Cody Collier

..   . *      .     .   * *. .        . .  .     .  *           `         .     . .      *    ..` ..          *
*         . .         ` `  *   . . .            . **              .*    * .  ` *     .  ..        .   `     *
     .           . .            ..     *  .               *    .   .           `       .         ` `  .    * *
                    . ..` .              *.  **    . * *  . . .     *   `             .     .  .     .    .
. `  .     .. .  . . ..     .       `.              . *          *   .      .   *. *. `  *        **`         *
          *         ..  .  `    `*   `    *   `.   .    `  .. .    . *   ..     `  .  `  . .   .*                *
                 ..  .   .          *     .  *        *  . *        *  .           `.  .          .        .
 *        *    .          .         *` **.       .* *.. .   *. .  `    .  . .                        `   . *
.. *    . .* .            .              *   .  .*. ..*  ..    `               .   . .*         `   .. .    ` .
       `  .      *    *  . . . . . *          .. .    .  *.         `                 *   . *.        .   *     ..

Logbook (LX)

lx.sysstack 14

Logbook (LX)

My informal notes, essays, and writing.

LX Mini

A feed of my shorter form social media posts.

Tuesday, July 28, 2026
I wonder what proportion of the training tokens come from scanned books versus the Internet at the frontier training shops.
Tuesday, July 28, 2026
Always Be Training Working out some problems with pre-training a GPT-2 model with C4 dataset on a Google TPU v6e1.

Open LX Mini