You know, I've mentioned this quite a few times before, but I'm like this old dog who tends to overthink everything 😄 I really do read into things a lot.
But here's my take on wheels—yep, I'm talking about wheels.
In life, and particularly in coding, nobody wants to reinvent the wheel. Coders love to stick to what's already there; that's the whole point of wheels, isn't it?
Our rule is pretty simple: if it ain't broke, don't fix it.
We only step away from this when we see that the existing wheels just don't cut it. They might be broken, not good enough, or just not right for what we need. Then, and only then, do we get into the whole cycle of reinventing the wheel, doing it over and over until we get it just right.
So, if you've got some newfangled idea for a better wheel, you really have to sell it to us. We're the ones who are going to use them, after all.
And let's not forget, we're also the ones who'll be footing the bill.
If we don't see a good reason to switch, we just won't. We'll stick to our old wheels, no matter how much you try to push the new ones on us.
Honestly, if we've got wheels that work just fine, we're not in the market for new ones. We'd rather find new roads to explore.
Better wheels? If they're so great, why aren't you using them yourself? That's what I'd wonder.
I might give those flashy new wheels a quick glance, but I'll always come back to my trusty old ones that work just fine.
Especially if these new wheels mean learning a whole new system, with a steep learning curve and not much benefit to show for it.
"The wheels we have aren't just good enough; they're fantastic on any terrain. That might be your selling point: the new wheels run well on your turf. But the old wheels perform just fine on all terrains, including yours." Some might argue🤗 Just saying.
Well, that would be my take on wheels, too.
"No matter how hard you pitch the new ones, if our current wheels seem too good to be replaced, I'll just stick to my Warthogs." That's what I'd say.
Just my two cents, from someone who's been on all kinds of terrain you can imagine, and even some you can't.
## Notes on Apple MLX
Some seem to misconstrue my intention when I say, 'You are Apple,' or 'They are Apple, after all.'
Once more, I'm not a fanboy of anything. I reserve the right to be critical whenever I encounter something that doesn't seem right.
As noted in my earlier essay, I've been on all kinds of terrain you can imagine, and even some you can't.
When I say 'You are Apple,' I mean that I want to hold MLX to the high standards associated with Apple's reputation. Therefore, I am earnestly appealing for enhanced support and development for MLX, reflecting the level of excellence that we associate with Apple.
The present circumstances prompt me to wonder: In this era of AI, how many at Apple are truly dedicated to this pivotal project? Certainly not as many as I would hope.
When we declare "Apple is the most valuable company on Earth," we're not typically using a metaphor. This statement literally conveys that Apple's market capitalization surpasses that of any other company worldwide. This could vary with time, but currently, as I write, it holds true.
This straightforward fact is often misunderstood, particularly by non-English speakers, like in Korea where the term '가치있는' (valuable) generally implies a metaphorical context. But in this instance, it's a literal statement based on figures, not a lofty depiction of Apple's attributes. It simply acknowledges Apple for boasting the highest market capitalization. Therefore, when I allude to "You are Apple" in relation to MLX, I'm making a similar comparison. MLX, with its less extensive documentation and support, is likened to a modest company when weighed against other frameworks.
I acknowledge that it's an emerging framework appreciate their efforts from the bottom of my heart, but once more, let's be clear: "You ARE Apple, for chrissake."
When I look at PyTorch, I do see reflections of Meta's current state: one of the 'Magnificent Seven' tech giants. One of the most valuable companies on Earth.
But can we genuinely see Apple in MLX? No matter its potential for growth? I'll concede it’s an emerging framework, but, to iterate, "You ARE Apple."
And that's the reason I've paused further documentation of MLX until I can ascertain Apple's genuine commitment to its development.
Please, don’t get me wrong.