Rising2 sources· last seen 10h ago· first seen 23h ago

AMD acquires Taalas, a startup that bakes AI models directly into silicon

AMD is buying Canadian startup Taalas, which hard-codes model weights directly into inference chips. That makes them extremely fast but locks each chip to a single model. A demo chip hit over 16,000 tokens per second per user running Llama 3.1-8B. Google is reportedly working on a similar approach f

Lead: The DecoderBigness: 32amdacquirestaalasbakesdirectly
📡 Coverage
50
2 news sources
🟠 Hacker News
0
🔴 Reddit
0
📈 Google Trends
0
Full methodology: How scoring works

Receipts (all sources)

AMD is buying Canadian startup Taalas, which hard-codes model weights directly into inference chips. That makes them extremely fast but locks each chip to a single model. A demo chip hit over 16,000 tokens per second per user running Llama 3.1-8B. Google is reportedly working on a similar approach f

[AINews] AMD buys Taalas
RSS · Latent Space · 23h ago
score 135

The Inference Inflection is HEATING up.