The idea that RAGs "extend their memory" is also complete bullshit. We literally just finally build working search engine, but instead of using a nice interface for it we only let chatbots use them.
To sum it up even more : this looks like standard end-to-end encryption, but any app user have the same network traffic, completed with fake data if no communication is needed.
Also, working a bit on developing my photos from RAW over last years taught me how we actually expect a lot of magic from a regular camera. The brain does a lot of work and low/high light compensation, color balance, etc... are required to some extend. Of course sometimes it becomes a bit absurd : most smartphone pictures seems oversaturated, with clear blue skies and I one took a photo of a blue-ish mountain because (I think) some classifier thought it was part of the sky.
I don't mind having a few screws to remove every few years when I need to replace my battery.
Although there is another thing, I'm not sure but I wonder if it has any impact. My FP3 has made a few very bad falls and nothing ever broke. I wonder if its "bad" integrity makes it very good at dissipating the fall's energy.
TBH I'm not sure wider adoption would worsen things ? Gaming distros would probably ship bullshit anticheat modules by default while the others would not, or at most provide some documentation on how to opt in.
I think it's quite similar to the situation with NVIDIA proprietary drivers? (I don't own a graphics card so I'm not super aware on this topic)
Back in ~2010, my first dual boot was an Ubuntu. It was fairly easy to run WoW from Linux and it gave me a solid >15fps while Windows ran at less than 10fps.
I was very young at the time but still aware that this was super impressive with extra compatibility layers. That definitely took part in selling Linux to me.
Yeah, I think there was some efforts, until we found out that adding billions of parameters to a model would allow both to write the useless part in emails that nobody reads and to strip out the useless part in emails that nobody reads.
In the curl git repository most files and most content are plain old ASCII so we can “easily” whitelist a small set of UTF-8 sequences and some specific files, the rest of the files are simply not allowed to use UTF-8 at all as they will then fail the CI job and turn up red.
Is your pin automatically updated to reflect inflation?