I really get sick of the comparisons to the iPhone and AWS and such.

Those things solved problems. Nobody needed to be convinced an iPhone was a good idea; it was an iPod, a PDA, and a newspaper stuffed into a device the size of a credit card that you carried around with you and worked no matter where you were. That's a GREAT idea. No one needed to be convinced about what made an iPhone (or Android, or even Blackberry) a useful and good thing.

Conversely, AWS made starting web businesses no longer require on premises servers, or really knowing anything about servers. You picked what you needed, and if you needed more at some point, you picked that. You could even dynamically allocate and deallocate servers on an incredibly widely available and robust data-center backend. That's HUGE. Numerous massive companies today may not have ever gotten started if not for AWS and it's now making far more money than Amazon's retail business does, and we know how huge it is because East-1 went down some years back, and a third of the freaking Internet stopped working correctly.

What problem even remotely on this level does an LLM solve?

LLMs can automate some tasks (especially basic coding), but with some major caveats:

- They have rapidly diminishing returns on higher contexts. They are very easy to overload with user information. You can post-train them, but that’s too much into actual computer science to be useful to the average office worker. They can work well as a co-worker assistant in a few cases, but they really can’t replace humans long term.

- LLMs only really work well when given the ability to call on-device tooling. Which, with a cloud system like Copilot, is going to be super tame and underwhelming because of lawsuits and various business deals. Tool calling is only really agnostic when you’re running the model yourself on your own device.

- On the topic of on-device models, there’s also the fact that AI provider companies have been caught evading things like robots.txt and causing so much crawling activity that it effectively becomes a DDoS attack. On-device AI doesn’t solve this totally, but most people likely won’t be pushing their gaming GPUs to the absolute limit 24/7 to constantly hit websites with crawl requests.

>> What problem even remotely on this level does an LLM solve?

Translation problem. Google Translate does job poorly, if you compare with chatgpt. Especially if there are mix of languages in input.

Basic search of common, Wikipedia-level knowledge, to explain something. LLM is good at it.

> Nobody needed to be convinced an iPhone was a good idea

It still took a couple of years, though.

Of course in hindsight its rather obvious. But there were other smartphones and they were (on paper at least) superior in quite a few ways.