You shouldn’t be comparing with DIMMs, those are a dead end at this point. CAMMs are replacing DIMMs and what future systems will use.

Intel likely designed Lunar lake before the LPCAMM2 standard was finalized and why it went on package. Now that LPCAMMs are a thing, it makes more sense to use those as they provide the same speed benefits while still allowing user replaceable RAM.

▇▅▆▇▆▅▅█

All ideas are made up

I agree, I bought my car in 2018 and its got a small screen and carplay / android auto. No OTA updated, no capacitive buttons, i don’t have to dig through touchscreen menus to change settings. I want to go electric soon, but everything i have driven is obnoxious with what you have to deal with in the cabin.

AWS has multiple teirs of storage options in s3, some replicate and some dont. by default those that do replicate do so in multiple availability zones, but not across regions. unless you turn on cross-region replication (CRR) which is an additional charge.

So, for example without CRR if your bucket is in us-east-1 and 1 availability zone goes down you can still access the data, but if all of us-east-1 is down, you cannot.

We still need base load of which nuclear is the best option.

it should probably stay in docker containers

As if managers even know what RISC-V is

What about making these massive trucks require a CDL to drive? Sure some would be dedicated enough to do it, but im sure most people would see that requirement and just get something more reasonable.

Its the server world that is demanding it. For most consumers 4.0 is more than enough, but servers are already maxing out 5.0 and will probably immediately max out 6.0 when devices actually become available.

the NSA (which lacks a mandate to act on US soil, and CF is a US company)

They absolutely do have a mandate to operate on US soil, that is actually the main mandate and there is a separate military agency (CNMF) that operates on foreign soil. They are both headed by the same guy though so they might as well just be one agency.

the thing about Recaptcha is that it didn’t always gate keep a google provided service, so that logic doesn’t really work. i agree though that we all benefit from less bots.

yahoogle

There is one extra step. I have an 6700xt, and with the docker containers, you just have to pass the environment variable HSA_OVERRIDE_GFX_VERSION=10.3.0 to allow that card to work. For cards other than 6000 series, you would need to look up the version to pass for your generation.

Here’s an example compose file that I use for ollama that runs ai models on my 6700xt.

version: '3'
services:
  ollama:
    image: ollama/ollama:rocm
    container_name: ollama
    devices:
      - /dev/kfd:/dev/kfd
      - /dev/dri:/dev/dri
    group_add:
      - video
    ports:
      - "11434:11434"
    environment:
      - HSA_OVERRIDE_GFX_VERSION=10.3.0
    volumes:
      - ollama_data:/root/.ollama

volumes:
  ollama_data:

have you tried the rocm docker containers that amd makes for your needs? it pretty much makes installing rocm on the base OS unneeded for me. https://hub.docker.com/u/rocm https://github.com/ROCm/ROCm-docker

its a play on words of the 2 main types of guitars (electric and acoustic) Basically a joke implying that if one type of bike is electric, the other must inherently be acoustic.

it doesn’t, what this is suggesting is the vpn was routing traffic through it so they could analyze snapchat traffic. not the contents of it but essentially meta analysis of the traffic. how often it was sending data, how much data, where it was going etc.

All OS updates are signed If what you are saying were possible, then all phones would always be susceptible to a man in the middle attack of someone faking apple’s server to send a fake OS update. All this seems to be is a qi charger that has the ability to also turn the phone on / off then it can just connect to apple’s server and update itself like normal.

just a small correction, /etc does get snapshotted when upgrades happen and will roll back along with everything else. you are correct though that home does not get snapshotted and is fully mutable.

I don’t have an answer to your nvidia question, but before you go and spend $2000 on an nvidia card, you should give the rocm docker containers a shot worh your existing card. https://hub.docker.com/u/rocm https://github.com/ROCm/ROCm-docker

it’s made my use of rocm 1000x easier than actually installing it on my system and was sufficient for my uses of running inference on my 6700xt.