To me the hoops that smoothbrains will jump through to avoid IPv6 and stay on legacy IPv4, especially when hosting, is pathetic. NAT, port forwarding, tunnels, blah blah blah blah.

I have something like ~1.2 trillion times the number of routable addresses that the entire IPv4 space has. Not all are reachable, of course, just the services that need incoming access and they're each on their own isolated DMZ.

Forgot to add: if you enable IPv6, be sure to use the browser extension IPvFoo. It shows a small 4 or 6 in your URI bar to let you know what protocol you're connecting to. Once enabled, you'll see just how much stuff out there actually is IPv6.

Started the move about 18 months ago when I decided to get off my lazy ass. My ISP gives out a /56 prefix, so that lets me run 256 /64 subnets/VLANs in the house, currently there are ~10 in use. Everything get a GUA through SLAAC and I use RAs (Router Advertisements) to give ULAs to everything. Any external facing services get their own VLAN and /64 for the system(s) as needed. Firewall blocks all incoming as they usually do by default and I punch a hole for the external-facing systems. They can't reach back into the network, they only answer the phone. All the systems update DNS dynamically if the prefix or full address ever change.

I have an SSH bastion set up. In all this time there has not been a single SSH attempt from the internet. On IPv4 it was constant background noice.
For those legacy IPv4-only systems on the internet, I set up NAT64. I have an IoT VLAN and IoT 2.4 GHz wireless network that are only IPv4 because a lot of IoT network stacks are junk.

I'm still farting around with it, but man oh man, there's no way I'd go back to IPv4. It was one of the best moves I've done in ages.

I think the idea is to keep shitting out 4.X releases until openai releases GPT-6, at which point anthropic will release whatever version of "mythos" they have mostly working that day as Opus 5.0 on the same day, or the next. They'll probably claim it's so operful they're skipping 5 and just calling it Opus 6 for marketing reasons

Everyone has to stop what they're doing for an entire day, travel to the training center, which costs money, they have to rent the training center, which costs money, they have to pay the training person to present the training materials, which costs money, and they have to develop the training materials/course, which costs money.
 
And then the next day is going to be complete chaos, because the training materials were developed against v0.7 of the software, and everyone is using v1.3 of the software, and nothing will get done for, minimum, 2 days, and you won't actually be at the same level of effectiveness for 3-6 months, and in some edge cases, 18-24 months, possibly longer.
 
I was at one company, and the 80 year old lady, Wanda, who ran payroll, worked on a specialty windows 95 computer, because she could not be retrained and didn't want to learn new software. And nobody messes with the payroll lady. This was in ~2014. I just looked her up looks like she passed away finally, probably still using Windows 95 all the way until 2023 bless her.
 
Anyways TL;DR for digital soveignry $2000/user is a trivial amount especially as a one time cost

Yeah I'm kind of done feeding Google every scrap of information about me for advertising purposes. I switched to private search and email last year for about $100/year and don't miss them at all.

output is about 6 tokens/s with 16k context window i'm not having any issues since it went live this afternoon. it's not sparkling like opus 4.5/6 but gets the job done

i generally send it a voice note via telegram while driving and then check back in like 1-2 min, or it is sending me a reminder about something on our shared calendar. it's still faster than texting my buddy about making plans for this weekend or whatever.

I'm using a $200 used ~5 year old (from the ebay listing) HP EliteDesk 805 G6 DM Desktop Ryzen 5 PRO 4650GE 3.3GHz 32GB RAM 512GB SSD WiFi in cpu mode... you don't need a gpu to run single user local LLM... just a bunch of ram. This isn't 2022 anymore

It's about 5 tokens/second which is totally fine for an async assistant. 20 tokens/second is about the lower limit for usable in realtime. You can also set it up to use a smaller model for quick questions (what are the next 6 items on my calendar/to-do list?) and drop through to the bigger slower model for harder questions (can you add this feature to my internal ticketing system and redeploy?)

I ordered 64gb of ram about an hour ago and i'm planning on running either qwen 35B-A3B 8 bit or 122B-A10B 3 bit in fully offline mode.
 
  >the actual cost of 'running the AI.'
 
is a fixed $200 cost (ram upgrade) + electricity

I cancelled my subscription overnight, and I'm using the free credits they gave me to wrap up some things and transition away. I am not going to be locked into someone's walled garden again.

I don't like ads either, but I do like that they (at least for now) have a paid tier with no ads. If there was an option to use google services at some paid tier, without being part of their ad network, I'd probably pay it. But there isn't and llm is as good as search these days (in many cases anyways) so I'm happy to jump ship. Piss off, google.

Trump in his first term was willing to go all-in on human spaceflight to mars...until he realized he couldn't get it done before the end of his term. Trump has always been interested in space stuff...but only if it's achievable within his term. This seems like a play to keep contractors employed and skills sharp until the next administration is seated, which will hopefully be willing to invest in goals longer than 4 years.