The actual blog post on this is more interesting.
One silver lining about the component restraints, hopefully it drives people to start creating efficient software again. Actually optimizing what you’re doing to save on RAM rather than a gig per tab in Chrome.
“Claude, make this software more efficient.”
Wish that was the case for a lot of new video games instead of relying on DLSS or FSR to cover up their poor optimisation
Picture the level of quality and performance we could have if we had both traditional optimization AND upscaling optimization AND frame gen optimization.
Cause holy fuck the number of times I’ve seen fsr/dlss slaped onto a game and no effort put into making it perform well.
Cause there IS work you can do to make fsr/dlss actually look good and perform better.
I’m in a community making open source firmware for the Centauri Carbon 3d printer. The thing has 128MB of of RAM. The smarter people in the community worked some magic getting Klipper working on these things, and we’re still fighting to shave off a couple more MB in average use to get things working smoother.
Talk to the folks running Linux on sn esp32! That’s some shrinking
Seems to me like everything is instead becoming even more bloated with AI stuff. A Google chat tab uses 2 GIGABYTES on my work computer.
Yes!
stupid RAM price benefit: reduce bloat
AI IS THE GREATEST THING TO HAPPEN TO MANKIND (ignore literally everything else)
I never even thought about this before, but what kind of hardware is a root level DNS server like 1.1.1.1 or quad 8 run on? Are they all roughly the same, or do they take different approaches to the load?
Look up how Cloudflare does anycast BGP. They have servers in data centers everywhere and leverage the fact that they can all be 1.1.1.1 so people’s queries only hit the physically closest server, it’s pretty cool
40 bytes are saved, exactly this amount.
