AE
6.5K posts

AE
@AlphaExtrema
All tweets are Opinions | LLMs are going to profile you and are now getting memory | Posts will be purged anytime | If I smell FURU I block.
Earth Entrou em Aralık 2020
57 Seguindo536 Seguidores

@AlphaExtrema @omshanti08 could be absolutely wrong tbh, they just seem primed
English

AE retweetou

Just implemented Google’s TurboQuant in MLX and the results are wild!
Needle-in-a-haystack using Qwen3.5-35B-A3B across 8.5K, 32.7K, and 64.2K context lengths:
→ 6/6 exact match at every quant level
→ TurboQuant 2.5-bit: 4.9x smaller KV cache
→ TurboQuant 3.5-bit: 3.8x smaller KV cache
The best part: Zero accuracy loss compared to full KV cache.

Google Research@GoogleResearch
Introducing TurboQuant: Our new compression algorithm that reduces LLM key-value cache memory by at least 6x and delivers up to 8x speedup, all with zero accuracy loss, redefining AI efficiency. Read the blog to learn how it achieves these results: goo.gle/4bsq2qI
English
AE retweetou

@AuclairsDad @USSOCOM markets also dont trust it .. else it would be trending up
English

Short $QCOM CEO has no vision
FinancialJuice@financialjuice
Arm are to sell its own chips for the first time in a bid for AI revenue. $ARM
Español





