I kinda get why they (and other companies) have to try AI at the moment though.
It’s not what people claim it is, but it could end up being an essential tool for the modern world, and if they don’t invest in it early their business might end up getting left behind.
We’ve certainly seen companies fall because they’ve not tried to stay on the cutting edge before.
We’ve certainly seen companies fall because they’ve not tried to stay on the cutting edge before
Best example I can think of is Kodak and digital cameras. They invented it then sat on it until it was too late because they didn't want to cut into their film scam.
It’s quite unbelievable that it was literally right there. The logistics were like 60% solved for them already, the remaining 40% was just making sure the online content remained linked with inventory and fulfillment, and expanding that capacity.
“We think online shopping will be just a fad” - the unimaginable hubris…
I was refreshing myself on wiki. They launched prodigy, but it was too early for online shopping to be popular. So they probably got a bad taste for that kind of thing. A concept in venture capital is that it's all about timing.
Ohh I forgot they did Prodigy. It was such a WEIRD experience; it seemed to drop you down into this BBS/DOS mode, and none of the navigation was very similar to everything else on any PC/Mac at the time. I enjoyed some of the games on it but I really didn’t know what to think otherwise. We didn’t keep it very long.
Mozilla has a finite amount of money. If they're (as far as I'm concerned) wasting it on AI nonsense, that's less development funds that can go toward Firefox.
While ML does have legit uses in many specific cases, this whole "throw 'AI' into everything" hype/trend is just blockchain all over again. IMO, the ones who are overreacting are the ones swept up in the hype.
You know there's two sides to hype, right? Both the positive and the negative. Constantly doomposting about how any mention of AI, or LLM, or ML, or whatever your preference is, is a sign of being a waste of everyone's time isn't any better.
Maybe instead of jumping from one extreme to another we can go ahead and have some kind of middle ground where we don't jump down the throat of everything we only have the barest concept of because it looks kind of like something we aren't fond of.
In that there is a finite amount of money, there is also a finite amount of development that can go on at once. If they just pile tons and tons of bodies on what you might call useful endeavors, it can lead to bloat and the right hand not knowing what the left is doing.
I hate to see AI (I suppose we mean specifically GPTs in this instance) trashed all the time, just because companies use it incorrectly. They shove it in every hole they can to hike the stock price. But it's a great tool, that arguably needs more time in the oven, which has legitimate helpful uses. Especially in the context of a browser.
For example, in Arc Browser I can semantically search the page/article for anything and it will show me the location of the information I need (ever tried to find the recipe itself in an article about the recipe?). I can also do some obvious stuff, like summarize and translate sections, which I could do by copying it into a dedicated service, but it's definitely much more convenient being built-in.
Would be much better if it ran locally off the NPU, but we are not there yet.
Because this comment reminded me of it, here's a website that strips the garbage out of bullshit recipe articles. Copy/paste the url of the recipe into the box on this site.
Downvotes from the people who believe that all “AI” is an LLM/GPT that must be trained on the collective stolen works of all humanity and requires all of South America’s collective power supply for just a day’s worth of queries