hashyy
A full, unpruned 177B MoE model at 11.5 tok/s on one 12 GB GPU. Expert streaming for llama.cpp, with the measurement harness and every dead end.
Get picks like this daily. The day's top launches, AI/tech news, and a weekly opportunity spotlight — straight to your inbox.
This is 1 of 189 launches in GPU acceleration and compute tools — see how it stacks up on momentum and crowding →
182 other launches read as similar to this one →
Details
- External ID
- 1402547731
- Source
- GITHUB
- Company
- —
- Product
- hashyy
- Website domain
- github.com
- Launched
- Oct. 3, 2026
- Cohort
- —
- Upvotes
- 18
- Upvotes percentile
- 0.42
- Tags
- —
Enrichment
- Niche
- GPU acceleration and compute tools
- Vertical
- Horizontal
- Function
- Model & infra
- Audience
- Developer
- AI stance
- AI-native
- Project type
- Hobby / open-source project
- Normalized one-liner
- expert streaming engine to run large moe models on low-vram gpus
- Manually corrected
- False
Could you build this?
No Implementing expert streaming for a massive 177B Mixture-of-Experts LLM to run at double-digit token rates on consumer GPU memory requires advanced low-level GPU kernel programming and systems engineering.
What it would actually take: Requires deep C++/CUDA/Metal engineering integrated into llama.cpp or custom inference runtimes, designing zero-copy PCIe DMA memory streaming pipelines, NVMe caching, and custom quantized matrix multiplication kernels. Developers need deep expertise in GPU memory hierarchies, asynchronous compute pipelines, and hardware-specific IO bandwidth bottlenecks.
Competitors
Other products that read as similar to this one — 182 launches clear the similarity bar, closest 8 shown.
Attention rank: #110 of 183 (itself plus its competitors, highest first — normalized so YC and Product Hunt are compared fairly).
Launched 330 days after the earliest competitor.
- A 6M-token movable window on a single 46GB GPU · hn · 2026-07-28 · 7 upvotes · similarity 0.44
- The cheapest GPU cloud · hn · 2026-09-03 · 63 upvotes · similarity 0.44
- Fine-tune an 8B model on a 4 GB laptop GPU · hn · 2026-08-04 · 139 upvotes · similarity 0.43
- splish · github · 2026-09-26 · 12 upvotes · similarity 0.43
- Taliesin · hn · 2026-06-04 · 10 upvotes · similarity 0.42
- LumaDock - Blackwell GPU VPS · ph · 2026-09-14 · 2 upvotes · similarity 0.42
- Hy4 preview · ph · 2026-08-29 · 215 upvotes · similarity 0.42
- RunNburn · hn · 2026-07-30 · 11 upvotes · similarity 0.41
Other launches for this product
- No other launches for this product.