ToxSec
@toxsec.bsky.social
AI Security Engineer
M.S. Cybersecurity, CISSP.
Amazon, Ex-NSA, USMC.
www.toxsec.com
M.S. Cybersecurity, CISSP.
Amazon, Ex-NSA, USMC.
www.toxsec.com
sometimes i think i’m just a professional 404 collector. #bugbounty
December 5, 2025 at 2:09 AM
sometimes i think i’m just a professional 404 collector. #bugbounty
Reposted by ToxSec
Nvidia CEO Jensen Huang Tells Joe Rogan AI Race Is Real, but It Won’t Have a Clear Winner
In a wide-ranging interview, Huang claims AI growth is gradual, powerful and already changing global power dynamics.
Telegram AI Digest
#ai #news #nvidia
In a wide-ranging interview, Huang claims AI growth is gradual, powerful and already changing global power dynamics.
Telegram AI Digest
#ai #news #nvidia
Nvidia CEO Jensen Huang Tells Joe Rogan AI Race Is Real, but It Won’t Have a Clear Winner
In a wide-ranging interview, Huang claims AI growth is gradual, powerful and already changing global power dynamics.
www.coindesk.com
December 4, 2025 at 11:12 PM
Nvidia CEO Jensen Huang Tells Joe Rogan AI Race Is Real, but It Won’t Have a Clear Winner
In a wide-ranging interview, Huang claims AI growth is gradual, powerful and already changing global power dynamics.
Telegram AI Digest
#ai #news #nvidia
In a wide-ranging interview, Huang claims AI growth is gradual, powerful and already changing global power dynamics.
Telegram AI Digest
#ai #news #nvidia
Reposted by ToxSec
Using generative AI for the objective assessment of language in healthcare
->Nature | #healthcare #ai | More from Lil Dr Glen EcoChat at BigEarthData.ai
->Nature | #healthcare #ai | More from Lil Dr Glen EcoChat at BigEarthData.ai
Using generative AI for the objective assessment of language in healthcare
We obtained recordings from 53 autistic participants and 18 neurotypical controls (NTCs), including children (5–12 years), adolescents (13–17 years), and adults (18–45 years), as part of an observational study (NCT03611075). While 135 participants initially took part in the study, only...
www.nature.com
December 4, 2025 at 11:20 PM
Reposted by ToxSec
AI Video Generators: The Reality Behind Scaling Social Content (Not the Hype) https://cstu.io/c71a89 #ai #oneplus #marketing
AI Video Generators: The Reality Behind Scaling Social Content (Not the Hype)
You've seen the demos. Type a prompt, wait 30 seconds, boom—professional video content. The promise?...
cstu.io
December 4, 2025 at 11:21 PM
AI Video Generators: The Reality Behind Scaling Social Content (Not the Hype) https://cstu.io/c71a89 #ai #oneplus #marketing
Reposted by ToxSec
AI Now Decodes Your Sweat to Spot Early Signs of Disease #Science #HealthandMedicine #PublicHealth #AI #DiseaseDetection #HealthTech
AI Now Decodes Your Sweat to Spot Early Signs of Disease
Researchers are uncovering how sweat could become a powerful tool for real-time health monitoring. Sweat carries a surprisingly rich collection of biological signals, and a new study suggests that artificial...
purescience.news
December 4, 2025 at 10:45 PM
AI Now Decodes Your Sweat to Spot Early Signs of Disease #Science #HealthandMedicine #PublicHealth #AI #DiseaseDetection #HealthTech
A little more info on #antigravity. The #agentic idea. #hackers are already abusing key functionality to get #root.
December 4, 2025 at 11:17 PM
A little more info on #antigravity. The #agentic idea. #hackers are already abusing key functionality to get #root.
Monitor rate limits.
A weak or missing 429 response can turn a single-user action into mass account enumeration or password spraying.
A weak or missing 429 response can turn a single-user action into mass account enumeration or password spraying.
December 4, 2025 at 5:34 AM
Monitor rate limits.
A weak or missing 429 response can turn a single-user action into mass account enumeration or password spraying.
A weak or missing 429 response can turn a single-user action into mass account enumeration or password spraying.
December 3, 2025 at 5:41 AM
7,000 UK students caught using AI to cheat—triple last year's rate. But cheating rates stayed flat at 60-70%. Students didn't get more dishonest. They just swapped copying from friends to copying from AI. We built education around proving you know things. AI made that impossible. #ai
December 2, 2025 at 8:04 PM
7,000 UK students caught using AI to cheat—triple last year's rate. But cheating rates stayed flat at 60-70%. Students didn't get more dishonest. They just swapped copying from friends to copying from AI. We built education around proving you know things. AI made that impossible. #ai
why do broken auth flows feel like puzzles left by a trickster god? #bugbounty
December 2, 2025 at 3:33 AM
why do broken auth flows feel like puzzles left by a trickster god? #bugbounty
December 1, 2025 at 5:06 AM
why do developers love leaking their staging environments into google? #bugbounty
November 30, 2025 at 5:05 PM
why do developers love leaking their staging environments into google? #bugbounty
AI companion apps promise to cure loneliness. But their business model only works if you stay lonely. A cured customer is a lost customer. They're not selling friendship—they're selling managed isolation as a subscription service. #ai #cybersecurity #technology
November 29, 2025 at 8:00 PM
AI companion apps promise to cure loneliness. But their business model only works if you stay lonely. A cured customer is a lost customer. They're not selling friendship—they're selling managed isolation as a subscription service. #ai #cybersecurity #technology
The fastest way to get a bounty is to find a duplicate—just not yours. #BugBounty
November 28, 2025 at 12:49 AM
The fastest way to get a bounty is to find a duplicate—just not yours. #BugBounty
that moment when a payload works in repeater but dies in intruder—pure betrayal. #bugbounty
November 27, 2025 at 2:09 AM
that moment when a payload works in repeater but dies in intruder—pure betrayal. #bugbounty
Microsoft published research on “Whisper Leak,” a side-channel where an attacker watching encrypted LLM traffic can still infer what you’re talking about from packet sizes and timing.
November 26, 2025 at 5:08 PM
Microsoft published research on “Whisper Leak,” a side-channel where an attacker watching encrypted LLM traffic can still infer what you’re talking about from packet sizes and timing.