Ankit Anubhav 

@[email protected]
625 Followers
86 Following
375 Posts

Cybersecurity Researcher, Physics, Astronomy, Linux, FOSS, contributing back to the community.

Other sub-topics of interest include time travel, parallel universes, generational space ships, mathematics, anaerobic and non carbon based life forms, voyager 1+2,chess, black holes, vaccum tubes, quantum computing, metaphysics, James Webb telescope, cryopreservation and hydroponics.

https://x.com/ankit_anubhav

Websitehttps://www.ankitanubhav.info
IOC sharinghttps://raw.githubusercontent.com/aanubhav-ioc/random/main/Verification.html
Ankit Anubhav Mar 7
Life is too short a simulation to take things too seriously..
Ankit Anubhav Dec 1

People who execute everything which the mcp server comes up with ( execute all setting in cursor etc.) are the new "person who clicks all the phishing links"

#ai #phishing #mcp

Ankit Anubhav Dec 1

While snake oil companies continue to get hyped and get absurd amount of financing, Carmat, one of the only companies in world which had implanted artifical heart in humans, had to file for bankruptcy due to not getting 3.5 million dollars.

We are truly living in a dystopian world..

https://www.carmatsa.com/en/news/carmat-announces-filing-today-for-insolvency-and-requesting-being-placed-in-receivership/

CARMAT announces filing today for insolvency and requesting being placed in receivership - Carmat

Carmat
Ankit Anubhav Nov 27
Just knew that the Humane AI pin, which raised 230 million $, already has become defunct (both the company and the device).
Ankit Anubhav Nov 26
Graham CluleyNov 26

Shadow AI - the use of artificial intelligence tools by employees without a company's approval and oversight - is becoming a significant cybersecurity risk, with Gartner predicting 40% of global organisations having suffered security breaches due to it by 2030.

What are you doing about it?

Read my article on the Fortra blog to learn more: https://www.fortra.com/blog/shadow-ai-security-breaches-will-hit-40-companies-2030-warns-gartner

Ankit Anubhav Nov 23

It's crazy how Atlas 3I is never going to come back due to its hyperbolic path, has so many anomalies, is the 3rd ever interstellar object, and yet we get a vanilla poor coverage by nasa.

The world's focus is only to make "Agentic AI", 15 second brain frying videos and other trivial stuff to keep humans inside the matrix.

#space #atlas3i

Ankit Anubhav Nov 10

If you summarize any documents using your AI, there is a risk that they might have a prompt to exfiltrate the AI memory. One does not need a post request as this can be done by URL browsing component itself ( exfil via GET). To avoid suspicion, the prompt can be hidden in 1 size font and white background, and a fake decoy "Request timed out" can be displayed. Read more of my research with Papon on the same.

https://t.co/RE7GXuurCe

#ai #cybersecurity

AI Summarization Gone Rogue: Memory Exfiltration via Documents

- Ankit Anubhav, Papon Trongprakorb One of the most popular uses of AI today is to save time—think of how often we use it to shorten or summarize documents. Whether it’s a report or an article, everyone wants the gist so they can stay efficient.

Ankit Anubhav Nov 7

"AI will soon supersede human intelligence.

This is not necessarily because AI is becoming intelligent, but humans using it are becoming dumber and dumber". - Dr. Avi Loeb

"Not necessarily dumber, but lazier, one should use AI for learning, not substitute learning for AI usage." - Joe Rogan

Interesting thoughts. #ai #gpt

Ankit Anubhav Nov 6

There is a lot of unnecessary criticism of Dr Avi Loeb. Yes he likes to keep it interesting, but he focuses on the real data and let it speak for you. Not a conspiracy theory guy but an interesting person.

#3iatlas #space #aliens

Ankit Anubhav Nov 5

Atlas 3I is becoming intriguing day by day. I'm just thinking how will voyager be different from a rock if it hits near a planet in 10 million years. Maybe a rock with slightly different composition?

#space #astronomy