Reiner Pope ✓

@reinerpope · Mountain View, California · joined 02 Jul 2011

CEO and founder, @MatXComputing, developing high throughput chips tailored for LLMs

18 763Followers

458Following

185Posts total

3.3MViews on collected posts

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

4.5K4 Oct

2.7K5 Oct

3.1M24 Feb

72.3K15 Apr

148K20 Apr

32.8K30 May

8.7K31 May

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

0.58%4 Oct

0.67%5 Oct

0.09%24 Feb

0.55%15 Apr

0.48%20 Apr

0.50%30 May

0.44%31 May

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

What the audience does

Likes52.9%3 453 in total

Reposts4.8%312 in total

Replies2.5%164 in total

Quotes1.6%103 in total

Bookmarks38.3%2 500 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

2^100 sounds impossibly large to count to. But $1B of chips and datacenter space and power can get you there in 2 years. 2^100 is only large if you don't allow parallelism. 8.7K views · 28 likes · 1 reposts · 8 replies 31 May 2026 there is a pretty (in?)famous story here from harvey friedman on ultrafinitism friedman goes up to esenin-volpin, a hardline ultrafinitist, after one of his lectures and then posits to him that if the number 2^100 doesn't exist, then surely he believes that somewhere on the line 32.8K views · 146 likes · 8 reposts · 6 replies 30 May 2026 Intelligence per picojoule, with @itsclivetime and @dylan522p (0:00) Intro (1:22) What is codesign? (2:49) Codesign example: Swish vs ReLU (4:22) Are DeepSeek papers codesign? (6:45) Predicting where ML research will go (8:06) Should researchers hate your chips? (9:34) Can you h 148K views · 624 likes · 55 reposts · 12 replies 20 Apr 2026 I chatted with @ysmulki about MatX, chip design and where silicon designed for LLMs is headed (8:17) Tightly coupling SRAM and HBM on one chip (14:03) More MoE FLOPS, smaller KV cache load (16:08) Numerics: from 32-bit to 4-bit (19:02) Targeting both training and inference https 72.3K views · 326 likes · 46 reposts · 16 replies 15 Apr 2026 We’re building an LLM chip that delivers much higher throughput than any other chip while also achieving the lowest latency. We call it the MatX One. The MatX One chip is based on a splittable systolic array, which has the energy and area efficiency that large systolic arrays 3.1M views · 2.3K likes · 199 reposts · 121 replies 24 Feb 2026 We serialize key/value maps and send them over the Internet or write them to disk all the time. Why don't we keep them in hashmap layout on the wire, so the receiver doesn't have to rebuild the table? https://t.co/1zmJD59OpE 2.7K views · 17 likes · 1 reposts · 0 replies 05 Oct 2025 Sometimes a more complex algorithm actually *is* faster in practice. This seems to be true of cuckoo hashing, provided you engineer it carefully. https://t.co/smB0Jx2f3w 4.5K views · 22 likes · 2 reposts · 1 replies 04 Oct 2025

Growth & engagement

Views per post

Engagement rate per post

What the audience does

Latest posts

Similar accounts