• 0 Posts
  • 317 Comments
Joined 1 year ago
cake
Cake day: April 13th, 2024

help-circle







  • I think it’s all performative bullshit, not good policy.

    Some decision maker has to appear innovative to his superiours, so he decides to have some number of locations assigned to a trial group and some bullshit installed. Even if it fails, just as long as he finds the right moment to start appearing critical of the experiment he can still pull off his play. After all moving fast and failing fast are also virtue in modern corporate bullshit lingo.





  • If you look at the desktop, there is AMD and there is Apple silicon

    You can get workstations with Ampere Altra CPUs that use an ARM ISA. It’s not significant in the market, more of a server CPU put in a desktop for developers, but it provides a starting point, from which you could cut down the core count and try to boost the clocks.

    There is also the Qualcomm Snapdragon X Plus with some laptops on the market from mainstream brands already (Asus Zenbook A14, Lenovo ThinkPad T14s Gen 6, Dell Inspiron 5441). That conversely could probably scale up to a desktop design fairly quickly.

    You’re right that we’re not there, but I don’t think we’re that far off either. If Intel keeled over there would be a race to fill the gap and they wouldn’t leave the market to AMD alone.




  • I don’t think the centralised approach works either. If you bake that grouping metadata of individual popular pages into Firefox you have an issue with keeping it current if page content changes. And you have a difficult trade-off between covering enough pages vs not blowing up the size too much. And the approach can’t work for deep web pages, e.g. anything people can only see when logged in.

    Ignoring all that: The groupings you could pre-process would be static and determined over some assumed average user behaviour, not an actual cluster of a specific users themes. You take some hardcore Warhammer 40k fan, and all his tabs on minis and painting techniques and rulebooks and fan media, and apply the static grouping then it all goes into “Warhammer”. However if you ran it locally it might come up with “Painting” “Figures” “Rules” “Fanart” or whatever. It would produce a more fine grained clustering for someone who is deep into a specific niche interest, and a more coarse grained one otherwise.

    So I think fundamentally it’s correct to cluster locally and dynamically for a usable result. They need to make it opt-in, and efficient enough. Or better yet they could just abandon the idea because it’s ultimately not that much use compared to the required inference cost.