I'm wondering: how can understanding gradient descent help in building AI systems on top of LLMs? To mee it feels like the skills of building "AI" are almost orthogonal to skills of building on top of "AI"
> There's no technical reason, other than thinness, for the way Dell soldered the RAM down, for instance.
That's a sentiment that's been repeated a lot, and it's not fully true.
One important property that soldered-on RAM has is increased security. There's been a demonstrably practical way to break full-disk encryption with physical access to a turned on or sleeping computer by re-attaching the memory quickly to another computer. The keys then can be read from the memory.
That's not a vector you have anymore if the memory is not removable.
Question from someone relatively clueless: Does that mean that NFTs also will use negligible amount of electricity once that's completely gone through?
Literally the first link in "Discover our content selection" leads to an expired domain name [1].
I'm not sure what to make of it but it does seem like a somewhat of a telltale for what's going to happen with my videos should I upload them to one of the PeerTube hosts.
One important data point there is Visual Studio Code.
It used to be really popular in 2017 (double the popularity of Atom/Sublime/Webstorm), now it's just eating everyone's lunch (triple the popularity of Sublime/Vim/Webstorm/Atom, almost equal to all of them combined).
Compare that app + Surface Hub to Google Jamboard [1] which is a single-purpose 4k TV capable of only fulfilling this use case and costing $5k + $600/year.
I think Apple might like that with their privacy stance. It really does increase the user's privacy by never sharing the location with external parties.
It will also simplify things for the developers in simple use cases and offer more integration like "jump to native maps" etc