A very simple reply, cataloging my experiences:

I am a programmer in healthcare working with data. Since the advent of code-capable LLMs, they have been incredibly useful to me (and my coworkers) in doing large amounts of relatively low-level programming or mapping work that is boring and unrewarding, but needs to be done. Hundreds of hours of my time saved per year, letting me focus on things that LLMs cant handle yet. Yes, this has resulted in tens of thousands of dollars of benefit to my employer, but it is not labeled “AI” in any way that an accountant could measure.

We use LLMs to take notes on our meetings and create a central documentation resource that keeps track of projects, frequently asked questions and documents that other coworkers would want. With the right skills document, they can automatically update project trackers and tickets from those statuses, saving time that we’d otherwise have to tediously document things.

Since I am a programmer first, there’s often a huge amount of complexities and jargon in healthcare that were not immediately apparent to me. The amount of text created in healthcare is enormous. LLMs let me distill that down to the relevant parts I need, and explain the details. Healthcare has massive, complex contracts regulating prices - LLMs can easily strip them out into simple lists that I can use in my code.

I design computer games as a hobby. LLM coding assistants let me automate away much of the tedious parts and let me focus on the fun parts. Or, use AI to generate images/text for them. I’m not shipping any units, it’s just fun, and LLMs make it more fun.

I use LLM advice to invest. Maybe I would’ve made the same decisions, but now I can collect the information faster, and also LLMs are great for explaining and comparing various tax implications. Definitely ROI both in time and money here, but you’ll have to take my word on the numbers.

I am not by… nature… a person who liked the outdoors, or nature. But when I started using the app iNaturalist, where I can take pictures/sounds around me and it’ll identify the animal/plant, it became a lot more fun to take walks and annoy my kids by taking a picture and identifying every silly little mushroom on hikes. So yeah, just trivial unmeasurable mundane life enjoyment.

My kids love using simple AI to make funny pictures of whatever they want, or do silly things with their faces (or mine). Or make silly songs. Or tell them what words mean in every other language.

When I research stuff online to write about, or discuss with people, LLMs provide excellent research ability, collecting together wide ranges of data/sources/papers in ways that would take a ton of time.

LLMs are excellent assistants for home-repair tasks, because you can talk it through with them. Sure, largely they are just collating previous excellent human-created stuff from reddit or youtube, but not having to do that search while you’re elbow deep in your furnace is pretty valuable.

I have a neighbor who puts up extremely esoteric flags. It drives me crazy if I can’t figure out what they are, so I describe them to LLMs to try to figure them out. Turns out he put up an Easter Island flag around Easter, and a Christmas Island flag around Christmas.

I got all of these from just looking at my previous chats, mostly with Claude (recently) and ChatGPT, Perplexity and Gemini in the past.

Definitely I can’t quantify this. But I can say that it is providing me with lots of value - and all of these cases (other than at work) are from the free versions.

Your mileage may vary. If you aren’t a programmer, then sure, I can see the benefit being a lot less. But it’s absolutely there if you want it, just mostly in consumer surplus rather than direct productivity benefit. And, it is very easy to see how the use cases will expand.