Percepta Blog company blog
Executing programs inside transformers with exponentially faster inference
We build a computer inside a transformer — executing arbitrary C programs for millions of steps with exponentially faster inference via 2D attention heads.
- Points
- 330
- HN comments
- 130
- Submissions
- 1
- Cited by
- 4 commenters
AI & Machine LearningSoftware Engineering2026
On Hacker News
| Submitted | Title on HN | Points | Comments |
|---|---|---|---|
| Mar 12, 2026 | Executing programs inside transformers with exponentially faster inference by u1hcw9nx | 330 | 130 |
Cited in comments
Linked from 4 Hacker News comments by 4 different people, from Mar 2026 to Oct 2026.
Most recent:
- giveexamples · Oct 29, 2026
- armcat · Sep 2, 2026
- Gomotono · May 29, 2026
- hackyhacky · Mar 13, 2026
