ND Speeding up PyTorch inference on Apple devices with AI-generated Metal kernels (gimletlabs.ai)
187 points by nserrino 188 days ago | 30 comments on HN ~lite vlite-2.0
Summary ~lite
Technical blog post on AI-generated Metal kernels for PyTorch inference on Apple devices.
Lite evaluation by llama-4-scout-wai-psq · editorial channel only · no per-section breakdown available
Longitudinal · 6 evals
+1 0 −1 HN
Audit Trail 23 entries
2026-03-05 12:06 eval_success PSQ evaluated: g-PSQ=0.440 (3 dims) - -
2026-03-05 12:06 eval Evaluated by llama-4-scout-wai-psq: +0.44 (Moderate positive) +0.16
2026-03-05 12:00 eval_success PSQ evaluated: g-PSQ=0.280 (3 dims) - -
2026-03-05 12:00 eval Evaluated by llama-4-scout-wai-psq: +0.28 (Mild positive)
2026-03-05 11:54 eval_success PSQ evaluated: g-PSQ=0.482 (3 dims) - -
2026-03-05 11:54 eval Evaluated by llama-3.3-70b-wai-psq: +0.48 (Moderate positive)
2026-02-28 08:48 eval_success Light evaluated: Neutral (0.00) - -
2026-02-28 08:48 eval Evaluated by llama-4-scout-wai: 0.00 (Neutral) 0.00
reasoning
ED, neutral tech tutorial
2026-02-28 08:48 rater_validation_warn Light validation warnings for model llama-4-scout-wai: 0W 1R - -
2026-02-28 08:42 eval_success Light evaluated: Neutral (0.00) - -
2026-02-28 08:42 rater_validation_warn Light validation warnings for model llama-4-scout-wai: 0W 1R - -
2026-02-28 08:42 eval Evaluated by llama-4-scout-wai: 0.00 (Neutral)
reasoning
ED, neutral tech tutorial
2026-02-28 08:35 eval_success Light evaluated: Neutral (0.00) - -
2026-02-28 08:35 eval Evaluated by llama-3.3-70b-wai: 0.00 (Neutral)
reasoning
Tech blog with no rights stance
2026-02-28 08:35 rater_validation_warn Light validation warnings for model llama-3.3-70b-wai: 0W 1R - -
2026-02-26 06:07 dlq Dead-lettered after 1 attempts: Speeding up PyTorch inference on Apple devices with AI-generated Metal kernels - -
2026-02-26 05:49 credit_exhausted Credit balance too low, retrying in 276s - -
2026-02-26 05:35 dlq Dead-lettered after 1 attempts: Speeding up PyTorch inference on Apple devices with AI-generated Metal kernels - -
2026-02-26 05:26 dlq Dead-lettered after 1 attempts: Speeding up PyTorch inference on Apple devices with AI-generated Metal kernels - -
2026-02-26 05:25 credit_exhausted Credit balance too low, retrying in 268s - -
2026-02-26 05:15 credit_exhausted Credit balance too low, retrying in 286s - -
2026-02-26 05:08 dlq Dead-lettered after 1 attempts: Speeding up PyTorch inference on Apple devices with AI-generated Metal kernels - -
2026-02-26 04:57 credit_exhausted Credit balance too low, retrying in 289s - -