> (Turbo) Pascal eventually gave way to Delphi which, despite its technical merits, never had the same impact.

Not sure I can agree with this. My time with TP7 was fledgling, Delphi was my first "professional" language. Over the years since, I still get cold-called and -emailed from recruiters who would do anything but pay for a Delphi developer.

An interesting parallel with NT: Dave Cutler was hired by MS to replicate his success with VMS and built NT.

Likewise, Anders Hejlsberg was hired by MS to replicate his success with Delphi and built C#.

I think that's pretty impactful.

As a 14B MOE I was rather expecting it to fit into smaller environments than 2x128. Am I wrong on that?

Mostly yes, you are wrong.

Active parameter count really drives inference speed not ram requirements.

For this model, assuming you want the full 1mm context length at 4-bit quants, you need something like 175 GB total VRAM.

~150GB of that will be weights (varies a bit by quant format) then ~3-5GB for kv cache (might be lower here, I think they're using mostly HCA compressed cache) and then I'll allocate a couple gb for overhead/concurrency (say another ~10GB as a safety buffer).

---

So the good news is that it comfortably fits into 2x128 at 4-bit, and we can realistically stretch to a 6-bit quant and make it just fit (~244GB needed for the full 1mm context).

What are you comparing this to, and for how much, and what sort of benchmark numbers is it getting? I'd love to run better cheaper hardware. I'll gladly pay you a reasonable finders fee.

its commonly used for smoking meats at 250F / 120 C.

Yeah, when I got to the "can't cook with it" part, I was like wait, what? I use it all the time for smoking brisket.

Ah well, TIL....

now - that said, I'm talking about pink butcher paper, I suspect the white stuff is unsuited for heat at all.

Uname is not enough here to identify what has or has not been done to your system to deal with the compromised certificates.

No, no, I meant it to show that the machine is running the current/latest macOS, and the results I am getting from curl on the two URLs.

Maybe you have another curl in your path?

    $ uname -a
    Darwin MacBook 27.2.0 Darwin Kernel Version 27.2.0: Wed Oct 14 23:47:21 PDT 2026; root:xnu-13432.41.4~25/RELEASE_ARM64_T6030 arm64 arm Darwin
    $ which curl
    /usr/bin/curl
    $ curl https://dv-g3b-emudhra.we-have-a-problem.com/
    Your client successfully connected to the site. That's bad!

Aha! Touche! Indeed, I am using brew's curl ("/opt/homebrew/opt/curl/bin/curl"). That's why; the one at "/usr/bin" connects just fine (facepalm).

likely "focus, and connect ( with other people )"

He even called out the unintended contact!

massive crushing pressure is not the only way to create diamond particles.

do you mind explaining more

Chemical Vapor Deposition (CVD)

Opencode works that way-ish.

Also - formalize what can be. There may be e.g. opportunities to write scripts that can be tested and linted. I realize this isn't always possible, but I do think it should be ruled out rather than ignored.