HNHacker News
TopNewBestAskShowJobs

throwaway879080

95 karma · joined October 18, 2021

submissionscomments
throwaway879080··on Twitter set to accept Musk's $43B offer – sources
quite doable, just pick up all the dead posts and comments
throwaway879080··on YC’s $500k Standard Deal
awesome, we might consider YC after all
throwaway879080··on Giant, free index to world's research papers released online
see this reddit thread on Scihub rescue mission https://www.reddit.com/r/DataHoarder/comments/nc27fv/rescue_...

there is a link there to 851 torrents , ~77TB of data (compressed zip files), and also an index (sql db dump)

throwaway879080··on The 'impossible' crane shot from Soy Cuba (1964) [video]
we don't smile because we are banned a lot
throwaway879080··on Why can’t the world’s greatest minds solve the mystery of consciousness? (2015)
wiring and signals that flow through the wires, i.e. the software

while the cells are just switches with memory, i.e. the hardware

throwaway879080··on Is Mark Zuckerberg a Man Without Principles?
more like his principles don't align well with author's point of view.
throwaway879080··on Vector-based pedestrian navigation in cities
Related:

How the brain navigates cities - https://news.ycombinator.com/item?id=28917033 - Oct 2021 (35 comments)

throwaway879080··on How the brain navigates cities
there are also "vector cells" https://www.nature.com/articles/s41593-020-00761-w

Neuronal vector coding in spatial cognition: https://www.nature.com/articles/s41583-020-0336-9?proof=t

throwaway879080··on At Netflix, a star and employees pressure a top executive over Dave Chappelle
if they cancel Dave Chappelle I cancel my Netflix subscription.
throwaway879080··on What to Learn
>working on a few things and not being particularly well rounded has worked for me in multiple disparate fields

arguably "well roundedness" is key to major breakthroughs in science & engineering

while most of us work on incremental improvements, the real innovation often happens on intersection of two or more disparate fields that until then were considered weakly related, and it does require knowing a bit of everything

the best scientists and engineers I know are "knowledge omnivores" and will digest anything. It does require well developed "filtering" capacity though, picking up signal from a lot of noise, and a sort of information hoarding mentality

throwaway879080··on Apple’s new M1 Pro and M1 Max processors
how does this new GPU and "neural engine" perform comparing to Nvidia GPUs, and do they support Tensorflow and something similar to CUDA SDK