301.
302.
The latest Gemma 4 models use a training trick to slash their on-device memory footprint
(androidauthority.com)
303.
Latent Agents: A Post-Training Procedure for Internalized Multi-Agent Debate
(news.ycombinator.com)
305.
A Man Who Reads Books for a Living
(news.ycombinator.com)
306.
A Man Who Reads Books for a Living (One Every Two Days)
(news.ycombinator.com)
307.
308.
309.
Phone signal on trains not good enough most of the time, research says
(feeds.bbci.co.uk)
311.
Brightline proved America wants trains. Can it survive?
(feeds.feedburner.com)
313.
314.
315.
316.
I made a kernel 2.2x faster. It made my training loop 3x slower
(news.ycombinator.com)
318.
319.
320.
321.
322.
323.
324.
Atherton spent $145K to delay train electrification. The rest of us paid $400M
(news.ycombinator.com)
325.
Opinion | We Should Starve Adversaries of AI Compute
(feeds.content.dowjones.io)
326.
327.
Your workforce doesn’t need more AI. It needs play
(feeds.feedburner.com)
328.
329.
330.
Shift will clean homes for free to train future robots
(news.ycombinator.com)