Reddit gains an irrevocable, sub-licensable, royalty-free license for user-created content posted on its site. So while the term "Reddit data" is somewhat inaccurate, it is a passable approximation of the truth. The more correct term would be "Data that Reddit has irrevocable, sub-licensable, royalty-free license to". But where's the fun in saying that? :-)

"Reddit data" does obscure the fact that the Reddit user retains full copyright ownership of their comments. So if a user wished to, they could independently license their content to GOOG, AMZN, META, MSFT et al. Theoretically, all the other users in a Reddit thread could make same decision, cutting Reddit out of the licensing loop entirely.

But the logistics of organising this are onerous. Even if a few crucial users in a thread refuse to license content, there will be 'gaps' and intelligibility and utility for AI training will suffer.

But if replies briefly quote the specific content they are responding to (as I am doing here), the context becomes much more clear. In that case, individual comments become much more intelligible and useful. Legally speaking, there should be no problem here because brief quotations fall under fair use.

Concievably, GOOG and other browsers manufacturers could offer to store your user-generated content in a browser repository ("keep a record what you wrote", like Windows Recall) or in the cloud, with the option of licensing the content to GOOG. They could also generate AI-summaries of the context you were responding to.

Why don't they strike their beloved South Korean home market with these "value-strengthening" ads first? :-/

Does it violate an unspoken social contract the Korean people have with their chaebols?

A few weeks ago, my 2 year-old Samsung 'The Frame' TV proposed new TOS to me. I think I rejected it... :-\

Your first assumption is correct, my usual ingredient panels are in English. Allergens are helpfully highlighed.

Yes, I occasionally point my phone at English ingredient panels - but only to blow up minute fonts, not interpret their information.

(No, I am not willing to learn Korean - I have an AI for that. Even if I did learn, a little knowledge can be a dangerous. Say I recognized Korean representations for milk and cheese and whey. But then missed the significance of the Korean word for sodium caseinate)

And as far as checking CT images go, I do have a medical consultant. It is a question of curiosity, to see how capable LLMs are.

As far as researching Pubmed papers goes, I do what good PhD advisors do - use free labour if you have access to it.

Looks like I have set off your LLM allergy. Or is it simply an Altman allergy? Well, there's no accounting for taste.

Yes, I have a learning difficulty, you insensitive oddity! I don't understand that things like whey and sodium caseinate derive from milk if the entire ingredient panel is in Korean. And have you ever seen the *amount* of ingredients in Korean processed food? Pages! Your typical 'Whole Foods' product it is not.

  So ... I just tell the AI my allergy, take a picture and have it sort everything out and give me a yea/no answer. And move on to the next product. I could lovingly decipher Hangul or/and brush-up on food chemistry. But I don't like unnecessary cruelty, especially on myself.

:-) Different strokes for different folks. My 3-year old $200 Android phone does nothing of that sort you experience with the default camera app. Moreover, I don't have to interpret "denatured whey protein" as pertinent to my diary allergy -- the AI picks it up and warns me.

My experience has been to successfully use plain text to instruct AI LLMs to:

generate poetry -- even blending unrelated genres that I am sure no one attempted before)
summarize documents
research diseases, medication and side-effects
review CT scan images for issues of concern
generate tour itineraries
translate languages (Hebrew , Korean, ...)
explain code
generate code (Perl, Bash, VBA...)
generate an Excel document with ISBNs and book summaries from pictures of my physical library

That's general enough for me. It should be general enough for you too - LLMs pass the Turing test.

Like you, I have more own interest in my experience than OpenAI's or other AI vendor claims. And my experience has been to successfully use plain text to instruct AI LLMs to:

1. generate poetry -- even blending unrelated genres that I am sure no one attempted before)
2. summarize documents
3. research diseases, medication and side-effects
4. review CT scan images for issues of concern
5. generate tour itineraries
6. translate languages (Hebrew , Korean, ...)
7. explain code
8. generate code (Perl, Bash, VBA...)
9. generate an Excel document with ISBNs and book summaries from pictures of my physical library

That's general enough for me. It should be general enough for you too - LLMs pass the Turing test.

Yes, LLM AIs tend to behave like a skilled, over-confident intern, whose work is useful but must be double-checked. Call it slop if you want, but its useful slop. It got me past the initial inertial friction with VBA for my test analysis project. I had to correct flaws in code structure and parsing strategy. But it ended my 20-year procrastination. Now I have a 1000-line VBA codebase that I am familiar with enough to reuse in other projects.

You ask how does "checking multiple slop hallucinations against each other" help? Comparing two hallucinations will cause the "disagree" alert to light up, won't it? It's not like the two AIs are coordinating via seance.

I was in Bali, Indonesia earlier on, and OpenAI translated to Balinese. Also, OpenAI claims it can translate Turkmenistani food labels. And Google indicate Turkmenistan seems to be catching up in food labelling standards. Entities in the world move ahead of the pigeon holes we slot them in :)

Of course AI makes mistakes. That's why you must exercise your judgement. Also, use multiple AI to check each other's work where important.

Yes, 20 years. What can I say? I really hate VBA. They made me employee of the month after I turned it in.

If AI can translate between me to VBA, Balinese and Korean, what else could I call it but general purpose?

Hmmm..thanks.. isn't a microsoft account the default now for home users?

Dude, with that low a UID have you superceded mortal talk and moved on to communicating in Haiku?

I get your point. But my point is LLMs are truly general purpose. I don't need "an app for this" or "an app for that" or an app for something no one thought of yet. It is intellect on tap. If the question deals with general knowledge and logic, you can ask it and typically get a decent answer.

Standing in the Korean convenience store after my flight came in at night, searching for food after everything around had closed, I didn't need to research OCR and label translation apps for Korean products. And then interpret the output. I could take a picture, ask a 1-sentence question and get the answer I wanted. And do it for a couple dozen products. If I go next to Indonesia or Turkmenistan, I probably will be able to do the same thing there. And a few months ago, I used Claude to write a 1000 line VBA script for Excel. It analyses results of automated software testing and does 'bisection analysis' of results (comparing result before and after a change date). I had been putting this off for 20 years, because VBA is not my skillset.

I dunno about that bro..I am paying OpenAI and Claude subscriptions.

I've given them as much money in a year as I paid Google for a quarter century of use.

Having a general purpose intellect on tap is very handy. I was in Korea recently and it was handy to take pictures of product labels of groceries and ask AI if it contained the things.i was allergic to.
.

If windows pop-up an unwanted app-store screen and I click on 'X' to close, am I an active user?

"Nice business you got there. It would be too bad if 2000 bad reviews happened to it"

As the other guy suggests - there are better ways. One is to trust neighbours over strangers.

So you have a browser plugjn or AI tool to review the reviews.

One heuristics is that eviews by a friend, or a friend of a friend (FoF), or a FoFoFoF.. get more credence the closer they are to you. Or the more their tastes match yours.

Of course, this requires some way to authenticate each review. But that could be as simple as a byline at the foot of each review. An identifier, followed by a digitally signed message hash.. automatically added by a browser plugin just after you click post.

The authentication could be based on the PGP web of trust model