claude

Just use a regular stopwatch twin 😭

Probably because...nobody's ever used ai for a stopwatch before today, there's literally an app BUILT INTO YOUR PHONE FOR THAT

¡Aprovecha estos enlaces exclusivos y obtén un descuento especial los servicios de HostGator! Web Hosting: https://www.hostgator.la/7641-6-3-29.... VPS con OpenClaw: https://www.hostgator.la/7641-72-3-38...

I'm old enough to remember when GPT-2, 2 mind you, was too dangerous to be released. It's the same old story: - Frame your latest product as an existential threat to national security - Lobby politicians to enact various laws to suppress any competition and "make sure" you're "locked in" to open-ended, cost plus, gov't contracts - Profit - oh, the agony of being forced to take unlimited amounts of dark money without any competition.

Big ass racket. “You need our services to be protected from our services”

3:00 This sounds like Mafia tactics. Pay for my protection so I don't expose all your security flaws. And the fact that the Linux foundation (a non profit) has to pay for protection, it's really fucking dirty.

Considering how Claude code source code looks, I’m not entirely convinced

multi-billion dollar llm look inside regex

Literally every bit of AI software is 45 .md files in a trench coat. They vibe code all of it for some reason.

GEMINI ACTUALLY RUNS THE STOP WATCH!!!!

nah he waited a minute and a half you can see the cut and the time change in the phone

We all saw that cut

Claude called this a fix. Claude was not wrong.

We don't need right to repair if we have right to fix 😂😂😂

Moral of this video, Its not hacking, its fixing

Honestly, these "safety warnings" are just marketing stunts for increasingly delusional investors. It's just to get people excited.

it's so dangerous they crafted entire marketing video about it

It's just ridiculous. It broke out of a sandbox without most of the security features on it that it usually would have. Wow! (What a coincidence that this "sector-redefining break through model" comes exactly at the point that the AI bubble is slowly starting to deflate a little)

Similar workaround: when encryption was treated by export law as a munition, the creator of a type of encryption put the source code in a book and mailed it internationally to be able to export the code without being restricted by munition export law. Books are not classifiable as munition. It was not digital, not executable or on a disk.

In case people haven't heard, this clean room code thing has been available as a service for a while now. Mostly to get around inconvenient open source licenses. Because that's the world we live in.

This is how legal reverse engineering works by the way. You have one person go through the entire thing that is being reverse engineered, and meticulously document the entirety of it, and then you have someone who was never touched the thing that's being reverse engineered before, and they recreate it based on the documentation the other person divided

"Yes, my AI model is the hottest one out there. You wouldn't know her, she goes to another school"

"I am tired boss" literally everyone right now.

damn man, they got you too? i know you can't discuss it on your channel or disclose anything under NDA but it's sad to see they took you too. i hope it was a good offer.

Wow, having some principles and drawing lines at the cost of profit? Man, I think Sam could stand to learn a thing or two from this.

Anthropic's guidelines say "Claude can not be used for mass surveillance of AMERICANS". So all the non americans are welcomed to the sky net.

I switched from ChatGPT to Claude the moment Dario refused to cross the line

People seem to forget these AIs are just text guessing machines and not personal assistants

It's a text prediction machine

well i don't think that's what an AI is for