an AI cannot exploit vulnerabilities in other sites if it cannot access the internet. Its entirely on you, when designing your toolset that will host the AI, to decide whether or not it needs internet access at all, and to put controls in place that will limit what sites it can even hit.
And right here is where you unwittingly revealed that you didn't read the fucking article, because if you had read the fucking article, you would've known that the toolset was sandboxxed from the internet, but the AI models "identified and exploited a zero-day vulnerability" to break out of the sandbox and "performed a series of privilege escalation and lateral movement actions in our research testing environment until the models reached a node with Internet access."
Our team had a pretty unique idea that we all were excited about, but after the pitch our Apple contact stopped answering messages about the promo. Sure enough, 3 months later Apple took the exact idea and used it in a promo for one of their own products on the App Store.
geez man, don't leave us hanging -- what was the idea?
he scraping abuse on reddit used to be quite terrible. Every subreddit I enjoy (hobbies + porn) was well moderated, so not a ton of bot bullshit on the actual site, but I saw a lot of my old posts content farmed on shitty websites.
[...]
I was really upset when I found there were sites serving my deleted posts in an easy to search format.
Most people wouldn't share ANY of the facts that (1) they actively make creepy posts on porn subreddits, or that they 2) got "really upset" when they found out that their creepy posts on porn subreddits were archived elsewhere, OR that they 3) consider copying their creepy posts on porn reddits as "stealing from me"
So when you found this out, did you demand to talk to the manager of the archving websites?
In all this vibe coding/agentic development/loop coding, HOW DO YOU KNOW the AI is doing something fundamentally wrong?
Presumably, the same way you'd figure out if a junior dev you gave a bunch of shit to do was doing something fundamentally wrong: by checking its work?
Even the summary is confused. Apparently a copy on your hard drive is not physical, and therefore can't be preserved or whatever. The copy of GTA VI you might download from Rockstar is also not physical no matter what you put it on? So a physical copy is what, printed on paper? Chiselled in stone?
So you found one braincell, but try as you might, couldn't find a second one to rub against it and figure out that maybe it's referring to the physical medium that people have been buying video games for over 4 decades -- on cartridges or optical discs?
And you really can't figure out the advantages of tangible copies:
- can be lent to friends
- isn't permanently tied to a specific account
- isn't dependent on a vendor's storefront needing to remain operational
- price competition between retailers vs fixed digital storefront pricing
- can be sold when you're done
- can be purchased used at a reduced price
I'm not surprised Rockstar decided not to provide their six terabyte game in those formats.
wtf are you talking about
"If we're building agents to help people complete everyday tasks using computers, our models need real examples of how people actually use them -- things like mouse movements, clicking buttons, and navigating dropdown menus," said a statement Meta issued earlier.
Early reports on the effectiveness of the training has shown mixed results -- the agents are REALLY good at mouse movements, clicking buttons, and navigating dropdown menus, but no matter the prompt provided to it, the agent just opens firefox and starts browsing job postings on Indeed.
So in this instance, you failed, because no one knows who the shit Rebecca Watson is, nor do they give a shit about whoever the fuck Rebecca Watson is, thinks
The rule on staying alive as a program manager is to give 'em a number or give 'em a date, but never give 'em both at once.
