Wednesday, August 7, 2024

Zombie AI Death Match

Zombie AI Death Match
Who would win in a fight between a zombie and an AI? 

The answer, as with most hypothetical conflicts between fantastical opponents, is: it depends. Consider the classic debate of bear versus shark - the outcome hinges entirely on where the fight happens. In the ocean, the shark always wins; on land, the bear.

I asked Claude Sonnet 3.5. For those of you who don't live in the nerdosphere, Claude Sonnet is an LLM. Here's what it said:

"In a hypothetical conflict between a zombie and an AI, the AI would likely prevail due to its superior processing power and lack of physical vulnerabilities. Unlike zombies, AI isn't constrained by biological needs or limitations. It could operate remotely, analyze the zombie's behavior patterns, and devise effective countermeasures without risking direct confrontation. However, this scenario raises deeper questions about the nature of intelligence, consciousness, and the potential risks of advanced AI systems that can outmaneuver and outthink biological entities, even hypothetical ones like zombies."

Of course, it would say that. It's an AI. 

However, on balance, I think it's right. Unless they were to clash in a post-apocalyptic world, where the AI is hosted on a single, antiquated machine [think Windows ME] and the Internet is unavailable. Then, the zombie wins every time.

Take that, Skynet.

What does this have to do with reality? I was going to title this post: AI Won't Take Your Job, But Someone Using AI Will. I think the main difference between AIs and humans is volition: we've evolved to desire things, and this causes us to reach out and get them.

The important word there is 'evolved.' Desire is an evolved characteristic, the result of the original protoplasm reaching out to acquire some kind of resource, thus gaining a survival advantage, and Chucky D took care of the rest.

Of course, you might not believe in evolution. If so, click here to win a prize.

What does that have to do with zombies? Zombies are a MacGuffin: a filmic device to solve any problem. They are a special type of dead human.

Our goal in this blog is to not become a dead human.

About me:



I'm Nikolai Ross. I'm an author and engineer. I write sci-fi adventures. My new novel A Lightship to Aldebaran is on Amazon now. Email me at nikolai.ross@proton.me for your free copy.

Tuesday, August 6, 2024

Welcome, my son.
Welcome to the machine.

Welcome to the Machine
With apologies to Pink Floyd.

We're all idiots now.

Computers have always had us beat when it comes to outright performance, from the Antikythera mechanism [built circa 100 BCE], through Alan Turing's Bombe, to ENIAC and modern supercomputers. That's why we invent them.

Even my Hewlett Packard HP25 scientific calculator [yes, I'm that old] would have outperformed almost any human in calculations. I only say 'almost' because a few exceptional humans might solve some problems faster than I can type them.

Today, no human can realistically hope to beat a computer at chess. As I write this [and yes, this is me, not ChatGPT], the Stockfish engine is rated at 3634, compared to Magnus Carlsen's chess rating of 2832.

What hope do we have? How can we survive in competition for resources, like shelter and food? How do we compete in a world where machines surpass us in every measurable metric?

This is the question that underlies much of the current debate about AI safety. There are detailed [and important] questions about competition for jobs and how well machines understand what we ask them, but the big question is this: is there a future for us on this planet? Have we evolved ourselves out of the picture?

Thankfully, computers don't think like we do. They're like an alien intelligence. They don't appear to desire things. They don't compete unless we program them to. Magnus Carlsen can still queue for a burger without fear of a computer cutting in line.

How long can we rely on that being true?

This blog isn't about AI safety. It's not about machine sentience, or the emerging AI singularity. It's about people.

Leopold Aschenbrenner writes in his fascinating [and rather scary] Situational Awareness paper that he's lucky enough to be in a small group of exceptional people working in Silicon Valley. Imagine you're not one of them. You might be an engineer, or a high school student, or a factory worker in some deprived northern England mill town.

What should you do? How will you survive - and thrive - in the brave new world of AI, where properly informed people are dealt a new hand?

Be aware, everything is about to change.

This blog is for you.

About me:



I'm Nikolai Ross. I'm an author and engineer. I write sci-fi adventures. My new novel A Lightship to Aldebaran is on Amazon now. Email me at nikolai.ross@proton.me for your free copy.




Zombie AI Death Match

Who would win in a fight between a zombie and an AI?  The answer, as with most hypothetical conflicts between fantastical opponents, is: it ...