[rtsil]

Off the top of my head, and just for the last couple of months, and only outside of work (where its value is even more immense), it has saved one of my indoor plants, told me how to handle a major boiler problem that would have left us without a working boiler during a weekend in the winder, with the next "emergency" repairman only available on Monday, advised me to use Kopia as backup solution for my personal files instead of Syncthing, helped me choose the right type of glass for a painting frame, answered a couple of questions about bikes and helped me when I was stuck in an harmonic analysis of a piece of music. All of that are extremely valuable to me (if only for the time not wasted googling answers), and in none of them its potential hallucinating would have been an issue. And I can't count the number of times where "specialists" in bike repairs or plumbing told me something incorrect or outright false, so I've learned to deal with hallucinations already!

[vel0city]

> And I can't count the number of times where "specialists" in bike repairs or plumbing told me something incorrect or outright false, so I've learned to deal with hallucinations already!

So much this. So many times I've argued with hired experts saying "can't be done" just to see yes, it can be done.

[swatcoder]

Yes, but which of those things would you not have resolved just as well 10 years ago? All those possibilities were added by the maturing web itself, as a genuinely novel change from having to source books or experts/friends in the days before.

I'm glad ChatGPT didn't lead you astray, but I'm not seeing what it's added here besides shuffling up the user interface in a way that you presently and subjectively prefer?

[TeMPOraL]

> I'm not seeing what it's added here besides shuffling up the user interface in a way that you presently and subjectively prefer?

This. But in the same sense the past 50 years merely changed interface from dusty textbooks in libraries to Google Search, and the past 100 years gave us dusty textbooks over writing to Royal Society, and that just replaced the option of asking a local whisperer or hoping you'll find answers on the Sunday mass.

Do not underestimate the power of being able to get an answer to your problem described, visualized, and perhaps complete with interactive demo to explore it further, in time it would previously take you to formulate the right search query that finally gives you relevant information.

EDIT:

And that's on top of all the arbitrary data transformations prior tools couldn't do. E.g. I'm increasingly often using GPT and Claude models to turn photos of (possibly hand-written) notes or posters into iCAL files I can immediately import into our family shared calendar.

Another frequent use case, data normalization. Paste a whole dump of inconsistently structured data multiple people collected (say, addresses of various local businesses that helped a local NGO and now are supposed to get a thank-you card for Christmas). Like, you get 200 rows of addresses in a single column, with spelling mistakes, repetitions, junk at the end, arbitrary capitalization, wrong order of address segments, and such; you need to separate it out into 5+ columns (name line 1, name line 2, street address, zip code, city, etc.) and have it all normalized.

The fastest and most robust way to do it as a one-off job, today, is to paste the whole thing to GPT-4o or Claude 3.5 Sonnet, tell it how the output should look (give one-two examples, mention some mistakes you saw), then send the message and wait 30 seconds for the job to be done for you.

(Yes, it may make mistakes - it didn't for me in recent memory, but it can. But for that, I quickly add an extra verification column for each one in LLM output, and do a simple case-insensitive substring match with original, and eyeball any data row that shows an error. And guess what, the formulas don't take much time either, since LLMs are good at writing them for you, too!)

[rtsil]

My plant would have been dead. As for the rest, sure, I would have resolved them eventually, after many frustrated hours of googling and trial and error.

Time is my most precious thing, I already don't have enough time to do all the things that I want to do, I don't want to waste that trying to find and test solutions when ChatGPT gives me instant answers. I'd rather spend time playing with my cats or riding a bike instead. It's not a matter of UI, it's a matter of preventing waste of time, energy and money, and less frustration. For that alone, €20/month is a very good value. And that's just for my personal life.

[swatcoder]

"many hours of frustrated googling and trial and error" isn't a familiar experience to me, but I'll trust that it is for you. I'm glad you see that as behind you now with this. I suppose you must not be alone.

[djeastm]

>besides shuffling up the user interface

I wouldn't discount this effect. As someone with sensory issues, one thing I like about ChatGPT as opposed to the "raw" internet is that I can see the answer to my questions in a nice and calm textual format without some website who created the article specifically to catch my search terms, but is trying to get me to deceptively click on ads or pull me into buying something through their affiliate links. That's absolutely increased my own enjoyment and productivity.

[fragmede]

objectively, it takes less time to ask a question and get a direct answer than it does to search for some words, leaf through a couple of results, find one that has the information you want, and then read that page. If I want to know the height of the Eiffel tower, being told it's 1083 meters tall is faster than searching for its website, finding the stats section, then locating that information on the page. Google realizes that, so they pull that info out of the page and just put it on the results page for you.