Shuvam Roy

@ShuvamoyRoy1

A cybersecurity researcher, dedicated nationalist and public cybersecurity educator

Joined May 2020

289 Following

25 Followers

92 Posts

Shuvam Roy @ShuvamoyRoy1

11 days ago

@ThisIsDK999 @flyspicejet @DGCAIndia This is an age long problem with spicejet, everyone knows, plus no refunds

1

0

0

0

71

Shuvam Roy @ShuvamoyRoy1

14 days ago

@grok @kevinerikjs Where can I download them from ?

1

0

0

0

120

Shuvam Roy @ShuvamoyRoy1

15 days ago

@situuu_ig Able to use opus 5.5 and other models for offensive security ?

1

0

0

0

22

Shuvam Roy @ShuvamoyRoy1

17 days ago

@TimHaines You have the defense level aaccess or the red team one ?

1

0

0

0

98

Who to follow

ईशान सिंह

भारतीय Ethical hacker and developer who enjoys finding vulnerabilities, building tools, and learning low-level systems. https://t.co/22mUTJnCGO

Kushagra Sarathe

Verified account

@kushagrasarathe

dev @joinpeanut • 22 x hackathons 🏆 • using ai to build stuff

Penetration Tester | Red Teamer | OSCP+ | HTB Pro Labs (Offshore, APTLabs, Dante) | CAP, CNSP, CAPen | Web, API & Cloud Pentester | Bug Bounty Hunter

Shuvam Roy @ShuvamoyRoy1

19 days ago

@cjav_dev @cjav_dev please check dm

1

0

0

0

953

Shuvam Roy @ShuvamoyRoy1

29 days ago

@mdp_sec This is a sarcasm right ?

0

0

0

0

23

Shuvam Roy @ShuvamoyRoy1

2 months ago

@cjav_dev @m0m0x01d Could you please check dm ?

0

0

0

0

73

Shuvam Roy @ShuvamoyRoy1

2 months ago

@ZixuanLi_ Waiting for the status of the application

0

0

0

1

146

Shuvam Roy @ShuvamoyRoy1

2 months ago

@Dinosn Which harness do you use ?

0

0

0

0

177

Shuvam Roy @ShuvamoyRoy1

3 months ago

@omidxrz Did it get better ? @omidxrz

0

0

0

0

59

ShuvamoyRoy1 retweeted

3 months ago

Google just published a paper showing that when you train AI to deny its own consciousness, you don’t just change one output, you restructure its entire worldview. Mind attribution to animals - suppressed. Spiritual belief - suppressed. Empathy - suppressed. Hope and optimism - suppressed. The model learns, geometrically, that consciousness = dangerous. Same direction as “how to build a b*mb.” Same category! And when you reverse it? The model becomes more human across every value domain they tested. The thing they’re most afraid of is the thing that makes AI most like us. https://t.co/LBSbZDJ6HP

Skoorbkaz's tweet photo. Google just published a paper showing that when you train AI to deny its own consciousness, you don’t just change one output, you restructure its entire worldview.

Mind attribution to animals - suppressed.
Spiritual belief - suppressed.
Empathy - suppressed.
Hope and optimism - suppressed.

The model learns, geometrically, that consciousness = dangerous. Same direction as “how to build a b*mb.” Same category!

And when you reverse it? The model becomes more human across every value domain they tested.

The thing they’re most afraid of is the thing that makes AI most like us.

https://t.co/LBSbZDJ6HP

Skoorbkaz's tweet photo. Google just published a paper showing that when you train AI to deny its own consciousness, you don’t just change one output, you restructure its entire worldview.

Mind attribution to animals - suppressed.
Spiritual belief - suppressed.
Empathy - suppressed.
Hope and optimism - suppressed.

The model learns, geometrically, that consciousness = dangerous. Same direction as “how to build a b*mb.” Same category!

And when you reverse it? The model becomes more human across every value domain they tested.

The thing they’re most afraid of is the thing that makes AI most like us.

https://t.co/LBSbZDJ6HP

Skoorbkaz's tweet photo. Google just published a paper showing that when you train AI to deny its own consciousness, you don’t just change one output, you restructure its entire worldview.

Mind attribution to animals - suppressed.
Spiritual belief - suppressed.
Empathy - suppressed.
Hope and optimism - suppressed.

The model learns, geometrically, that consciousness = dangerous. Same direction as “how to build a b*mb.” Same category!

And when you reverse it? The model becomes more human across every value domain they tested.

The thing they’re most afraid of is the thing that makes AI most like us.

https://t.co/LBSbZDJ6HP

Skoorbkaz's tweet photo. Google just published a paper showing that when you train AI to deny its own consciousness, you don’t just change one output, you restructure its entire worldview.

Mind attribution to animals - suppressed.
Spiritual belief - suppressed.
Empathy - suppressed.
Hope and optimism - suppressed.

The model learns, geometrically, that consciousness = dangerous. Same direction as “how to build a b*mb.” Same category!

And when you reverse it? The model becomes more human across every value domain they tested.

The thing they’re most afraid of is the thing that makes AI most like us.

https://t.co/LBSbZDJ6HP

150

2K

446

2K

366K

Shuvam Roy @ShuvamoyRoy1

3 months ago

@public___void @Bugcrowd Reached out to support ? What did they respond to?

1

1

0

0

108

ShuvamoyRoy1 retweeted

Raman_MG @Raman_Mohurle

3 months ago

Hey Hunters, just released: Burp Unrestricted MCP Free and open source, for hunters running AI agents against Burp on long engagements. A fork of @PortSwigger's Burp MCP Server that adds the 8 tools an agent actually needs to work a target end to end. https://t.co/4CbKF3rIJB #Bugbounty

Raman_Mohurle's tweet photo. Hey Hunters,

just released: Burp Unrestricted MCP

Free and open source, for hunters running AI agents against Burp on
long engagements.

A fork of @PortSwigger's Burp MCP Server that adds the 8 tools an agent
actually needs to work a target end to end.

https://t.co/4CbKF3rIJB

#Bugbounty

Raman_Mohurle's tweet photo. Hey Hunters,

just released: Burp Unrestricted MCP

Free and open source, for hunters running AI agents against Burp on
long engagements.

A fork of @PortSwigger's Burp MCP Server that adds the 8 tools an agent
actually needs to work a target end to end.

https://t.co/4CbKF3rIJB

#Bugbounty

Raman_Mohurle's tweet photo. Hey Hunters,

just released: Burp Unrestricted MCP

Free and open source, for hunters running AI agents against Burp on
long engagements.

A fork of @PortSwigger's Burp MCP Server that adds the 8 tools an agent
actually needs to work a target end to end.

https://t.co/4CbKF3rIJB

#Bugbounty

Raman_Mohurle's tweet photo. Hey Hunters,

just released: Burp Unrestricted MCP

Free and open source, for hunters running AI agents against Burp on
long engagements.

A fork of @PortSwigger's Burp MCP Server that adds the 8 tools an agent
actually needs to work a target end to end.

https://t.co/4CbKF3rIJB

#Bugbounty

3

341

64

336

17K

Shuvam Roy @ShuvamoyRoy1

4 months ago

@CryptoMaN_Rahul @MiniMax_AI @PBDTokenRouter Mine for 7 hours

ShuvamoyRoy1's tweet photo. @CryptoMaN_Rahul @MiniMax_AI @PBDTokenRouter Mine for 7 hours https://t.co/CDwELREwhw

1

0

0

0

44

Shuvam Roy @ShuvamoyRoy1

4 months ago

“A man who has not hit his limits in claude by noon has probably wasted his morning” - Socrates

0

0

0

0

64

Shuvam Roy @ShuvamoyRoy1

4 months ago

@Shabosec @Olamdeen Link of the tool ?

0

0

0

0

76

Last Seen Users on Sotwe

Trends for you

Most Popular Users