posts total

1.1MViews on collected posts

Latest posts

Haven Feng ✓@HavenFeng16 Jul

@HavenFeng thinking like a physicist is all you need? 6.5K views · 46 likes · 0 reposts · 2 replies Open on X →

Haven Feng ✓@HavenFeng16 Jul

@guanningzeng @Jiani_Wang_ @wenjie_ma @shaofeng_y27736 @arcprize 16.6K views · 98 likes · 0 reposts · 3 replies Open on X →

Haven Feng ✓@HavenFeng16 Jul

[schema]'s saturation of the ARC-AGI-3 public set is only a starting point. There is much more to explore! Full blog: https://t.co/A13KTXczUX Agent traces: https://t.co/Gc6hP4OitW Amazing team effort with @guanningzeng, @Jiani_Wang_, @wenjie_ma, @shaofeng_y27736, https://t.co/b

25.9K views · 223 likes · 14 reposts · 3 replies Open on X →

Haven Feng ✓@HavenFeng16 Jul

[schema] handles the state and mechanism in one editable program, a symbolic world model. It designs experiments to verify hypotheses, backtests the program against history, and plans inside its world at zero action cost. 24.7K views · 143 likes · 6 reposts · 2 replies Open on X →

Haven Feng ✓@HavenFeng16 Jul

ARC-AGI-3 gives an agent a 64×64 grid plus legal actions, no rules, stated goal, or reward. The agent must discover both what the world is and how it works like a physicist: 1. State grounding -> identify objects, relations, and goals. 2. Mechanism discovery -> infer how these h

31.3K views · 203 likes · 4 reposts · 7 replies Open on X →

Haven Feng ✓@HavenFeng16 Jul

Today, we’re introducing [schema]: a harness reaching 99% RHAE with Opus 4.8 + Fable 5 and 95.35% with GPT-5.6 Sol on ARC-AGI-3 Public set. [schema] makes an LLM think like a physicist. 🧵 https://t.co/U0C0JFj62e

0:47

950K views · 2.5K likes · 297 reposts · 125 replies Open on X →

Against accounts of the same size

6 posts from the last 90 days, next to the under 10K follower range. shown widely, but few of those viewers react.

Median views25 306this account4 175median for under 10K

Reach, %484.04%this account250.82%median for under 10K

Engagement, %0.66%this account1.41%median for under 10K

Metric	This account	Median for under 10K	Ratio
Median views per post	25 306	4 175	6.06×
Reach (views ÷ followers)	4.8× audience	2.5× audience	1.93×
Engagement rate	0.66%	1.41%	0.47×

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

950K16 Jul

31.3K

24.7K

25.9K

16.6K

6.5K

Last 6 collected posts, oldest on the left. The scale is logarithmic: one post can outrun the rest a hundred times over.

Engagement rate per post

0.32%16 Jul

0.69%

0.62%

0.94%

0.61%

0.74%

Reactions — likes, reposts, replies and quotes — divided by views. Median for under 10K accounts is 1.41%.

What the audience does

Likes51.5%3 207 in total

Reposts5.2%321 in total

Replies2.3%142 in total

Quotes2.3%143 in total

Bookmarks38.7%2 409 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

Against accounts of the same size

Growth & engagement

Views per post

Engagement rate per post

What the audience does

Similar accounts