The buyouts were offered to their aging orgs. Datacenter workers, quantum, MAI, sales and marketing, and cybersecurity were not included. https://finance.yahoo.com/technology/articles/microsofts-buyout-offer-employees-shows-135510739.html

Tide pods.

The cloud servers are actually literally Xbox series x’s.

Ah, I may have to clarify that technically Microsoft isn’t the AI bro I’m referring to, though they do partake in the insanity. I’m referring to OpenAI specifically. “circular financing” and all that, but Microsoft is more of the hosting provider for OpenAI.

Xbox cloud servers are literally 5 motherboards: 1 management board and 4 Xbox ones. Developers can optimize for a single configuration rather than bake in support for multiple ones. Once it works on an Xbox dev kit, it works on all Xboxs including cloud, console, and handheld. SDKs are making that a thing in the past, but it’s not perfect.

I feel that it is more in-depth and more accurate than most media on it. They acknowledge that they had to personify it to appeal to humanity. Despite being in the thick of it, I do think it’s a good video to share to those lesser in the know. They explain it in easy to digest language without glossing over the technical workings of it. They are anti-Ai purely out of self-preservation and admit it in another video (which is manipulative).

They are sitting in the warehouse because there aren’t enough other parts (e.g. DIMMs), power, or data centers for them. And as stupid as AI bullshit is, they are paying more than gamers by orders of magnitude.

Xbox and PC versions aren’t exactly 1:1. Plus, those are extremely expensive GPUs to not be selling for top dollar to the AI bros instead.

Edit- I’m leaving out things like architecture, shaders, tensor cores, NVIDIA vs AMD, SXM vs PCI-E, etc.

They aren’t the same GPUs that are in Xboxs. those are nvidia GPUs eg the nv72 or gb300. There’s no way to make them compatible with Xbox. Source: my unc works at Microsoft data centers.

A quick glance shows that I’m in the minority. I grabbed a Bosgames M5 with 128GB UMA RAM when I discovered openclaw (at the time claudebot). I’m running Lemonade, which is a wrapper for llama.cpp, sd-cpp, flux, and others. I’m happy to have paid a “mere” $2400 compared to what similar systems cost now, and now I’m toying with fine tuning my own models. Though, I primarily run Hermes Agent paired with Qwen3.6 35b A3B (using a heretic ablitieration for 0 refusals).

Hermes Agent does a lot of random stuff for me right now. It manages my calendar and to-do list in nextcloud via MCP server. It created and maintains an llm-wiki based on my Obsidian Vault. Then it creates a report every morning about both and the weather.

The server has tailscale installed, so yes I can access it anywhere.

I got Lasik last year. I could smell it, but idk about that being the worst. I did not care for the calipers that held the eyelids out if the way, or the machine physically suctioning onto the eye itself. I got a lil Valium (thank goodness) and a teddy bear. The bear was forced into my hands, and I remember thinking “oh, the fluff is missing from the middle. I wonder why that is.” when the suction was applied, I squeezed it so I didn’t move or anything. Hands fell right into the cavity of missing fluff. “oh, that’s why” I thought.