The limitations of AI - Ed Zitron article poses questions.

AI is at a dead end, and the dead internet theory

[Whitey]

tbh - I've racked my brains, and plugged into competitors developments in AI and I haven't found much that we could run with in our business on any scale. Sure, chatbot's and all the usual, but nothing that turns the internet on it's head. That said we're investing in AI architecture, bots etc, but it's more to be strategically in place for moving quickly when needed. And that's not yet arrived.

Or am I completely missing something?

Here's an interesting article from Ed Zitron that's worth reflecting on:

I have been warning you for the best part of a year that generative AI has no killer apps

I shared concerns in July that the transformer-based-architecture underpinning generative AI was a dead end, and that there were few ways we'd progress past the products we'd already seen

A few weeks ago, Bloomberg reported that OpenAI, Google, and Anthropic are struggling to build more advanced AI, and that OpenAI's "Orion" model — otherwise known as GPT-5 — "did not hit the company's desired performance," and that "Orion is so far not considered to be as big a step up" as it was from GPT-3.5 to GPT-4, its current model. You'll be shocked to hear the reason is that because "it’s become increasingly difficult to find new, untapped sources of high-quality, human-made training data that can be used to build more advanced AI systems," something I said would happen in March, while also adding that the "AGI bubble is bursting a little bit," something I said more forcefully in July.

[wheresyoured.at...]

I wonder if all this play's into the "Dead Internet Theory" [en.m.wikipedia.org...]

That should be enough to tease your thinking.

www.wheresyoured.at Godot Isn't Making it Before we get going -” please enjoy my speech from Web Summit, Why Are All Tech Products Now Shit? I didn't write the title. What if what we're seeing today isn't a glimpse of the future, but the new terms of the present? What if artificial intelligence isn't actually capable

[ronin]

Crumbs. That’s the most colossally negative diatribe against Generative AI I’ve ever read.

I don’t think Zitron is entirely wrong. But I think his Cassandra-like predictions are way off.

Towards the start of his commentary he underlined his contention that no killer-app is coming. But I wonder if this misses the point of what Generative AI has enabled: the most accessible computing interface yet.

In the early eighties we typed commands.

In the early nineties we clicked on icons.

In the early 2000s we tapped on icons with a stylus and in the early 2010s we tapped on icons with our thumb and index finger.

By the late 2010s we were talking to devices, albeit at a very primitive level and using a fairly limited vocabulary.

Now, in the mid-2020s, we have all the jigsaw pieces in place to give sophisticated, natural-language instructions to devices, without needing to pause, re-phrase etc.

And what if that’s it?

If that is it, it’s very far from nothing.

From this perspective (the perspective that the value of Generative AI is in using it as an phenomenally versatile human-to-computer interface), asking “What’s the point of Generative AI?” is a bit like asking “Where’s the profit from developing WIMP (Windows, Icons, Menus, Pointer) environments when we already have DOS?” or “Will there ever be a killer-app to justify using a mouse or capacitive glass?”