Brian Maucere

You know, and stuff like that.

But it does bring up the larger fact about that.

So to me, I don't know.

Going back to Spark, it sounds to me like lighter, faster, you know, lower latency because there's going to be tons of stuff that when you're using codecs or, you know, a clog code or otherwise.

That just don't require full power.

And this allows us to move faster through those iterations and then bring in the big guns codex when necessary.

Is that kind of the gist you guys got is the it's not one or the other.

They say an example where they build a snake game very quickly, Beth.

wait for my snake game right you need a snake game and you need it now you're on a plane you've got your phone out you got your laptop and they're like we are pushing back from the gate beth and you need to go on airplane mode and you're like i'm out of time but i need to have this snake game i am no veruca so like i don't um

Well, I think for real time stuff like Andy's talking about with translation, that's another one of the examples I'm looking and it gives us side by side between codex and codex spark.

And it said, um,

it gave it a document and sold it to translate it into Spanish, German, Italian, Polish, and there was one more on there, maybe French.

It went too fast.

And it completed all of that in the JSON files and exported it in 18 seconds.

So if you look at like ultra low latency a la speech translation and near real time, we're kind of there anyway.

You could see it was maybe like, maybe Spark has something, there would be benefits to code running that quickly, probably on device too if I had to guess down the road where my phone is able to run codec Spark or some future generation of this, right?

Maybe we're not there today.

But it's processing ultra fast on my phone and not having to go to the cloud to do ultra fast, near perfect or perfect translation.

I could see something like that maybe.

Because then you're literally running net new code always.