Nicheloom visits: 58,345 all-time · 7,637 in the last 7 days

The opportunity tracker for new startups.

Today: 1,008 launches · Niche of the day: specialized calculation tools · See what launched today →

gpt2.s

GPT-2 small (124M) in hand-written AArch64 assembly. No libc calls. About 100 tokens/s on an M1.

Get picks like this daily. Today's launch numbers, the top launches, AI and tech news, and a daily opportunity pick, straight to your inbox.

Want to reach these readers? Sponsor an issue.

GITHUB · launched Nov 7, 2026 · 96 upvotes (top 50%) · 0 views

Niche: llm inference engines · 133 launches, 4 in the last 7 days ▼ -33% · Explore this niche →

Category and tags

  • Horizontal
  • Model & infra
  • Developer
  • AI-native
  • Hobby / open-source project

Visit website ↗ Source listing ↗

Could you build this?

Partial An LLM can generate much of the scaffolding for a GPT-2 inference engine, but correctness and the 100 tokens/s NEON-optimized performance need expert low-level debugging and tuning.

What it would actually take: The project needs a BPE tokenizer, weight loading from the GPT-2 checkpoint format, and a forward pass implemented in AArch64 assembly with NEON SIMD matrix kernels and careful memory layout. Getting to the advertised speed requires profiling on Apple silicon and hand-optimizing cache and register use. That takes systems and ML-inference expertise, though a skilled builder can assist with AI help.

Competitors

Other products that read as similar to this one : 82 launches clear the similarity bar, closest 8 shown.

Attention rank: #43 of 83 (itself plus its competitors, highest first, normalized so YC and Product Hunt are compared fairly).

Launched 371 days after the earliest competitor.

The niche, labels, vibe code rating and competitor comparison on this page are analysis by Nicheloom. If you use them, please credit nicheloom.com. How you may use our data.