the lisp/ directory. No package-install, no straight.el, no
A few weeks ago my friend Harrison (@hktouw) and I did our yearly Tesla FSD cruise around the Bay Area — seven hours of letting the car drive while we talk about whatever comes to mind. This was the first year we never had to take over the wheel, which meant even more time for conversation. We covered AI adoption, investing, and then landed on something that’s been bugging me for a while. Why do we still store credentials in plaintext .env files?
。新收录的资料对此有专业解读
The optimal configuration was $(45, 52)$: layers 0 through 51 run first, then layers 45 through 79 run again. Layers 45 to 51 execute twice. Seven extra layers, near the middle of the 80-layer stack, bringing the total parameter count from 72B to 78B. Every extra layer is an exact copy of an existing one. No new weights or training, just the model repeating itself.
In one month, she watches what her friends would take a year to watch. "But at least I'm not online shopping!"
Shortly after 5 p.m., the Flyer lifted off from the parade ground. The aircraft was a skeletal structure of spruce and muslin, its twin propellers driven by a small gasoline engine. Unlike modern airplanes, it had no enclosed cockpit, no seatbelts, and no protective structure. The pilot and passenger lay side by side on the lower wing, exposed to the wind.