7b gemma nano models are... not terrible. I can see having a 30b model that uses 15gb on a phone that outperforms early (1H 2024 era) GPT 4 models by the end of the decade. Maybe in a year or two. Things are moving pretty fast. GPT4o mini was good enough for almost anyone already.

And (also, but not limited to) why?

In the UK, which (I think that's like consumer reports in the US) has basic high performing models for £30, so 40 bucks?

With that said, it'll get near daily use and probably last a decade. Even at $200, that's 2c per slice toasted. Though apparently often the expensive toasters aren't that good.

"TSMC and Samsumg have fabs in the USoA"

How long before the tangerine terror fucks that up?

There's no reason to kill ARC because it's only taking sales Nvidia didn't want.

It also depended on low speed interfaces to processors. You used to be able to accelerate computers which came with a 68k chip by slapping a daughterboard with a faster CPU on it into the CPU socket. That's not practical today.

Side note, anyone want a radius accelerator for Mac SE?

"Biden's effort to reduce student loan debt has ended."

Too bad he didn't fight as hard for that as he did to prevent us from discharging such debts through bankruptcy, that fake fuck.

You can do that with any modern phone but people don't do it....

Kiiiinda I guess, but somehow everything is shit.

"Sounds like you are salty because you aren't getting the score you want."

I'm salty because a bunch of clowns who provably don't know how credit scores work are trying to tell me how credit scores work while I have proof that they don't.

Yep. It basically just does the job. Straightforward, no strange takes on interaction, plays well with non XFCE things, low resource, decently configurable, unopinionated and works decently out of the box.

"Simply paying off 1-2 credit cards in full every month while keeping your utilization low for ~5 years will get you into the high 700s, which is all you need for any practical purpose."

I did literally all of that and I'm in the low 700s where I don't qualify for first time home buyers assistance, so no and no. Your faith in the system working as advertised is just sad.

My credit score is well over 800 and I don't see how I'm exploitable. I haven't paid any CC fees or interest in decades, and have no debt anywhere else. But maybe I'm missing something obvious. Can you explain a bit? (serious question).

The literal definition of a parallel circuit is one where the circuit divides and the current is split between two components - look it up.

You know I actually reached behind me and grabbed Horowitz & Hill off the shelf just because.

Looks like someone has h4x0rized it so you can look too:

https://kolegite.com/EE_librar...

Page 2, Figure 1.1, "parallel connection". See the lack of an EMF in that diagram?

When the OP said "capacitors in parallel", that's exactly what he means. Those wires on the left and right? They're not connected to anything. They can be sure, by implication, but need not be. They're just floating there in diagrammatic or real space.

If there is not more than one path for the current the circuit is not parallel.

OKey dokey, since you keep dodging this question I'll ask again. I'm beginning to suspect you're dodging because you don't have a good answer that also is consistent with your definition.

So take two capacitors connected in parallel, yeah with one end connected to + on your battery the other end connected to -. Cut the wire to +. Are they no longer in parallel? What if you remove the wire completely? Now replace the cut wire with a switch. Turn it on. Turn it off. Do they change from parallel to undefined as you flip the switch?

This is not a physics vs. engineering definition, it is THE definition of what parallel means.

So you say, but that sounds like a definition with holes in. Somehow I've got this far in life without ever having a super precise, pedantic definition that fails when you flip a switch.

WTF is wrong with UI designers these days? I have my fair share of 'get off my lawn' moments, but I really don't think my hatred for new UI paradigms falls into that category. Designers today seem to have no familiarity with the 'form follows function' concept.

My guess would be that it's kind of reached if not an optimum, a deep local minimum. A huge amount of the awkwardness of early GUIs has been thoroughly hammered out, and the WIMP idea has generally settled into a number of core interaction mechanisms, but with a lot of similarities. The problem is that most of the big variations have been done, and we've kept the ones that worked.

But the only way for a designer to do something to make their mark or stand out is to do a big variation. Trouble is there's not a lot of options left for big changes that improve things.

Also, see Zawinski's CADT: people in their zeal for newness forget all the very hard won lessons of the past. Like GNOME bringing back bits of MDI. MDI was an interesting, but ultimately (IMO) failed experiment. One might argue tabbed interfaces are a remenant of MDI, but in firefox, I can make multiple toplevel windows and move tabs between them, which is somewhat outside of the scope of MDI. I suspect the GNOME people never lived in the Win 3.11 era where MDI was somewhat popular, so never got past the stage of "hey this is cool grouped things move together" to see how that can fundamentally defeat the point of windowing in the first place.

As are billions of others all around the world.

LOL, they or I may indeed deserve it, but you get to suffer it too, so enjoy your schadenfreude. :)