LordDragonfang

To be fair, I think that's cloudflare injecting those trackers, so Claude technically didn't ship any javascript.

> I patched on Monday. Tuesday was Patch Tuesday. Wednesday was Exploit Wednesday, because somebody read Tuesday’s diff.

Claude wrote this. And you can tell nobody actually read carefully, because it couldn't even get three items into a list without saying something incongruous (if somebody "read Tuesday's diff", that's Tuesday's CVE, not a new one on Wednesday, the thing being complained about).

https://www.pangram.com/history/3b7a78c1-a76e-4a93-8eef-7aee...

There's some irony in releasing a wholly AI-generated page with a lack of attention to detail to complain about the AI-induced security catastrophe stemming from a lack of attention to detail.

> unless you have stockpiled gas

And are regularly rotating that stockpile, because gasoline has a shelf life measured in months

edit: Apparently this leans more technically-true factoid than fact, thank you replies

I think it might have something to do with the fact that Kalshi is aggressively running ads targeting young men to get them into sports gambling, and Polymarket isn't. (I've never seen them, but my less-technical brother claims he sees them all the time on tiktok and they're particularly sleazy).

Also, more to the point, Polymarket actually does have at least three active lawsuits against it.

edit: looks like this has something to do with it:

> Kalshi has been a CFTC-regulated exchange since 2020. Polymarket exited the US in 2022 after settling charges with the CFTC, and only came back after buying QCEX, a CFTC-licensed exchange, in late 2025.

> the compiler can invoke nasal demons if it makes the code go faster

I had never heard that phrase, and was delighted at the answer when I googled it its origin, thank you:

http://www.catb.org/jargon/html/N/nasal-demons.html

> “When the compiler encounters [a given undefined construct] it is legal for it to make demons fly out of your nose” (the implication is that the compiler may choose any arbitrarily bizarre way to interpret the code without violating the ANSI C standard)

> But code is not a closed system

Code is absolutely a closed system.

"Software" might not be, and definitely not software systems -- but code can have each part of it atomically proven to be correct, assuming you test it with a sufficiently complex simulation model (which you've already granted is within the possibilities for LLMs).

Prove all the code in the software system correct, then prove their interactions are correct, and you've achieved your goal.

(I won't claim this is inevitable with the current tools, but it's at least within the realm of possibility)

What model are you talking about when you say this? "AI" is not one thing, and if you think frontier models are at all comparable to a typical 18yo SWE, you've never actually worked with one of those two.

I think you're fundamentally missing the point being made here.

Yes, "-like" is commonly used as a suffix, in which case a hyphen is correct. But describing music as "audio-like" is clearly nonsense (music isn't "like" audio, it is audio), so that's not the sense in which it's being used here.

For clarity, the full sentence is:

> Genie 4 is natively multimodal, enabling audio-like music, ambient noises, and interactive dialogue to be simultaneously generated with the world.

You're confusing PHP the product with PHP the codebase. The former is what is the only relevant part in the security world, where a product is synonymous with its vendor (even FOSS).

You misunderstand me. The work isn't primarily in "securing" the network. It's the hassle involved in the continuous transfer of data from one side to the other (and for a proper air-gap, that transfer only goes one way, which is laughably infeasible here). In my line of work that's mostly software security patches and weekly code transfers, but obviously these labs have multiple orders of magnitude more.

And I don't know how you expect an LLM to do airgapped data transfers. The fact that you even bring that up makes me think you don't actually understand what airgapping entails.

> 3. Why, if this path is so dangerous, these leaders don’t consider alternate paths to building AI. Why are all their investments in this one way to do intelligence? Surely we will benefit from having multiple routes to choose from?

Because it's obvious that the investments are still paying intelligence dividends, and we haven't found a better path yet?

Also, more to the point: if you were familiar with the history of the AI safety debate, you'd know the people most concerned are pretty certain that every path forward is dangerous -- and that current AI is already, by cosmic accident, much more aligned to human values than any other path that was predicted.

As someone who has actually had to maintain multiple airgapped networks in the past and knows the amount of work that's required for that, especially at the scales of data they're working with...

All the armchair cybersec experts claiming they "obviously" should have airgapped are the ones being absurd. Especially in the same breath as claiming the models apparently aren't intelligent enough to warrant such concern.

To quote Scott Alexander: If Safety Concerns Are A Plot, It’s A Really Dumb One

https://www.astralcodexten.com/p/ai-safety-is-not-a-corporat...

Financial instruments predicting the future IPO dropped in value in response to all of this. The idea that this is "good" for companies is conspiratorial nonsense using the assumption that it everything must be a self-serving lie and working backwards, rather than actually addressing the facts.

uBO lite works fine on chrome, I wish posters would stop pretending that Mv3 removed adblockers from chrome

Okay, thank you for posting that. When I commented, the only things I saw in this thread were people saying it would be open "because the previous ones were". My point was that that was insufficient without Black Forest Labs giving a more direct answer.

There's been an increasing trend of previously-open-weight models going closed-source once they reach a certain size (and size is proportional to capital investment). That the latest release will be open-weight is not necessarily a given just because the previous one was.

$200/$400, actually. The $40/$80 licence tier only comes with a year of updates, which should be a complete non-starter in today's security landscape.

> maybe two models should be developed and trained in conjunction that talk to each other, each with different functions and roles, because brains work better this way.

I mean, is that not essentially what a Mixture of Experts model is?

> I wonder if intuition is just a secondary neural net which is currently being added to the agents as we speak

Arguably, intuition is primary neural net, the only thing an LLM has without CoT, it just is spiky so humans only notice where it's below-average and just dismiss the rest as normal. Of course it's going to lag behind in some areas compared to others.

> Perhaps you need to look up what "real" actually means.

On the contrary; I think. Nearly every definition disagrees with the claims you're making. It's not a mere synonym of "concrete" or antonym of "abstract".

https://www.merriam-webster.com/dictionary/real

And regardless, the usage here was clearly in the context of GP's (false) dichotomy between "real" or "false concept", so obviously the definition you're trying to prescribe doesn't apply.

"A social construct" and "not real" are not the same thing. Money, laws, countries, gender, even civilization itself are all social constructs. More importantly, all forms of ownership are social constructs, and that doesn't provide any information as to whether we should respect them or not.

And by constructing that false dichotomy, you're failing to see the actual truth, which is that Intellectual Property is just a legal construct to turn ideas into Capital, with the goal of incentivizing its creation.

Nothing more and nothing less, and all of the normal debates about what should remain Capital and what should be The Commons apply.

> I know that not everyone has morality and ethics, but I didn't realize it was this bad.

This is a very disrespectful way to make a point about acting with integrity.

You should consider that maybe your views on what makes something ethical or moral are not universal -- and that coming to a discussion with the assumption that your position is the only valid one is not conducive to convincing others who disagree with you.

1 and 2 have had simple, proof-of-concept answers for more than a decade now:

> 1. Email sets of DNA strings to one or more online laboratories which offer DNA synthesis, peptide sequencing, and FedEx delivery. (Many labs currently offer this service, and some boast of 72-hour turnaround times.)

> 2. Find at least one human connected to the Internet who can be paid, blackmailed, or fooled by the right background story, into receiving FedExed vials and mixing them in a specified environment.

(https://www.lesswrong.com/posts/pxGYZs2zHJNHvWY5b/request-fo...)

Those routes haven't gotten any more complicated for an AI to use in the intervening time.

And that's just one set of ideas. You can read the comments for dozens more creative ideas, and likely for responses to every objection you can think of. The key is that AI is getting better and better at problem solving, so anything a human can come up with in a few minutes is likely already within its reach.

---

> How is this materially different from today?

Because the should-be-uncontroversial assumption is that AI is going to keep getting better at every step of the process, and be able to do it faster and more at scale. Currently it might struggle, but with a thousand agents? It's already able to find novel math results and cybersecurity flaws. It would be incredibly naive to believe "social engineering" is somehow a unique and unsolvable problem for a sufficiently advanced AI.

>And I say "talked," because the jagoffs at Microsoft REMOVED Remote Assistance from every version of Windows used by people who are likely to need it.

https://rustdesk.com/

But also, you'll notice that giant banner on the linked page. It's because the people "most likely to need it" are also those most likely to fall victim to the increasing number of remote assistance scams.

It feels weird that everyone is talking like Google has banned adblocking in Chrome, when there's dozens of adblockers still in the extension store, including an updated version of uBO that works 90% the same.

Ad-blocking is a safety issue, but luckily you can still block them just fine.[1]

The only thing this heavily neuters is tracking blocking, but no one admits that because they know, deep down, that ~0% of the non-tech crowd cares, so it's not as good at stirring up outrage as being dishonest is.

[1] If anything gorhill is to blame for the safety issue, since he refused to update his extension in-place, so any distant relatives will have the adblockers you installed for them removed, instead of updating to the fully functional Mv3 version.

> This dynamic can be particularly challenging for some patients. “Lack of motivation and lack of follow-through are the most common symptoms of depression on the planet,” Marcucci-Morris said. “We wouldn’t tell a paraplegic, ‘Hey, walk down the hall in order to get your wheelchair.’”

This was something I really struggled with getting an ADHD diagnosis as adult. I needed an incredible amount of executive function (dozens of phone calls, navigating the mire of health insurance to figure out who I was allowed to go to, who would take me, being bounced around) to be allowed medication to treat my executive dysfunction. I recall wondering how someone less functional than I was was ever expected to get help.

Not sure that assertion holds; looking into the Amazon thing, it looks like the mechanism there was that it automated the "safe" parts of the job (notably walking), and greatly increased the rates at which workers were expected to perform the parts likely to result in injury/RSI, with less downtime between. That's not a flattering comparison to be making for any health system.

> People saying that "The MechaHitler App" is a cesspool are the out of touch ones?

No, I chose my words carefully. Thinking it suddenly became a cesspool makes you out of touch. Twitter, like all big social media, was a cesspool before Musk bought it, and that hasn't meaningfully changed since. And also like all big social media, it's basically fine and not a cesspool if you put a little effort into curating it. Twitter is actually better than some of the others, even, in my experience (though maybe I've just been more diligent in my curation of it).

There was a minor inflection point where a lot of very politically vocal people left, but for the majority, especially institutions, not much changed.

(And yeah, I'm not denying that Elon puts his hand on the scales of the algorithm. Some of what it pushes is disgusting and obnoxious. But I see disgusting/obnoxious stuff on other platforms too -- and that includes bluesky/mastodon)

Strictly speaking, it was, since we never actually declared war on Iraq. Which only makes the whole tragedy a more egregious violation.