SIGNAL GRIDv0.1

Launch HN: General Instinct (YC P26) – Frontier models on edge devices

1 sources1 storiesFirst seen 6/5/2026Score28Breakthrough
Single SourceContradictory Claims
CoverageRecencyEngagementVelocityBignessConfidenceClipability
Bigness
28
Coverage
13
Recency
88
Engagement
13
Velocity
0
Confidence
49
Clipability
60
Polarization
0
Claims
5
Contradictions
1
Breakthrough
100

Sentiment Mix

Positive0%
Neutral100%
Negative0%

Geography

North America

Expert Signals

guanming0717

author1 mention

Hacker News

source1 mention

AI-Generated Claims

Generated from linked receipts; click sources for full context.

Launch HN: General Instinct (YC P26) – Frontier models on edge devices.

Supported by 1 story

Hey HN, Guanming and Bill here from General Instinct (https://general-instinct.com/).After years of working in robotics, we kept running into the same problem: the best models never fit the hardware we actually had available.The models that performed best were usually designed around datacenter assumptions: large GPUs, lots of memory bandwidth, and reliable network access.

Supported by 1 story

But most physical systems have the opposite constraints.That led us down the path of figuring out how much of a frontier model could be preserved while still making it practical to run on edge hardware.As part of that work, we recently open sourced InstinctRazor (https://github.com/General-Instinct/InstinctRazor)One result we're excited about is compressing Qwen3.5-122B-A10B, a roughly 245 GB BF16 MoE model, into a 48 GiB GGUF.

Supported by 1 story

The resulting model is actually smaller than Gemma-4-26B-A4B while outperforming it on benchmarks like MMLU-Pro and GPQA-D etc.

Supported by 1 story

we preserve the parts that are always active (router, norms,...

Supported by 1 story

Claim Contradictions

negation mismatch

A: Launch HN: General Instinct (YC P26) – Frontier models on edge devices.

B: Hey HN, Guanming and Bill here from General Instinct (https://general-instinct.com/).After years of working in robotics, we kept running into the same problem: the best models never fit the hardware we actually had available.The models that performed best were usually designed around datacenter assumptions: large GPUs, lots of memory bandwidth, and reliable network access.

Paper to Product Links

Related Events

Timeline (1 stories)

Receipts (1)

Bias Snapshot

Center
Left 0%Center 100%Right 0%