Technology
Mixture of experts
An architecture in which only some of the model's specialised blocks are activated for each input.
3
mentions in the archive
0
central to an episode
3
episodes with a mention
3
fragments with a timecode
Feb 2025
first mention
Feb 2025
peak · 1 in a month
104
links to other entries
Timeline
Mentions by month · May 2024 — October 2026
central to the episodesubstantial discussioncontext and comparisonbrief mention
202420252026
First
Feb 2025
Peak
Feb 2025 · 1
Months on air
3
Latest
Dec 2025
February 2025
peak month · 1 episode · click a bar for another month
central to the episode 0substantial discussion 0context and comparison 0brief mention 1
#043DeepSeek Did Not Kill NVIDIA—It Made Powerful AI Cheaper and Expanded the Chip Marketbrief mention