Desktops shouldn't be simple. They should work reliably and predictably like every good procduct. Everyone wants something simple until they want to do something that's not supported because it's not a simple use case. So unless you are a very simple person who always only does very simple things and lead a uniquely simple life (i envy you), a simple desktop is not for you.
These issues pop up everywhere, not just with JS. You can (and people do) write horrendously complex and laggy software (e.g. using "frameworks") in any language. Lots of excellent pieces of software are written in pure JS.
There is a good solution. It's actually a great solution: Write everything in plain JavaScript. You get a great language to develop in. All your problems will go away. No dependency hell. Excellent load times and performance. Superb compatibility.
I don't use any of facebook, instagram, LinkedIn, tiktok, X, bluesky, mastodon, snap, ... and it's not even hard. On the rare occasions I took a look at those I found them boring beyond believe. Even mastodon following people I am actually interested in feels like a 90% waste of time.
The paper confirms that even the biggest and most advanced models can still be misled by adding irrelevant information that can be easily filtered out by humans.
Based on my own experience it is still fantastic. Not long ago I had an issue with an order and a real human called me back right away outside normal office hours. No waiting in a phone queue. And he was actually able to help me. If that is not excellent I don't know what is. I don't know of any other large company with a remotely similar customer experience.
I don't think it's hyperbolic at all if you look at the published data, development of past and planned future energy requirements for AI. And as if efficiency gains ever stopped anyone from using even more energy. See https://en.wikipedia.org/wiki/Jevons_paradox
> I think it's important to keep in mind that we're very early in the AI journey.
That's what I am saying. At the moment there is this one really dumb idea, that bigger is better.
Let's just assume for a moment that the hype is real and that these LLMs are incredibly intelligent and will replace us all soon.
Then the model shouldn't be any less intelligent if we remove facts like Uma Thurman's measurements and other vapid information. If the model already has the capability to use tools than all of that crap is redundant anyway.
And while we are at it let's remove a ton of other junk like languages I will never use and which also doesn't make the model any smarter.
So how small can this kernel get while still being clearly intelligent, able to communicate flawlessly in english and apply logical reasoning.
That would be a worthwile endeavor and maybe even possible without boiling the oceans.
Well if they are paying real people to answer my questions I would call that a pretty good deal. That's exactly my point. As a user I don't care how they come up with it. That's not my problem. I just care about the content. If I pay a human for logical reasoning, train of thought type of stuff, I expect them to lay it out for me. Not just give me the conclusion, but how they came to it.
This was so obvious from the start. Whenever big tech creates "standards" now you already know it's going to be total horse shit. Look back at the old threads when passkeys launched. HN was full of fanboys thinking it's the best since sliced bread and passwords are so yesterday. Managing your passwords takes a bit of effort like everything in life where you don't want do give away control completeley to some corporate aholes. Whenever you let someone else manage your stuff you set yourself up to getting ducked.
Of course all animals are conscious to a degree. Otherwise you would have to draw an arbitrary line somewhere which would contradict the gradual nature of evolution.