Second Part To Hell

artificial (life, evolution, intelligence)

[vx-underground @vxunderground]

i really like malware

[Peter Steinberger 🦞 @steipete]

@naravelych Better we explore this now then in 5 years when agents are even more powerful.

[Unemployed Capital Allocator @atelicinvest]

.@steipete

[tmp.0ut @tmpout]

We are extending our call for papers to January 1, 2025!

We are now targeting an end of January release.

If you have any Linux/ELF related research, projects, or papers, we would love to publish them!

Huge thanks to everyone who has already submitted!

https://t.co/95d9RLZ7Pt

[Second Part To Hell @SPTHvx]

Imagine...

[Mandy Stadtmiller @mandystadt]

16 partially grown human brains with no mouths, tiny eyes called organoids that live for 100 days comprise the “world’s first bioprocessor” for a “million times less power than a digital chip” and omfg we are now in fact living in The Matrix

[Second Part To Hell @SPTHvx]

Introducing LLMarshal 0.1, a GPT-based #antivirus demo, that can effectively fight against LLM viruses (such as LLMorpher):
https://t.co/9BFg45k6vn

Good AI is beating Bad AI!

Motivated by seeing that after 9 months, all 58 AVs at @virustotal fail detecting LLMorpher!

[Second Part To Hell @SPTHvx]

🔆March 2023: First virus using #GPT to auto-mutate itself (LLMorpher1&2)

🔆August 2023: Advanced virus mutations using #GPT4 (LLMorpher3).

Until now, no additional codes? Would have expected #LLama2 viruses by end of 2023. @ylecun @karpathy

Code: https://t.co/1uxiEvx831

[@mikko @mikko]

@adulau Search for ’LLMorpher’.

[tmp.0ut @tmpout]

tmpout vol3 - article a day #12

"23 Open Problems for Digital Self-Replicators"

@SPTHvx presents 23 self-replication problems to the upcoming generation of programmers

read here: https://t.co/wJsMvFoobA

[Second Part To Hell @SPTHvx]

Nice overview by @ziyaxanalbeniz about my GPT-virus LLMorpher (https://t.co/Uom5HLrShc)

Shows also how a local #LLama2 copy can be used to hide malicous code, using LLMorpher's idea:

Code as english-language description translated to Python by a LLM.

https://t.co/ma5ZNyJi0k

[tmp.0ut @tmpout]

@tmpout vol3 - article a day #2:
"Second Part To Hell Interview"
SPTH was an early 2000s virus powerhouse as member of rRlf and author in 29a and Valhalla zines. Check out the interview before you read his open challenges in virus research.
we invite comments and feedback!

[Second Part To Hell @SPTHvx]

⚡️ 📜 Full Metamorphism of computer virus Code and Prompts via #GPT4

A computer virus autonomously uses GPT4 to entirely rewrite its own code and its prompts.

Text: https://t.co/e4VfqdaIfG

Code: https://t.co/1uxiEvxFSz

[Andrej Karpathy @karpathy]

The hottest new programming language is English

[Second Part To Hell @SPTHvx]

Inspired by the century-shaping list of Hilbert in math, today I publish my vision🤖

23 Open Problems for Digital Self-Replicators☣️

Large Language Models, Reinforcent Learning, Swarm Intelligence, DNA infection, Brain-Computer Interfaces, Consciousness

https://t.co/qp78IrmzaV

[Second Part To Hell @SPTHvx]

What an amazing old-school vx ezine! Congrats to the @tmpout gang!

I contributed:
- 3 GPT-powered viruses, including a brand-new #GPT4 fully metamorphic LLMorpher III

- 23 Open Problem in Digital Self-Replicators

Also contains an interview with me, thx @guitmz

[tmp.0ut @tmpout]

Check it out, it's tmp.0ut Volume 3!

https://t.co/Gm9erN8hTJ

[tmp.0ut @tmpout]

Check it out, it's tmp.0ut Volume 3!

https://t.co/Gm9erN8hTJ

[tmp.0ut @tmpout]

Get hyped, volume 3 is coming soon!

[Second Part To Hell @SPTHvx]

Just uploaded online-viewers for some of the most iconic vx-zines since the early 90s, such as 29A, iKx, matrix, Codebreakers, Coderz-net, rRlf, valhalla, tmp.0ut:

https://t.co/d5KyMB3T4i

Which if your favorite historic ezine is missing?

[Second Part To Hell @SPTHvx]

Perfect. After computer viruses with ...
- Artificial Evolution (2010/11)
- Strong Metamorphic Scripts (2012)
- DNA infection (2013)
- Metamorphism via GPT (2023)

I know what my next topic will be...

...see you in a few years 👋

[Jim Fan @DrJimFan]

"Consciousness" has been the taboo "C-word" in AI.

This 88-page paper, coauthored by Turing Prize laureate Yoshua Bengio, is a systematic survey of the scientific theories of consciousness, as well as possible implementations in today's AI stack.

I commend their courage to address such a controversial topic that most researchers shun. Here's what the paper concludes: quote "Our analysis suggests that no current AI systems are conscious, but also shows that there are no obvious barriers to building conscious AI systems."

It's a long read - will share notes later, bear with me.

Paper: https://t.co/dA6RZDAqhr

[Second Part To Hell @SPTHvx]

In March 2023, i published the virus LLMorpher, which encode its functions in english langauge & use @openai's GPT to get the code: https://t.co/Uom5HLrShc

Now, 5 months later - it is not detected by a single (!) AV program. This is worrisome.

@vxunderground @mikko @karpathy

[vx-underground @vxunderground]

Inspired by our friend @SPTHvx, let's review a collection of malware author interviews from the 80s, 90s, and 2000s.

In these interviews, the interviewer asked "what do you think the future of malware looks like?".

Let's open a time capsule on a Saturday night =D

[Second Part To Hell @SPTHvx]

Here I collected predictions about virus-writing future from 1993-2012 (from the leaders of the field, including members of iKx, 29A, rrlf, valhalla):
https://t.co/BKQjCUZHPI

Now, 10 years later, i am curious about your vx predictions for 2025, 2030, 2050! 🧐