DeusExMachina

Anthropic is going to watermark all text that comes out of Claude and give companies the tool to detect it, but not the users.

Now they are also blocking specific uses of their models, with a blast radius that also affects people with similar but unrelated use cases.

Add that to all the safeguards they already have added to their models for copyright infringement, security, and other "safety" features.

It does not seem to be a great strategy when faced by fierce competition from cheaper and/or open-weight models from countries like China that do not care about any of these.

> As someone who is aggressively pro AI

But then

> I've also noticed how awful everything is to read lately

> A site I read stories on has been flooded with AI generated content, all of it terrible

I'm not taking either side, but I'm curious about how you reconcile being aggressively pro something that, in your own estimation, is making everything terrible.

Are they integrating it well with AI overviews then? It looks like they haven't found a way to monetize it yet.

> I didn't even realize I had signed up for a one-year commitment.

I think that's handled by the rule. From their webpage:

> Companies must clearly explain subscription terms to you

Basing layoff decisions on citizenship is likely illegal, and that's why no company does that. It does not mean, however, that they can't do it by proxy. If you lay off people above a certain salary and justify it as cost cutting, yoiu are likely targeting mostly nationals. The few H1B caught in the net are just collateral damage. Then, you can hire imported workers that are willing to work for far less just to have a chance to come to America.

There are a thousand steps between models generating machine code and the future you envision, none of which is necessary, especially since there will be humans involved that will prefer alternatives.

> Cancelling visas or deporting them will not get American workers their jobs back.

When Microsoft lays off 6.000 American workers and then refills those positions with H1Bs, I would argue that cancelling visas will definitely bring those jobs back.

That might be true if you are the one exposing fraud and you are trying to reach the largest audience. That might also not work at all because when something is trying to appeal to everyone, it appeals to no one.

On the other hand, there is an equal amount of close-mindedness, and sometimes malice, from the side that demands "a better source" because they don't like where the information comes from, regardless of its veracity.

That doesn't really matter though, and it just makes the argument stronger.

This is a technology with an inherent tendency of making up false information AND we don't even understand how or why.

That's enough not to entrust these sytems with critical decisions that could start a war.

Isn't "deal with the difference" just renaming the file?

What AI reviews? I thought AI needed to be bound by Bend laws.

The AI either knows what the user wants perfectly and is able to prevent problems, or is a rougue agent that needs to be bound by laws to not misbehave. You can't have both.

> LAW: "the sum of all balances in this contract must be zero"

What prevents AI from writing an arbitrary calculation that makes sure that sum is always zero while breaking a whole other bunch of assumptions?

If the answer is more laws, then this does not address the parent comment.

Relatedly, a friend told me that graphical design gigs dried up because everyone can make event fliers and posters with AI with results that are often better than those of designers.

Except that the UDHR is not the rule of law.

It is a mere declaration, as specified at the top of the page, that "is widely recognized as having inspired [...] the adoption of more than seventy human rights treaties"

That does not make it an enforceable law. Moreover, treaties are agreements between countries, not laws, which a country can violate at will.

People might create supranational organizations like the UN or the ECHR, but that has, untimately, no enforcement power because that always belongs to sovereign countries. Violating a treaty can also lead to retaliation by other signatories or even war, but it still does not make a treaty law.

It might well be that the US has specific laws that prevent charging a citizen when they wipe their phone. But that must be a law passed by the US legislature, not a "universal declaration".

I agree with the conclusions of the article, and I am no fan of Amazon or search ads on any product.

However, I would like to play devil's advocate for a moment to brainstorm another issue.

Let's say I build a product that is better than others in the market. We all agree that competition is good becuase it leads to better products and lower prices.

However, the algorithm favors entrenched players. A product that has been bought many times and has positive reviews will be the first in the search results and people usually buy that. Which means I have no way to compete other than paying my way at the top of the list.

I'm not stating this as the ultimate solution or even a good one. But I think it's worth to ask the question.

What is the definition of major?

Globally, Firefox has ~2.2%.

On desktop alone that's still only 6.4%.

Where is the threshold?

While the calculation is not wrong, there are other factors at play in Meta's decisions.

First, how likely is it that similar rulings happen in other states or countries? While proportionally to the population the fine is significant, it isn't over Meta's global market. It can be only if similar fines will happen at a high enough rate.

Second, what is the global cost of adapting the whole platform to the ruling? (Assuming they don't build exceptions for users in New Mexico like Apple does)

Both factors have an influence, so if the fine is unlikely to be replicated, which it probably is, and the cost of adapting is exceedingly high, which certainly is, then Meta will continue their practices.

Wait, aren't they just licence plate readers? How do they help with all these other crimes?

Still, that defeats the purpose.

You analogy makes no sense.

"I complain about some people doing X openly while I ignore others also doing the same X behind the scenes"

is absolutely not equivalent to:

"Complaining that X sucks does not imply I am ignoring that some other unrelated Y also sucks".

I don't think the comment you are replying to is giving up leverage. It's simply pointing to the OP that what he seems to be upset is not the thing itself, which everyone does, but only that he knows about it or ignores others doing the same.

> "murdering" is universally considered as "bad"

Societies that practice(d) slavery and human sacrifice clearly do not share such a "universal" value.

Overall, intent is quite hard to prove. You would need to access internal documents and communications that explicitly direct the military. That's probably not going to happen.

However, from what I have observed, Israeli government officials have often expresses that intent publicly, despite claiming otherwise at the UN.

> Adult men have the highest casualties

And 30% of casualties are children. The two are not mutually exclusive and this is the data point currently under scrutiny.

> Hopefully it doesn't get captured by one clique or another.

Given that it's the most-voted post, it looks like it already happened.

There is nothing ironic in using the current status quo to spread something we hope will replace it. It's called pragmatism.

Most of the reactions I am seeing on various social media are negative.

Granted, that is not the market that would buy a Ferrari, but one of the point of buying luxury cars is the status they grant. There is not much status in a car mocked by the public.

> There are enough people who would pay any money to have an electric Ferrari.

Are there? That's a pretty bold claim.

I'm sure they think the same at Ferrari, but plenty of successful companies create products that flop miserably based on the wrong assumptions.

I would personally think that the public interested in high-performance combustion-engine luxury cars is not interested in generic-looking electric cars even if they come from the same company.

I think that the problem arises because people equate things that are not the same.

One person learning something is good. At scale, that becomes everyone learning something. That's even better.

Machine learning is not scaling up people learning. It's completely different even if it's called "learning".

As the article argues, it's plagiarism at scale. In that sense, one person plagiarizing content is bad. Everyone plagiarizing at scale by using LLMs is even worse.

I don't understand the endgame here. Websites let Google crawl their content in exchange of traffic. If Google cuts that out completely, what incentive do websites have to not block the Google crawlers?

I understand that Google is feeling an existential threat from other AI products that provide answers directly. But they must also understand their symbiotic relationship with the web.