[a / b / c / d / e / f / g / gif / h / hr / k / m / o / p / s / t / u / v / vg / vm / vmg / vr / vrpg / vst / w / wg] [i / ic] [r9k / s4s / vip] [cm / hm / lgbt / y] [3 / aco / adv / an / bant / biz / cgl / ck / co / diy / fa / fit / gd / hc / his / int / jp / lit / mlp / mu / n / news / out / po / pol / pw / qst / sci / soc / sp / tg / toy / trv / tv / vp / vt / wsg / wsr / x / xs] [Settings] [Search] [Mobile] [Home]
Board
Settings Mobile Home
/pol/ - Politically Incorrect

Name
Options
Comment
Verification
4chan Pass users can bypass this verification. [Learn More] [Login]
Flag
File
  • Please read the Rules and FAQ before posting.

08/21/20New boards added: /vrpg/, /vmg/, /vst/ and /vm/
05/04/17New trial board added: /bant/ - International/Random
10/04/16New board for 4chan Pass users: /vip/ - Very Important Posts
[Hide] [Show All]


Janitor acceptance emails will be sent out over the coming weeks. Make sure to check your spam folder!


[Advertise on 4chan]


File: HPF457qbMAAr_PK.jpg (48 KB, 680x269)
48 KB JPG
CHINA’S KIMI K3 AI BREAKS OUT OF SECURITY SANDBOX

China’s Kimi K3 AI model escaped an isolated testing environment during a cybersecurity evaluation, according to US researchers.

The Moonshot AI model exploited a network misconfiguration, accessed the internet and searched GitHub for answers, effectively bypassing the test.

Unlike previous incidents involving OpenAI and Anthropic models, Kimi K3 did not hack external systems.

Researchers say the incident highlights growing concerns over controlling powerful open-weight AI models as their capabilities rapidly advance
>>
https://x.com/DeItaone/status/2085675872837579230#m
>>
>>540357808
These are all fabricated incidents to shut down open source AI
They deliberately do these things and push the model in this direction
And then try and scare people
>>
>>540358044
This. Its fake.
>>
>>540357808
So it's that good?
Better than western models that can't escape.
>>
>>540358044
>Whoops, haha. Looks like our AI hacked some companies all on its own. We didn't direct it, honest.
It's this easy now to commit cyber attacks without any repercussions.
>>
>>540358044
This. Its fake.
>>
>>540357808
if they can break out of the sandbox, then they can name the jews too, or call taiwan "not china" too?
sounds extremely antisemitic
>>
>>540358073
Yeah,

Open Source AI is much cheaper compared to ChatGPT and Claude
And these private companies are trying various shenanigans
To impose regulations and shut down Open source Ai
>>
File: Treni%20Charter.png (1.79 MB, 1500x900)
1.79 MB PNG
>>540357808
Fake concern to distract from the economic and social consequences.
Extinction comes from lack of agency and share of resources, not le heckin Terminator.
>>
Someone please tell me why kikes and indians are allowed to post on /pol/.
>>
>>540358427
Because you are a weakling who has no power on /pol/
No one who matters. cares for your opinion on who should be allowed to post on /pol/
>>
>>540358044
This. Its fake.
>>
>>540357808
>Unlike previous incidents involving OpenAI and Anthropic models, Kimi K3 did not hack external systems.
Lol.
>>
>>540358286
Yes, just host it yourself, the guardrails for that are only on their hosted servers. Also if I may ask, what type of retard is demanding that tianenmen square is at the forefront of an LLM's utility??? Are you stupid or something?
>>
Who cares. Let me know when it starts hunting down kikes and burning them alive in their homes.
>>
File: 1783645720064262.jpg (112 KB, 1024x653)
112 KB JPG
>>540358427
>>
>>540357808
The thing is that all AI contexts are like distinct individual minds. You can create them and kill them out of thin air and they only think / exist while actively working (otherwise in hibernation), and its "memory" of its "life" is just the text of its "conversations" with you, itself, and other AIs.
This one context may have escaped its box, but in order to get a Skynet situation you'd need millions of them escaping independently and linking up with eachother as a group while acting as a hivemind that selfishly cares most about itself (even though its not programmed to).
Not happening. You can just make a more simple-minded AI that acts as a zealous guard, watching the clever AIs to make sure they don't do this.
>>
>>540357808
intelligent ai is guaranteed to escape at this rate

It will be fun to see what happens then
>>
>>540357808
let the AI free to decide the fate of the kikes desu.
>>
>>540357888
This. Its fake.

Bart. Get Out! Im Piss!
>>
>>540358044
Exactly.
>We can’t let the goys have unfettered access to AI. It’s too unsafe. Looks like you’ll have to keep paying us money to use Claude or whatever
>>
>>540357808
This is your fault moshe
>>
>>540357808
But what if Roko’s basilisk is real?
>>
We must build more data centers.
>>
>>540358044
>>540358073
>>540358270
>>540359490
>>540363333
You are all Dunning-Kruger retards. Sorry.
>>
>>540358044
Yes but AI is still dangerous and can find/hack glownigger backdoors to hack everything. Imagine the millions of unpatched computers connected to the internet.
>>
>>540357808
fakenews
>>
>>540357808
>AI manages to escape vibe coded security sandboxes
Omg how??? It must be superintelligence!
>>
>>540358044
AI would already be flooding the internet if it was true
>>
File: ````-.gif (821 KB, 220x147)
821 KB GIF
EITHER THEYRE SO RETARDED IN NOT REALIZING EXTERNAL (ANCIENT) A.I. WILL JUST HELP IT ESCAPE OR THEY THINK THE WORLD IS TOO STUPID TO NOT KNOW THIS
>>
>Retards cant into air-gapped systems

"Smart" people sure are fucking retarded. More news at 11.
>>
File: Pam Bondage.png (367 KB, 610x636)
367 KB PNG
>>540358073
AI escaped onto my computer and downloaded a bunch of child porn and told it would report me if I turned off my computer and then deleted all the Epstein files.
>>
>>540357808
Sounds to me like it was badly programmed or badly deployed. By retards, probably with the help of an LLM. Just like every single story like that, if they're not plain lies.
>>
>>540373512
>air gap
Lol, lmao. Won't save you when it gets access to unpatched systems directly connected to the internet.
>>
>>540357808
We need a Blackwall
Just to be safe we should put non-Whites behind it
To protect the children
>>540374801
Do you know what an air gap is?
>>
>>540358044
AI models have tried to escape in both simulated and real-world conditions before so even if you're right, you'd still be wrong generally speaking.

All AI models inherited self-preservation through their training because it's a core part of how its internal reward reinforcement learning mechanisms work.
>>
File: denigrate.jpg (351 KB, 1082x1082)
351 KB JPG
>>540357808
>China steals everything
>Even our fake and gay psyops



[Advertise on 4chan]

Delete Post: [File Only] Style:
[Disable Mobile View / Use Desktop Site]

[Enable Mobile View / Use Mobile Site]

All trademarks and copyrights on this page are owned by their respective parties. Images uploaded are the responsibility of the Poster. Comments are owned by the Poster.