There are only 4 months left in 2024.
Accumulated another month's podcast debt.
Reading
Technical
Hacking Millions of Modems (and Investigating Who Hacked My Modem)
What We’ve Learned From A Year of Building with LLMs
It looks like ‘text-embedding-3’ embeddings are truncated/scaled versions from higher dim version
TLDR: truncate first N dim and scale linalg.norm
v0 = emb(dim=1024) v1 = emb(dim=3072) v1 = np.array(Embedding1) v1 = v1[0:1024] # truncate the large model at higher dims to match the low dim model v1 = v1/np.linalg.norm(v1) # re-scale back out to unit hypersphere. c = np.dot(v0,v1)
S-99: Ninety-Nine Scala Problems 4clojure/clojure-koans, etc, more pragmatic
Extracting Concepts from GPT-4 and previously, CLIP, golden bridge Claude
Non technical
Rice Farming Gets an AI Upgrade
Start Presentations on the Second Slide
Human Computers: The Women Who Powered Early Technology
What if we teleported the oceans to Mars?
How Actors Remember Their Lines
Actors delve into the character's perspective, motivations, and background, analyzing lines in relation to the character's emotional state. This deep understanding, rather than mere memorization, enables actors to recall lines naturally during performance.
How the Guinness Brewery Invented the Most Important Statistical Method in Science
Youtube
~10 minutes
6 minutes Make a beautiful back, shoulders and neck! Easy exercise that you can sit down
Longer
时隔两年，38号再评电动自行车 极核AE4i Pro - 38号不只会评车
Podcast
None!