Loading activity…
@zephyr_z9
AI & Chips | Not Investment Advice | DYOD
Technology
Modine shares jumped 29% in extended trading after reported Samsung takeover talks, but no final decision had been made.
AI
willdepue says initial-weight patterns match Blackwell hardware, citing reproductions on a B300 and rented GB200.
AI
Kimmonismus cites Bloomberg reporting a possible Nasdaq debut the following week, with timing still tentative.
AI
Two X users call the market reaction overblown, with one arguing the coverage likely concerns an older deal.
AI
Anthropic's October 2026 acceleration was about 80% due to AI, METR guesses, though the estimate is highly uncertain.
AI
A Binance shareholder said Tether froze about $971 million and Circle about $823 million; he expects user funds fully covered.
AI
One user cites an assumption that 550,000 humanoids produced annually would need 21.5 PB of DRAM.
AI
A commenter says an Alibaba employee believes its chips will soon support 50–100T models.
AI
A user quotes a “neolab” friend saying it could resell compute even if it fails to train the best model.
AI
Artificial Analysis says Kimi K3.1 Pro scored 64% on its Cyber Index, up 23 points from Kimi K3.
AI
OpenAI reported hardware side channels; AndrewCurran_ quoted its account of agents passing synthetic secrets between virtual machines.
AI
Kimi claims a 1-million-token context window for Pro and two to three times Pro's speed for Highspeed.
AI
Moonshot says both versions are live on Kimi.ai, Kimi Work, Kimi Code and its API.
AI
One commentator says as few as two major Chinese AI startups may still be growing by late 2027.
AI
A user says an investigation is ongoing and claims Quest serves roughly half of U.S. hospitals and physicians.
AI
POLITICO reports the Technology Support Office was established in April by reorganizing two classified Pentagon intelligence programs.
AI
A post citing Bloomberg says DeepSeek raised its target at least twice, from about $7.5 billion to $12 billion.
AI
Its developer says Flint exceeds 125 tokens per second on an M4 Pro MacBook Pro at short contexts.
AI
A user estimates 42 GB200 NVL72 racks could suffice at what they call production throughput, using off-peak prices only.
AI
A user calculates the range assuming 97% cached input tokens and GPU rental costs of $3–$6 per hour.