They do get released. I need a source of high quality rips for the NAS to stream from.

Still no blu ray last time I checked.

Globally, you probably are. It doesn't take that much to qualify. There is a website to check via income.

Too bad we're already over 1.5 C though it will take a decade for it to be official.

Ok, if you don't use their web site you won't see the UX dark patterns. Trust us, they there and fit with the overall garbagefication theme. Annoys the living shit out of me. At least no more Prime Video UI and ad trainwreck.

If you haven't noticed, you've been not paying attention. I canceled Prime a while ago and they try very hard to get you back. And they try to sneak on you billed expedited shipping when over minimum gratis shipping quota. Dark patterns galore.

It would be a major pain for me to boycott them completely so I don't, yet.

How would a national government (not TLAs) target particular individuals in a large number of users and what information can they gather given e.g. https://mullvad.net/en/help/no-logging-data-policy ? So perhaps not quite as easily as ordering a tap.

My national government has no business knowing which protocols I use to contact which endpoints and tamper with that traffic. Wrapping up that information in a tunnel is a good first protection layer.

You forget that nation-states control your ISP. And of course you can choose your VPN provider or run your own.

The provider and national TLAs will see all traffic that is in cleartext and meta traffic which is even more valuable. It can also actively tamper with that traffic. So you're technically incorrect and you assume your threat model is universal. It's not. And, of course, there are use cases for Tor, whether with or without VPN.

Nobody is using MMF these days even for local runs. As to sfp, check https://fs.com and pick a matching pair that is cheapest. These days it makes sense to use 10G or 25G rather than 1G. Some people run 100G for their homelab, but even used it's pricy and noisy.

A free running cellular automaton (CA) approach in hardware would work, but each cell would be a much souped up SRAM cell, the interactions would be all local and 2D. Considering Cerebras is 40 G SRAM on the 300 mm WSI and is about at the cooling limit I'm afraid you do not have 5 orders of magnitude. Perhaps reversible spintronics can help with the power draw, but you still have to splat a higher dimensional network so not just local interactions into a 2D array.

You seem to not be using open source software packaged for multiple architectures or which can be built for your binary target. Most people will be just using a browser and an office suite.

No, that captures just the neuroanatomy. Not the properties like density of ion channels, type, value of the synapse and all the things we don't know yet.

You seem to trust Nvidia. I don't.

They write

"Of course, AMD is trying to get into the the AI training and inferencing game itself with the Instinct MI300 chip. And that, perhaps, is the main if modest cause for hope. If AMD can gain some traction in that huge market, it will not only be making lots of money, it will be in a position to do a similar thing to Nvidia and push some of that technology across into its gaming GPUs."

which strikes me as incorrect. AMD MI is pretty widespread in HPC. With margins lower in the consumer market it makes sense to focus on HPC.

Me, too.

I would look into thin clients and Lenovo etc. tiny PC for office on eBay. I run old low power low noise rackmount Supermicros which are nice but hard to find at low prices.

Yes orders of magnitude, but not too many of them. The real estate of a 300 mm wafer is limited, the structure shrink is saturating and you can't get too many layers. You still need a packet switched network on the wafer even if the rest is mostly analog. Perhaps spintronics can limit the power requirements too.

And these power lines into the EU are very easy to disrupt and disrupt again after repairs.

Factor in power bills and heat and noise into your calculations.

Secure has no meaning.

Sadly, you can't download the hardware to run models that size.

There are diminishing returns in semiconductor photolitho. Moore scaling is long over, absolute real estate see WSI with Cerebras, DC costs and power envelope are all sending a clear message. Quantization is there, so you can go from digital multipliers to analog and go spiking networks, but transformers and Co have little power there.

Also, the kind of economy that can carry Gen AI as business model is not a given, long term.

https://h01-release.storage.googleapis.com/data.html

Make sure you have spare 1400 terabytes.

There is a reason 1984 was written by a Brit.

Neat. At 750 Wp I wonder what the EROEI of the installation is.

Not enough sanctions. More are needed.

Interesting, wasn't aware. It remains illegal in Germany then, as long as you seed. Notice in Germany you will receive a nastygram in any case as long as you've been identified as a member in a swarm, even if you do not seed. It would be interesting if someone would contest that, though you will likely lose since technical details don't matter in most courts.

No port mapping, no upload, so technically legal where only sharing is penalized. IANAL.

Since most VPNs no longer have port mapping it's a pure download which would make it technically legal e.g. in Germany.

Potentially double use for space launches.

Without VPN or Usenet subscription it's going to be difficult and/or illegal.

Anemic as it is it's even fake growth. See https://surplusenergyeconomics.wordpress.com/ for some more real world-relevant metrics.

Too bad Iris production is unaffected.

LineageOS with Jerboa over Mullvad VPN here. Not many options on mobile devices.

"First, will the world be able to sustain such a large consumer class within its planetary boundaries?" indeed.

360 km or 194 nautic miles is the inter-port distance.

Nah. He's that Youtube dude.

Funny they don't mention the reasons for the deficit in Suomi: anti-russian sanctions.