By far the biggest news item of this past week has been the death of David Lynch, my recently-designated 'favorite director of all time' and source of much personal metaphysical reckoning over the last year especially. I'm so happy my podcast-and-movies hobby moved me painstakingly through his full body of work over the past many months. I am a richer person for it.
Like most who track the AI industry, my (very short-feeling) week has mostly gone to following DeepSeek's momentous release of their new reasoning models.
Most of what I've read this week comes from that news, with some holdovers from the thoughts I've had rolling in my head over the past few weeks about finding a healthy balance for my online life going into '25 and the upcoming, oh let's say four, years.
For DeepSeek and their breakthroughs' consequences for AI and the broader world:
- A solid run-down of what's so special about what they did,
- Why this and other developments are cause for pessimism when looking at stocks like Nvidia's (what a prescient article this one was!),
- An accessible & short academic book on "Foundations of Large Language Models" (which I admittedly just skimmed through).
On finding online balance:
- This short video on kicking phone addiction through understanding the design behind skinner boxes, which has me trying out ideas to automate the suggestions she gives,
- A personal post that resonates with my own efforts to limit doom scrolling by deleting apps and putting in purposeful friction (great site URL - one of my all-time fav films),
- A nostalgic piece on the death of personal computing, the rise of extractive business models, and how to reclaim the parts of computing that have more-or-less been beneficial over the past decades,
- Watching Protesilaos' demo of organizing tasks & agenda with Org mode, which I am adopting for work,
- And finding it funny that it has taken humanity about 50 years to get to yet another piece of what Douglas Engelbart and his team did in the late 60s at SRI.