It's important not to assume that LLMs are giving you an impartial perspective on any given topic. The perspective you're most likely getting is that of whoever created the most training data related to that topic.

I'm sure there are people who spend their time doing this, but I don't understand the motive. Doesn't one post in comment threads because one wishes to share their thoughts with other humans?

That's one reason why I post comments (and none of my comments are AI-generated). But I think some people cut-and-paste AI responses because they like winning upvotes and running up an upvote counter.

Ehh, you really can't imagine any way that an AI system might escape human control and act autonomously?

    while crypto_balance > 0:
        generate_scam()
        send_out_emails()
        deposit_proceeds_into_crypto_wallet()
        pay_cloud_bill()
        spawn_new_instance()

And who is the owner of the crypto wallet in your example?

Probably a LLC in Nevada owned by a corporation in Vanuatu or the Cook Islands, organized so that voting control is via a blockchain-managed asset.[1]

[1] https://www.311institute.com/ownerless-companies-on-the-rise...

That's a lot of words to say "people".

It has no owner. It's controlled by a self-replicating (or at least self-propagating) AI-based system.

Take my baseless speculation for what it's worth, but could it be that you were depressed because your life was too easy? We humans are meant to struggle through adversity. Can you really appreciate your financial security if you've never faced financial insecurity, or appreciate companionship if you've never experienced loneliness?

It’s a reasonable question but I doubt it. We weren’t affluent at all and I worked my butt off for everything. And that’s good, because I agree that if things are too easy it turns into a curse.

> The top 10% income bracket of the US is broad enough to include basically all US software developers, isn't it?

I wish! My salary is a bit below the median US household income.

"80% as good as the real thing, at 20% of the cost" has always been a defining characteristic of progress.

I think the key insight is that only a small fraction of people who read recipes online actually care which particular version of the recipe they're getting. Most people just want to see a working recipe as quickly as possible. What they want is a meal - the recipe is just an intermediate step toward what they really care about.

There are still people who make fine wood furniture by hand. But most people just want a table or a chair - they couldn't care less about the species of wood or the type of joint used - and particle board is 80% as good as wood at a fraction of the cost! most people couldn't even tell the difference. Generative AI is to real writing as particle board is to wood.

Particle board:

- degrades faster, necessitating replacement

- makes the average quality of all wood furniture notably worse

- arguably made the cost of real wood furniture more expensive, since fewer people can make a living off it.

Not to say the tradeoffs are or are not worth it, but "80% of the real thing" does not exist in a vacuum, it kinda lowers the quality on the whole imo.

How about

- There are 8 billion people on the planet now and there isn't enough high quality furniture quality wood to make stuff for all of them.

Up until the time of industrialization there just wasn't that much furniture per person in comparison to what we have now.

The reason 'real' wood furniture is more expensive is not that there isn't demand or artisans creating it, there are likely more than ever. Go buy hardwood without knots and see how much the materials alone set you back.

The trade off isn't 'really good furniture' vs 'kinda suck furniture'. It's 'really good furniture' vs 'no furniture at all'.

Knotty softwoods can make perfectly suitable furniture. They can (and are) grown at scale.

I’m sympathetic to the viewpoint that the supply particleboard furniture has suffocated the marketplaces for mid- and low-end wooden furniture. Such pieces definitely exist affordably (I’ve bought them at places like Marshall’s, for instance). But they seem comparatively underrepresented in the market.

Maybe a consumer preference for flatpack furniture is enough to explain this? But then again, wooden furniture can be flatpacked too—ikea has plenty of it.

If you make better furniture, it will last longer, and you don't need as much wood to serve the same number of people.

It will cost more, sure, but that keeps people from just throwing it out; they sell it instead of throwing it out. The amortized cost is probably similar or even better, but less wasteful.

Yep I own a rocking chair that my great great grandfather built on a lathe and a dining table my grandfather built. Meanwhile I’ve eventually had to replace almost everything I’ve bought from IKEA.

In some cases that "eventually" has been before putting the damned thing together because the low quality particle board they use can't even survive shipping.

(per capita) buy one cabinet every time you move (they break if you try to move them), or buy one quality piece of wood furniture and resell it when you don't want it.

it's disposable plates vs dishwasher ones, but particle board vs actual furniture

You did not read my comment very well. I was not commenting on the the particle board tradeoff, or even the AI tradeoff we find ourselves in now. I was saying that reduction to a lower common denominator (80%), even though it seems innocuous, actually does have broader effects not usually considered.

Who said anything about particle board. There is factory created furniture that uses long lasting high quality wood. It will last generations and is still less expensive than handcrafted furniture.

The main issue with AI is that it still relies on the industries it destroys for its training data. You still need people to cover the news, to review products, to write about their feelings and to find new things to talk about. AI just denies these people an audience or attribution.

This is where the particle wood analogy falls apart. IKEA creates its own goods. AI relies on the work of the industry it's destroying.

> Generative AI is to real writing as particle board is to wood.

Incredible analogy. Saving this one to my brain's rhetorical archives.

One law I would like to see if expected durability. Food has an expiry date and ingrediant list. Something similar should accompany all products so consumers can make an educated choice how long it's gonna last and what's gonna break

"Nice metal <thing> you have there, would be a shame if one of the critical moving parts inside was actually plastic."

Sure it's awful but look how much you get.

After spending some time prompting LLMs and then talking to actual people, I've once or twice been tempted to phrase questions along the lines of "Comment on x, with an emphasis on y; do not mention z."

Surely, someone must have already thought of this. Perhaps what we really need is a curated index of curated indices.

"Risk of internalizing spurious explanations" is an excellent way of putting it. LLM output is, essentially, a polished-looking, authoritative-sounding summary of what the top few Google results probably say about a topic. Nine times out of ten, the explanation may be spot on. But "the first few google results" are not, in general, a reliable source. And after getting nine correct answers in a row from the LLM, it's unfortunately very tempting to accept the tenth at face value without consulting any primary sources.

I've been finding that ChatGPT is helpful when taking a "first dive" into an unfamiliar topic. But, after studying the topic at greater depth through primary sources, I'll start to see many subtle errors, or over-simplifications, or claims stated as facts which are actually controversial among experts, in the ChatGPT answers. Overall, I'd say ChatGPT can provide a good approximation of truth, which can speed up research by providing instant context. But it should not by any means be the final destination when researching a topic.

The social security "tax" should really be conceptualized as an investment, not a tax. The typical fast food worker has probably not passed the first bend point in the Social Security PIA formula, meaning that social security is giving them 90 cents on the dollar*. You, with your fancy tech job, are likely well past the second bend point: social security is only giving you 15 cents on the dollar* (and nothing, obviously, for earnings beyond the payroll tax ceiling).

It's a progressive system overall - but it wasn't designed for the purpose of wealth redistribution, hence the payroll tax ceiling.

* More precisely, their monthly benefit at full retirement age increases by 90 cents for each additional dollar of pre-retirement average monthly earnings, whereas yours only increases by 15 cents.