So $150 million to train the two models?

Seems like a reasonable capital cost if it worked and was legal.

Charge subscriptions of $10,000 year to keep something up to date. You wouldn't need a profound amount of customers to cover it.

I'm skeptical this company is doing it properly (or even has their own models), but I think you could do this with two models.

The documenter is trained on all available data.

The coder is trained but without any copy left code.

Clean room reverse engineering actually seems like a place where AI will be extremely capable.

With open source you can document a lot more than with closed source.

Generally if the implementors have seen the original then it's not clean room.

It seems to me if one AI writes documentation and then another that never saw the code in question writes code it'd be legit.

How is that illegal?

NSA pressured NIST to include compromised parts into elliptical curve encryption.

It allowed for a private key (presumably held by the NSA) to greatly reduce the difficulty of breaking things using that part of the suite.

Though Apple started getting relevancy again with the iMac and iPod of the late 90s. It was stuck with OS9 which made Windows 98SE and especially Windows 2000 look great (Windows ME not so much).

OSX came out 25 years and 4 days ago (I just checked), I think your timeline is optimistic by a couple of years when the whole software ecosystem (except quark xpress) got moved over.

Early on the best example of OSX native software was MS Office and not much else.

Yes.

Using the kernel developed by Next was a huge deal. It doesn't change the fact that 25 users ago Apple was an absolute mess (thus the need to purchase Nexr for the OS (and CEO).

Os9 was an absolute outdated disaster.

25 years ago Apple was still OS9

It was constant crashes that required reboots.

OSX came out almost exactly 25 years ago and within 18 months (as software moved to it) changed the whole story.

It's super sad.

I purchased my Vizio TV because it's only smart feature was that it had a built in Chromecast.

No apps, no ads, it just showed up on my phone apps to let me put them on the TV.

Then after 18 months a huge update and it became annoying just like every other TV.

Cars already do this. I'm not sure why they're acting like it's new.

My 2006 Honda lets me know when it's time for service starting a few thousand miles early.

Exploding in the launcher is bad.

But if it's a matter of 1% vs 0.1% failing mod trip. Or even 10% vs 0.1% failing mod trip this is a game changer.

I'm picturing something as hard to intercept as the better Russian misiles that delivers a payload similar to shaheed for the price closer to a shaheed.

Such a weapon would be a game changer even if it isn't some type of unstoppable munition.

The only feature you list I'm remotely interested in is the mirroring to the dash.

But there's no reason Android auto/apple car play can't do that without deeper integration.

WRT to the other controls. Buttons please. I can afford an extra $1,000 or so to have cruise and climate control be buttons rather than voice and a screen.

I use voice to control my home HVAC, but a car is a small controlled space and buttons are far superior for that.

And yet Ukraine is having some success with it's slow af cruise missiles that cost 5x as much (significantly larger payload I assume too).

At $99k these are starting tondrop towards high end one way drone prices.

If I were a country with small GDP I'd love to be able 40x the cost of interception by having the other side blow a patriot.

At such a low cost these don't need to be unstoppable to be useful, just requiring top tier anti air vs interceptor drones makes them a huge game changer.

I thought no this is the correct response.

The people that whippee up something quick for a tech demo failed, but in the long run artists will be able to properly use it to make things look better.