yellow_lead

> There might be a huge potential for an AI which scrapes your (and my!) data and concludes what I can safely delete, saving the world tons of resources.

Thinking...

User has 15 pictures of his dead mother, deleting 14.

Saved 20 MB! :)

Ah ok, I misunderstood what NAT-PMP was. Thanks for explaining.

I was kind of assuming the engineer didn't have full access. But maybe the other factor(s) were additional exploits

> I'm guessing here, but my money would be on freezer ICS systems being made available to the world by accident by way of NAT-PMP on a consumer router.

Why would there be accidental port mappings on their router though? I could maybe see owners setting it up so they could monitor from home, but even that is a bit of a stretch for rural supermarket owners

For those who live alone, how many of you watch YouTube while eating at home?

If you watch or do something else while eating, what is it?

I found the "new to you" tab on YouTube, which worked well once or twice. After those initial uses though, I couldn't get good results again

Do you have a source? Because I've only been hearing about the laptop.

> Binance has said the incident was a “complex attack” that stemmed from multiple factors.

> Its initial investigation suggests that the security incident may involve a compromised DevOps engineer’s laptop, which allowed an attacker to gain access to a service container and submit unauthorised withdrawal transactions.

I know it says "multiple factors" here, but there are no other factors I've found.

> OpenAI confirmed both frontends rejected the reported messages, resolved the issue, and awarded a total bounty of $5,000.

It requires the user to click a link. But I still find this bounty very low. Others may not report them if such a small bounty is given.

Azure? Really?

> The attack is remote, needs only the victim's number, and shows them nothing.

err, how can I send an SMS like this? Would a mobile network even accept/forward such an SMS?

Little hard reading through the vibe slop here

Crazy to hear. So you look in user's sandboxes, does this comply with GDPR and other regulations?

Remember "choose boring technology"?

I had to evaluate Deno vs. Bun vs. Node for a project two years ago, and chose Node. I think Deno has good ideas but often times these projects cannot reach escape velocity, especially when they're constrained by VC motives.

Reselling APIs seems lazy to me. I wonder if CF plans to make their own provider. That's what I had assumed when I read the title.

> I can think of no stores I'd rather be in, if there's a fire, than that Apple Store.

How about a fire extinguisher store?

Funny how the showcased "Tidal house" has unreadable text on an image.

https://tidal-house-retreat.openai.chatgpt.site/

> Quiet cabins. A view of your own.

Yep, they're spy glasses / perv glasses.

A site I work on serves dynamic language content based on the accept-language header. We do this for some pages and others we redirect to i.e /en/page

But we can't cache those dynamic pages since they vary based on accept language. Maybe with this we can

I might try it for work if it weren't made by meta.

> account payout is frozen

Frozen payouts are different than money being held for longer.

> Singapore is a controlled economy

Singapore is a mixed economy [1].

[1] https://en.wikipedia.org/wiki/Economy_of_Singapore

How dare employees do something without asking for the full backing of the company's resources. Incredibly unethical!

Both can be true:

1. OpenAI couldn't have solved the problem without the researchers' private data for training.

2. OpenAI models can solve math problems

> But this guy? He's timed his exit, waiting for the IPO

What exactly do you mean by this

If the model can understand neuralese why can it not convert it into English for monitoring or review purposes?

Or we believe the model will encode secret messages like "don't reveal this information" into the neuralese. But as the author mentions, they could have been doing that all along

> Models can omit key information in their visible thoughts, as this Anthropic 2025 paper shows. We are also worried about steganography

I don't know but I'm basing my assumptions on this article: Teenager’s Gemini mistake locks entire family out of Google accounts [1].

1. https://www.pcworld.com/article/3104521/teenagers-gemini-mis...

You may also be banned by asking "sensitive" or "inappropriate" questions. So normal usage of Google AI products could also be risky.

I don't care if my Claude or OpenAI account is banned, but risking a Google account is different.

> Two mathematicians at Anthropic studied and validated Claude’s paper, and produced an informal note for experts stating Claude’s proof concisely.

No, experts are still needed

High Anubis difficulty is annoying the hell out of me for several sites. And it's starting to not block LLM bots anymore?

> 33% are now solving the math and getting through to the main site — because apparently what we have to offer is worth spending a ton of cycles to calculate the Anubis challenge.

Is there anything new in this article? Yes, experts use AI better than non-experts for tasks in their domain. See LLMs reward expertise [1] and Terrance Taos conversation with LLM [2].

[1] https://www.seangoedecke.com/llms-reward-expertise/

[2] https://chatgpt.com/share/6a5fdc7a-d6f8-83e8-bbea-8deb42cfed...

Misplaced modifier? Wow