Loading activity…
@dylan522p
SemiAnalysis AI Infrastructure Research and Consulting DMs are open for consulting, quotes, or to talk shop, Opinions my own
AI
A user described FlashAttention as a 'mega kernel' in an October 20 post.
AI
Mistral says Mistral Large 4 is live via API in public preview now, with open weights due at the end of October. Reuters says the launch is part of Mistral’s push to narrow the gap with leading Chinese open-weight models.
AI
Social Capital said on October 5 that those activities now account for roughly 90% of frontier lab compute.
AI
SemiAnalysis said its “5x+ more value” claim depends on a workload-based, API-equivalent comparison of subscription usage, and argued Claude plans beat OpenAI’s on that measure. Reactions in the packet disputed whether that metric supports the headline’s broader framing.
AI
Reflection says Beam uses 23B active parameters and plans to publish its full weights later in October 2026.
AI
Matan Grinberg attributes the decision to undisclosed recurring Cognition talks. He says Factory does not know how much information Degnan shared.
AI
A post linking to Reuters claims Anthropic lost $42 billion in 2025 and says its IPO prospectus warns that its AI could pose “existential risks to humanity.”
AI
A tongue-in-cheek post puts a fully populated GB300 NVL72 rack at roughly $5 million, or $3,165 per kilogram, and predicts racks could outprice wholesale cocaine by weight in 2030.
AI
A 2025 New York Times report cites intelligence officials saying employees used chats meant for sensitive security matters for discussions with sexual themes.
Technology
Artificial Analysis ranks Opus 5.5 first in its Coding Agent Index, but reports a $13.04 cost per task—21% above Opus 5—despite lower token prices.
AI
OpenAI says Sol and Luna’s API prices are 50% below GPT-5.6. A launch-day recap cites Anthropic’s claim that Opus 5.5 costs around 40% less than Opus 5 on typical workloads at default settings.
AI
The owner says their RTX 5090s also rose in value from $3,500 to $7,000, revisiting a June recommendation to buy rather than wait for lower computer prices.
AI
SemiAnalysis describes rapid expansion of the TPU stack to outside customers, a growing customer base and a weakening competitive advantage for CUDA.
AI
SemiAnalysis founder shares a brief remark on AI chatbot choice via social post.
AI
OpenAI engineer Kevin Liu announces release of internal data on coding agents.
AI
Cerebras DevX head asks investors if AI-generated pitch decks warrant rejection
AI
Former OpenAI engineer Will Depue shared the view on X.
AI
Assistant professor shares interest in a post from EnactraAI on X.
AI
AI researcher posts observation on human brain efficiency in hardware terms.
AI
The remark appeared in replies to a tweet by developer Ben Awad on X.
