24GB RAM for 2599 USD.

My best results in intelligence and speed in my MacBook Pro 2019 with an i7-9750H and 32GB RAM are with Qwen3.6 MoE version. I use it for my Hermes Agent:

llama-server \
  --hf-repo majentik/Qwen3.6-35B-A3B-RotorQuant-GGUF-Q2_K \
  --hf-file Qwen3.6-35B-A3B-Q2_K.gguf \
    --host 0.0.0.0 \
    --port 8080 \
    --threads 4 \
    -tb 4 \
    -b 2048 \
    -ub 2048 \
    --ctx-size 65536 \
    --n-gpu-layers 0 \
    --flash-attn on \
    --cache-type-k q8_0 \
    --cache-type-v q8_0 \
    --reasoning off \
    --batch-size 2048

It takes around 20 minutes for Hermes to start using this as a local model. But after that it is quite good.

Imagine if the EU start demanding electric cars batteries to be of a standard size, Current and Voltage. Something like 0.5m wide, 0.5m long and 0.1m tall, less that 40 Kg with a standard C13 plug, 480V and 10A. I don't care about the chemistry you use inside. As new chemistries get released, you can replace your old battery for newer with more range, without throwing the car away.

Drones can crash on your roof or fly through your window, and then burn your house down with a lithium battery fire.

Other drone problems:

1) Dogs will attack drones. From a Slashdot story: "... they hover 23 feet in the air and lower their cargo to the ground on a tether." Dogs will grab the tether.

2) Some people will shoot at drones.

3) Over houses, they will be noisy, scarily noisy, for the local neighborhood.

4) If there is a technology failure, someone in the neighborhood may die, or a house may be burned down.

Remote control over drones can ALWAYS be eliminated or hijacked by radio frequency interference.

I don't want drones near where I live. Will drones be allowed near where Jeff Bezos of Amazon lives? No, of course not.

Iron-Air batteries gets over 10.000 cycles easily. Sodium as of today are around 2.000 cycles, with a promise to get to 10.000.
Long term cost Iron-Air makes more sense.

hahahha

Axiom number 1 of selling: "Selling is always emotional, and once the buyer is convinced of the purchase, they take it upon themselves to find the technical reason to justify their purchase."

For instance, you can do many things with KVM (RedHat, Oracle and some others sells KVM support), just like with Nutanix. A very common pattern I've used in several companies is for the KVM nodes to mount storage via iSCSI, and then each node performs a DRBD replica of the iSCSI volumes.
For example, at one company, they have two sites 10 kilometers apart. Each site has a 10TB iSCSI volume. Each site has 4 KVM nodes.
All 8 KVM nodes mount both 10TB iSCSI volumes and use DRBD + Pacemaker.
This way, regardless of which site fails, all the data remains synchronously replicated (thanks to DRBD), and I still have 4 nodes left to run the most critical services.

I am sure that you can find out some obscure Nutanix feature that is not pressent in KVM to justify what you want to buy.

"We're currently looking to jump ship to either Microsoft's Hyper-V or Nutanix". You alredy got screwed by Broadcom and your plan is to move to another propietary options... AMAZING.

Or even use an LLM Deep Research to look for alterntatives.

Remember This is the company that forbid GameStop trading during the stock meme days. Move along. Nothing to see here.

They should make it easy for comma.ai to plug into their cars and pay comma.ai to support their cars. Sending liabiities to them.

Yep. I had a big chunk of code with a typo that I was not able to detect. I referenced the wrong variable.
I describe the problem to Gemini and gave it the code and it fixed it.

"hill gradient" is the issue. On a perfect flat leveled surface you can add as much battery as you want and the car will still move. The acceleration will suffer though. Resistance increase on a flat leveled surface for a car when adding more weight is really low.

Something really interesting is the Netflix documentary called "take your pils". At the end, they showed a double blind experiment of people taking Aderal vs. placebo.
Although there were ZERO improvements in the Aderal group, one strange result is that almost 100% who took Aderal instead of placebo, checked one box in the test. The box asked "do you feel the pill helped you perform better"?. That didn't happened to the control placebo group.

Patents are the core problem.

Economy: Science that studies scarcity.
Patent: A way to make scarce a resource it is not scarce.

So by definition, patentes reduce the world economy. Even more importan, Mariana Mazucatto (Italian economist) managed to study all drugs create after WW2 up to know and she was able to prove that at least, 75% of the drugs were created with government funds, and then a pharma took it. modified an atom of the molecule and patented it.

Take for example Ketamine as a single shot lifetime antidepressant ( https://www.ted.com/talks/rebe... ). It was proved to work real well, but since it cannot be patented, no pharma wants to work to get it approved by the FDA as antidepressant. Finally, a pharma was able to modify the molecule and patent it (SPRAVATO). The modification was done so it is no longer a single shot lifetime antidepressant (you need to take it constantly) and to be able to patent it. It was the fastest drug ever approved by the FDA at the time.

Government and patents are at the core of the problem. If a new drug that promises to solve depression forever for every person in the planet with a single shot, and no pharma wants to test it because it can't be patented, government role is to test it and approve it.