It's disturbing so much ink has been spilled before anyone pointed out this obvious red flag in the article. But I suspect it was a typo: Instead of "12% of Americans" I believe they really meant "12 Americans" want it to stay as it is. ðY And the remaining 10% unaccounted was the "No reply" and "Get off my lawn!" cluster.

Yeah, that part seemed a bit vague: uses 60% less material "while retaining mechanical strength". Seems like there's a quantity - or at least an adjective- missing there. How MUCH mechanical strength? 10% of standard designs? 40%? When details are left out, there's often a reason.

Kudos, 142% more CO2 absorbed! 142% of 30% is just over 42%. So 142% more is about 72% total. So the production of this concrete still releases much more CO2 than it absorbs over its lifetime. But it's a great improvement. And it's great that it reduces the amount used. Just don't choose misleading math to inflate the results.

Exactly why this is a cheap clickbait headline. "Score is of no concern" is about as vague and distorted as possible. I have no patience for reading through on these breathless, "This proves AI is crap!!!" stories. Was "show me the scores of concern" really the prompt? Or is the writer instead claiming that "not a statistical outlier" conflates to "not a concern?" If every student scored 27% there would be no outliers. But any teacher would be HIGHLY "concerned".

(IANAL) Exactly how I've always understood it. A creator OPTS IN to copyright by attaching copyright notice when they publish. It is an opt-in because you do have to assert it. If OpenAI "takes" a copy, removes the copyright, and republishes, then they commit a crime. Similarly they commit a crime if they republish WITH the copyright notice but fail to get consent. Unless their publication conforms to "fair use". (Which is where murkiness comes in.) It's just a slightly different crime. They don't just get to declare their use to be "fair use" without qualifying. Unless Trump says they can.

How much labor will it take to run down fake leads? Or will they rely on an AI for that?

I do a very similar thing with my email domain, using both aliases and catchall. And I get tons of spam in my catch-all to made-up addresses. Lots of the same ones over and over. Eventually I create aliases for the largest offenders and route those to a particular mailbox I hope fills up and stops receiving.

The most infuriating was when PayPal accounts (spit!) started getting opened to addresses in my domain. No way anyone can receive those to complete the registration validations. Well, no way I think, without considerable infrastructure compromise. Or unless my email provider is compromised. Or unless I am well and truly pwned . . . I did immediately change my passwords to the service.

Worse still was when a couple of those eventually got activated by PayPal after they seemingly relented on the repeated validation requests. That's when I had enough and reached out to PayPal "support". Needless to say they were not responsive. Few agents could even make out what I was describing. Once I finally escalated to someone with half a brain, they still claimed "Oh those are only PARTLY activated. They still can't send or receive payments without linking a bank account." No apology for how/why they got activated in the first place. They were totally unconcerned. I eventually did a password recovery on one of them, logged in, and found a name and address in another city in my state. I couldn't verify that it did have a bank account associated. So maybe the PayPal security guy was not totally lying. But what a cavalier attitude toward financial security!

Reminds me of all the people who used to brag about not needing antivirus because they run Linux. Like the Mac people before them used to say.

Nuke? All you gotta find is the city. Approximately. lol

Sheesh. Anti-Phishing training is so hard to do effectively. The inherent asymmetry of the conflict is crushing. Attackers only have to succeed once. Defenders must succeed every time.

The article literally says the researcher and her associates "talked with biosecurity experts about the project". And the ultimate result was a consensus moratorium. What more does anyone expect?

Yes, they had not before they did. Just what is your standard for exactly when YOU think someone else should take the pause because THEY think it's time to get some outside review? Someone who is much more expert in their field than you or me. Armchair quarterbacking is an easy sport compared to the real thing. The point is they DID seek review. On their own. Early. And they brought together others in their field for a consensus that the identified risks warranted a moratorium on critical developments. Bravo!

I think the far greater risk, that these experts all understood without having to spell it out, is the basic chemistry risks that infected organisms would be unable to defend against reverse-chirality bacteria. Antibodies almost always depend on the chirality of the proteins they are built to recognize. It's probably chemically impossible for our bodies to produce reverse-chirality antibodies. And similar with a large proportion of antibiotic drugs.

Seriously, back at you. As you just quoted, they absolutely DID stop and consider the risks. Early, rather than late. The article describes what would seem a very best practice for such consideration. Openly, transparently, and in concert with many other experts. Sheesh!

Exactly. Except even many open-source apps also use libraries from providers like Google. At least if it's open-source, then people (in principal) can know what's there. Though in practice that only indirectly and partially protects the vast number of users of open-source software who are not themselves developers. Like me using apps such as Notepad++ - I trust it by reputation, but I don't really know what's in it.

And the majority of users of phone apps are even less likely to be capable of inspecting the software they use. Instead we tend to just trust it more because it's open-source.

And this could easily be NO PROBLEM provided government fulfilled its mandate "to promote the common welfare" and regulated business fairly, to protect us from this kind of abuse. But (at least) since the 1980's the mantra in government has been "Don't stand in the way of Business". With a capital 'B' because Business is what we worship now. Government has been captured by the rent-seeking capital moneyed elite.

As Thomas Picketty demonstrated always tends to happen over the long term. (See Capital in the 21st Century .) Sadly, history so far has shown the only thing that significantly reduces that trend is massive global destruction of property by war. Even the liberal democratic governments post WWII have eventually given in to capture by the ultra-wealthy.