OldManYellsAtCloud

Who has the original 2011 Daily Show clip of John Oliver's Freedom Packages? Pretty perfect for Venezuela?

@nixcraft Compiz Fusion still works today provided you haven't succumbed to the #waylandmafia

only old linux desktop users will remember this desktop effect. can you name it? #linux

@svpino Try QEMU for x86 VMs on your M3. Also try Asahi Linux if you want Linux natively. Performance is actually pretty good and you have more than enough RAM.

This is the best computer I've ever owned. But I’m thinking of buying a Linux box to remote into it and test code in an x86 architecture. What would you buy if money is not a problem?

Let he who is without Photoshop cast the first Tweet. #KateMid

@abacaj Interesting use case. Do you encounter any ECC correction errors with these cards?

This is how I save over $1000 a month on compute costs for my AI side projects and experiments. I’m able to fine tune models up to 72B parameters on this node (with room for more 3090s) Renting and having access to a node like this 24/7 would cost you over $3k a month

@PicturesFoIder Which one of them touched the other's wife?

Cop arrests fire fighter in the middle of tending to a wounded civilian because fire truck was 1 mm over the line.

Can anyone explain why @TCS is still managing the websites of @British_Airways after 3 days straight (and years before) of ridiculous failures? End TCS now. British Airways deserves better.

@devfacet Yes you're right the Pi consumes a lot less power (and also lasts a lot longer on a battery backup/UPS) but the compute and RAM of a C7000 or rack server can emulate hundreds of Pis though they consume a lot more energy. A breakdown would be interesting.

@YellsOld Could you share a link for any of those servers? I like to look into power consumption for continues AI inference like I'm doing. My current electricity cost <$30 (including network equipment).

This is how I save over $1000/mo on compute, storage, and bandwidth costs for my AI side projects: a total of 52 CPU cores, 104GB RAM, 3TB NVMe, unlimited ingress/egress, and less than 100W power consumption. The taller cluster runs API servers built with the Transformer.js library and the NATS services framework. The shorter cluster runs a Postgres database server (with streaming replication) and some services such as pgpool, prometheus, etc. They are all connected to the cloud via NATS using a secure WireGuard tunnel. No DNS server, no port forwarding, no load balancers. #raspberrypi #pi5 #AI #edge @nats_io @xenovacom @PineberryPi

@British_Airways it's time to burn down your constantly down website and app and start over. This is the third time in a week I've needed it and the third time it's down. Get rid of this garbage and start over.

I don't know which LLM you're using but as usual Google sure takes the cake for variety and naming skills. I'm not sure if I'm using Bard or Vertex or Gemini or Deep Mind AlphaCode 2 powered by Gemin Bard Vertex Bison Hoobynot.

@AKBrews No it really f*kin isn't. Just more inspiration for the subpar filmmarkers of tomorrow.

The movie Bad CGI Gator might be the greatest thing ever:

#GPT4-4 Turbo nomenclature takes me back to a time when things were simpler.

Here we go again.

USA you seriously need to get it together. The international community no longer even flinches when 9 year olds get shot in US schools. Business as usual for USA. #schootings

I really don't know what to think about investing anymore.

@Schuldensuehner Depends on how many times they halt trading.

WTF? The markets are now pricing in a probability of default of 47% for Credit Suisse. What have I missed?

I think I just watched #WakandaForever at home on an OLED TV. Is this how Hollywood tells us to go back to the cinema?

Contrast that with just this one example of insanely powerful iron you can buy from Dell. The first R6525s have 256GB RAM, 3TB NVM, 2x10G net, 2x AMD EPYC 7513. Second, same, but 2x AMD EPYC 7443. So that's a total of 288 vCPU, 15 TB NVM, 1.3TB RAM for $1,287/month over 3 years!