Much less than seeing new software products being shipped, we havn't even seen attempts at them being shipped, or built for that matter, again excluding this sort of meta software around the AI itself like harnesses and tooling.

We could speculate that attempts have been made but not shipped, and the details of why not being revealed, but then we're just back to it being talk with no demonstrable stuff happening in the world that backs it up.

There's been this continual allusion to ever-narrowing circles of influence and ever-accelerating capabilities behind the curtain, but in fact things do keep coming out from behind said curtain and while their capability is indeed advancing it doesn't really seem at all like it's accelerating, even, let alone vertical.

Also at what point do we raise the fact that despite the ever narrowing circles of access, nobody within those circles seems to be able to actually do anything with it that's ostensibly what it's sold to do? All they seem to be able to do frankly is build more models and harnesses and tools designed to consume more tokens from the models they build, but not actual directly useful new or innovative software that generates what could be very real revenue.

Eventually when you keep talking about exponentials but not that much seems to be demonstrably changing, the narrative is going to run out of steam.

There's no way I'd imagined it wasn't done in a professional setup by a full team.

A real standout part is how he commuted by bus to the studio for years, to work on it mostly alone. Truly amazing dedication to a passion project.

I haven't watched it in probably 25 years but I'll find it now and appreciate it all the more for knowing this!

I keep seeing this idea and I don't understand the reasoning behind it.

I think it could be a bit like saying if you showed someone 500 years ago a smartphone they would likely conclude at first it was magic. But once you had some time to let them use it and tell them how it all worked on a high level they would eventually obviously realise, no, it's not magic.

I guess just in the same way if you presented current LLM tech out of nowhere a few years ago to someone who'd never seen it, I concede they may be likely to imagine it was AGI in that first conversation, depending on their background.

But after using it for a bit and learning what an LLM is etc they'd land exactly where everyone is today - a great technology useful for some things, not AGI, not magic.

Double decker trains I just find fun for some reason, and you get slightly better views riding up top over fences and walls and such.

It does seem like a bit of a trend in engineering that the first to get there set slower standards that others then copy, speed up and otherwise improve upon. See Britain and trains.

On that note, the very same Jesse W. Reno tried to build a spiral escalator in London in 1906 in Holloway Road tube station, the shaft for which still exists today. It never made it into production aparrently, but goes with the trend.

It makes the future feel fun, and feels like a return to adventurous aesthetics for the pure sake of it. At the very least, it's noticably different and interesting. I hope architecture follows suit.

To be honest it is worse than this. Spotify over the years just seems to have gotten worse and worse at the one job I want it to do well which is just play the music I already know I want to play, which should be incredibly obvious from my play count and what I've downloaded.

Instead I'm regularly surprised by random noise such as popover in-product ads for extra features I don't care about, and simple AI features I might want like detecting I am in this certain place at this certain time where I usually listen to these albums or playlists so maybe weight them accordingly in a quick access list on the home screen or something aren't, for whatever reason, done.

I mean this doesn't even need LLMs it's just an algorithm, but you see what I mean. Using intelligence in the ways I should obviously care about from the way I use the product, instead of trying to get me to change the way I use it (I won't, I wish there was a way I could just definitively tell them this).

Alas, after years of annoyance with this it's still just about above the bar of useful enough, so I keep paying, and just can't be bothered to invest the time switching to some other service like apple music. So in a sense I'm probably encouraging this behaviour by not churning.

I might be misunderstanding your point but I think you're in agreement with the gp, as in they are arguing for a locked down interface behind which the AI sits, narrowed to use within whatever particular operations the user can access, as opposed to a wide open AI interface with access to any operation (in principle).

So in your example there's no way to ask the AI for any old discount because the interface for doing so is locked down to just the discounts you can access, although the AI may be able to apply a discount you didn't ask for dynamically, if allowed, to give the customer a better experience.

Of course both versions can be implemented securely, it's just probably in general a smaller attack surface if you provide a tighter interface first.

One thing you notice in kids in their native language is they get the trickier/irregular parts of grammar wrong all the time and repeatedly, as in they are corrected but just carry on making the mistake, I guess until one day finally the right pattern gets learned through sheer brute force of repetition. In the meantime, they're perfectly well understood by everyone anyway.

Obviously as an adult you can somewhat shortcut this because of better understanding of grammar and knowing how to learn, and being quicker to absorb and make the changes, but regardless I think the basic premise still applies - you learn the right way over time just by doing it over and over and over until you can't not remember it is the right way.

And, to be honest, if the wrong way is understood, how practically wrong really is it? (you will always always have tells you are not a native speaker no matter what, some of these can even be charming, I don't know if it should even be a goal to get this to zero at all).

I think it's pretty likely that in, let's just go for the round number and say 80% of cases, everyone will conclude that just existing well used frameworks / code generator templates where everything's understood and most of the hot paths are well tested is pretty fast to build with (and more importantly, maintain and operate) and was really all we needed for most useful business software.

Everyone else also thinks like that and so isn't normally unambiguously critical, so you end up not knowing who really thinks that way and who's just doing it because of the game theory.

It could even be practically nobody. There's no way to tell until the deadlock breaks.

I do it too so it's not a criticism, but I find it kinda interesting culturally. It's a reflection I think of the environment that we've created for ourselves, or has been created for us (or some combo) where somehow you can't point out an unqualified negative of this tech.

This to me is actually one of the clearest bubble signals, just because if it really were as good as all that, it'd be completely redundant to remind everyone of that when criticising it.

I think it might be a sort of weird collective psychological thing where we must not admit what seems pretty clear, just for fear of breaking from norms.

But, of course, yeah it's very useful for some stuff and I use it all the time :-)

Comments URL: https://news.ycombinator.com/item?id=49126118

Points: 2

# Comments: 0

It's probably just the old 'fake loader' psychology, to be honest. Waiting without knowing when the elevator will arrive is boring/frustrating, where getting into an elevator that moves, even if the wrong direction, feels like progress.

Things are happening, and even if it takes longer ultimately that's a less frustrating state to be in.