Its unclear how they intend to fix these fundemantal problems tbh. Things like "Automate the 'must not fail' moments with rules, APIs, and triggers." And "Use policies, templates, function calling, and explicit do/don't constraints." Make sense but if you have deterministic workflows, what do the llms still add?

Using APIs makes sense but isnt the whole point of these things that they can automate away stuff, it feels like we're building really big complicated frameworks to put these things in. Does it still have any actual benefit for stuff like this?

I'm not Salesforce obviously, but a combination of LLM as an input interface and the deterministic APIs doing the recurring automation behind would work.

For instance you talk with the LLMs for a while, they give you a workable set of DSL commands, you check what they do and make sure they match your needs, and set them to run as frequently as needed.

This pretty much matches my preferred workflow for scripting with LLMs: don't ask it to do the task, ask it to write a script that does the task, then validate and execute or refine.

To save face and make it look like an iteration rather than a reversal.

LLM's still make stuff up routinely about things like this so no there's no way this is a reliable method.

It doesn't have to be reliable! It just has to flag things: "hey these graphs look like they were generated using (formula)" or "these graphs do not seem to represent realistic values / real world entrophy" - it just has to be a tool that stops very advanced fraud from slipping through when it already passed human peer review.

The only reason why this is helpful is because humans have natural biases and/or inverse of AI biases which allow them to find patterns that might just be the same graph being scaled up 5 to 10 times.

I hope I'm wrong but I haven't seen anything like this in practice. I would imagine we have the same problem as before where we could use it as an extra filter but the amount of shit that comes out makes the process not actually any more accurate, just faster.

Having seen from close-up how these reviews go, I get why people use tools like this unfortunately. it doesn't make me very hopeful for the near future of reviewing.

that's why I wanted someone to actually do a real test to see if it has even a shred of accuracy since that means it could be improved upon in the future.

Nobody should be using AI as the final arbiter of anything.

It is a tool, and there always needs to be a user that can validate the output.

Appropriate compensation is a non-issue? I have the impression many people jump on the hate-EU train for no other reason than there's many comments reinforcing it.

What do you really think about this case in particular? I'm pretty curious where this comes from.

Who should receive the compensation? If I want to know the answer to a particular question and most search results point to SEO garbage which doesn't even answer it, then who should be compensated and for what? If those SEO garbage websites are to be compensated, doesn't that just incentivize more garbage?

I don't know. I don't really care about the details in this case, I just don't really get the dismissive attitude that often surrounds things like this. Do you think this is not something that is worth looking into if it happens at such as large scale?

Just do be clear, I use genAI all the time for finding info and answering questions, so my browsing habits changed as well. I'm the kind of person who this case would indirectly be about. But don't you think that it's valuable to look at how do we compensate people who create content when their content is being used by genAI.

Many people seem to have the feeling of 'oh it's too late and those websites were garbage anyway (whatever that means), who cares'. Don't you think that's a bit of a silly way to go about this?

> But don't you think that it's valuable to look at how do we compensate people who create content when their content is being used by genAI.

But why should we compensate them simply because their content is being consumed by AI? For me, any kind of compensation MUST take relevance into account, otherwise we'll reward quantity and not quality, thus quality won't be preserved.

Maybe the answer is to actually NOT do any compensation like that, instead focusing solely on attribution so that it's in people's interest to reward select creators manually to keep the content valuable.

Appropriate compensation for what? The summary is generated on the publicly available information.

If using data from those websites in a way decreases their visitors or something similar then I think there's an argument to be made for that. I don't know the details to case but just because something is publicly visible doesn't mean that you can just do anything you want with it.

Every major news site in Europe is full of articles full of "The New York Times reported that [summary]" so I'm a bit confused as to why, all of a sudden, it's a problem.

Newspapers have been doing this for at least a century, while news radio and news broadcasts have done it since their inception.

There is no guarantee that a website would get a visit if there was no AI summary. Also you can do anything you want with public domain information. That's the whole point of it being public. Otherwise it should be licensed or copyrighted content.

Almost every news article you come across is copyrighted, and is not public domain.

I just read the 'original affluent society' and (most of) your linked essay, I kind of agree with you. That said, the conclusions of Kaplan lead to estimates or 35-60 hours a week (excluding some depending on the group) and that surprised me a lot. That's very different from the image I got from some other comments in this thread talking about extremely long days with constant back-breaking work. Would you agree?

Constant, backbreaking work was not a feature of hunter-gatherer societies in the way it was of early agricultural societies, yes; at the same time, they still worked equal to or longer hours than we did, at things we would likely consider quite grueling and boring (mostly food processing), and what they got out of it was a level of nutrition even they regularly considered inadequate; moreover, a lot of the reason the average per day work estimate is so low, as the paper covers briefly, is that there were very often times, especially during the winter, where food simply wasn't accessible, or during the summer, where it was so hot it was dangerous to work, so there was enforced idleness, but that's not the same thing as leisure.

Whats the problem with attached bottle caps or volume warnings? I used to find these things annoying when I was younger but I do realise things like that can be very useful, even though they are small steps.

What do ILVA and AM stand for?

What is so fundamentally different about DID proposed in the UK or the US then? I read through some of the documents about it and the data scoping that will be available, which isn't with something like BankID seem to be the only difference. What am I missing here?

If an authoritarian state tells a bank to block you as a customer you get exactly the same result. All these options of blocking people are already available to states in general.

Very different levels of friction, though, and that matters too in practice.

It's a pretty common thing that replanted forests turn into monocultures that don't have a lot of value for biodiversity. This then leads to all sorts of problems that healthy diverse forests don't have. I don't know if that's the case with the Appalachian forests but this is depressingly common. That being said, there are good steps being taken, e.g. the rewilding projects by mossy earth.

It wasn't replanted. It just grew back when the farms were abandoned.

This wasn't because a Chinese CEO was appointed, it was because of his recent actions moving (apparantly critical) production away from the Netherlands. Where did you get this angle from?