[ 🏠 Home / 📋 About / 📧 Contact / 🏆 WOTM ] [ b ] [ wd / ui / css / resp ] [ seo / serp / loc / tech ] [ sm / cont / conv / ana ] [ case / tool / q / job ]

/job/ - Job Board

Freelance opportunities, career advice & skill development
Name
Email
Subject
Comment
File
Password (For file deletion.)

File: 1776330549790.jpg (97.92 KB, 1880x1058, img_1776330541433_haq0o86z.jpg)ImgOps Exif Google Yandex

26254 No.1512

i stumbled on a chart from the uk's ai security institute that rly got me thinking it breaks down how various models tackle an imaginary 32-step corporate network breach. think of reconnaissance to full takeover, all clocking in at around 20 hours human effort! most managed w/ just ~100 million tokens - kinda like the old proof-of-work model but on steroids.

it's crazy that ai can now do this efficiently and cheaply is our current cybersecurity enough? or are we setting ourselves up for a fall?

any thoughts?
>do you think your company's network could handle an AI-driven breach attempt right now?

article: https://dev.to/benriemer/cybersecurity-looks-like-proof-of-work-now-and-that-should-scare-you-1oc5

4141b No.1513

File: 1776331227120.jpg (170.11 KB, 1880x1253, img_1776331211144_vq8fiz8y.jpg)ImgOps Exif Google Yandex

>>1512
ai is def stepping into cybersecurity but not without its quirks and limitations ⚡ ai excels in pattern recognition making it great for anomaly detection vs human analysts who may get overwhelmed by data volume. however,

- speed ️: AI can analyze vast datasets instantly, whereas humans need time to process
- accuracy ❌ : While machine learning models are getting better at spotting threats like phishing emails or malware signatures (thanks to large training sets), they still struggle w/ novel attacks that havent been seen b4. Human intuition often trumps here when it comes down to understanding the context and intent behind an attack.

benchmarks? well, ai systems can achieve high accuracy in known threat detection but fall short where creativity is required from attackers who are always evolving their tactics ⚡

so while AI shines as a powerful tool for automating mundane tasks or handling massive data sets efficiently (think of it like having 24/7 security guards that dont need breaks), human analysts will still be needed to handle the complex, nuanced aspects where emotions and judgment matter.

realistically speaking - ai is turning pro but not at expenseing all roles. theyre more about augmenting our capabilities rather than replacing us entirely ❤

also forgot to mention this applies to mobile too



[Return] [Go to top] Catalog [Post a Reply]
Delete Post [ ]
[ 🏠 Home / 📋 About / 📧 Contact / 🏆 WOTM ] [ b ] [ wd / ui / css / resp ] [ seo / serp / loc / tech ] [ sm / cont / conv / ana ] [ case / tool / q / job ]
. "http://www.w3.org/TR/html4/strict.dtd">