Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

AI Joins Hunt for ET: Study Finds 8 Potential Alien Signals

blogs.nvidia.com

Post reply on HN
Print view View original

61–70 of 161 posts

Re: AI Joins Hunt for ET: Study Finds 8 Potential Alien Signals

#61

Post by edgyquant »

godelski wrote: post #55
readyplayeremma wrote: post #26

Earlier quoted context omitted.

Our existing computing platforms are very specific and very limited. Without knowledge of that specific design and the accompanying limitations, how could you pre-craft a non-interactive "exploit" that could be executed by such a system? I do think that it becomes more possible if we have AGI-like systems doing detection and analysis, but we do not. In addition, any universal exploit against AGI systems would probabl…

For a clearer example, there's a reason a virus that affects Windows doesn't affect Linux. Or a virus that affects Windows XP doesn't affect Windows 11. There are of course counter examples, but this is the trend. To understand why the counter examples exist requires expert knowledge (or rather that there are basically the same things running in the same way).

We are currently discussing exploits designs by foreign intelligences to infect artificially intelligent systems. We are not talking about privilege escalation on a Mac.

Re: AI Joins Hunt for ET: Study Finds 8 Potential Alien Signals

#62

Post by edgyquant »

throwbadubadu wrote: post #28
edgyquant wrote: post #19

Earlier quoted context omitted.

>it seems unlikely What are you basing this on?

We've built enough stupid complexity* that either the aliens must already have spied it out completely (but if they done that they could likely overtake is in any way quickly..), or it just won't happen. (*) I mean seriously, how would you design a virus in some limited signal that has a chance of overtaking any arbitrary specific system someone may have invented, that's impossible. Never say never maybe my imaginati…

It would likely be pretty “easy” to design a signal that tricks convolutional networks etc. no need to care about the underlying hardware

Re: AI Joins Hunt for ET: Study Finds 8 Potential Alien Signals

#63

Post by akkartik »

jerf wrote: post #34
O__________O wrote: post #2

Curious, assuming such an signal is found, are there safety measures in place to isolate it from other systems in off chance it contains “alien malware” that attacks systems receiving it?

It's a bit of a science fiction myth that it is impossible to process a bit of data without executing it, or that data can somehow force itself to be executed by its sheer swaggering potency as data. And a thing that is more true than it should be, because people keep insisting that they are super awesome and totes capable of writing C that can be put on the network. If we'd stop doing this there'd be hardly any trut…

Focusing on the non-scifi angle here, there's one other reason besides C that we end up constantly running data when we think we're parsing it: the depth of our computational stack.

https://www.sitepoint.com/anatomy-of-an-exploit-an-in-depth-...

https://gitlab.com/gitlab-org/gitlab/-/issues/371098

So I'd say we need 3 things:

* Stop using unsafe languages

* Use languages that separate parsing from execution. Just follow Lisp, have equivalents for `read` and `eval`. Lua is a notable offender here: https://www.lua.org/manual/5.4/manual.html#pdf-load There is no way to parse a table while guaranteeing no code execution.

* Use languages that forbid monkey-patching, because that's one vector for turning `read` into `eval` because someone had a bright idea.

Re: AI Joins Hunt for ET: Study Finds 8 Potential Alien Signals

#64

Post by jjk166 »

retrac wrote: post #44
jerf wrote: post #34

Earlier quoted context omitted.

It's a bit of a science fiction myth that it is impossible to process a bit of data without executing it, or that data can somehow force itself to be executed by its sheer swaggering potency as data. And a thing that is more true than it should be, because people keep insisting that they are super awesome and totes capable of writing C that can be put on the network. If we'd stop doing this there'd be hardly any trut…

Assuming the bandwidth exists, somehow describing enough mathematics to describe a virtual machine, and then sending a program for that virtual machine, might just be the best way to communicate. A package that can help the recipient interpret it by learning and responding to the local environment. I can't put my finger on any specific book but I have the impression of this scenario coming up in science fiction. I've…

I wonder if in the same way you can do statistical analysis of human languages to determine things like which characters are vowels, perhaps if you received enough programs written in some alien brainfuck-like language you could use statistical analysis to guess which symbol is which.

If nothing else you could always brute force it; with 8 logical operators there would only be about 40,000 possible combinations. Maybe figure an order of magnitude larger number to account for idiosyncrasies in their implementation of the language. Running "hello world" a half million times shouldn't be too hard. Of course figuring out which ones are spitting out gibberish and which ones are spitting out perfectly intelligible results for one fluent in alienese would be hard. I presume they'd send a bunch of example programs that compute pi or something similarly universal, but there are a lot of potential gibberish we could fit a pattern to. It might even make sense to just keep running every code sequence in every possible combination until we get something cool.

Re: AI Joins Hunt for ET: Study Finds 8 Potential Alien Signals

#65

Post by 1970-01-01 »

O__________O wrote: post #2

Curious, assuming such an signal is found, are there safety measures in place to isolate it from other systems in off chance it contains “alien malware” that attacks systems receiving it?

Not sure how this could even theoretically work. Without having a partial understanding of the intended recipient, a radio signal can't cause harm. Something like 'this signal is false' and 'divide by 0' need to be decoded and processed before they are malicious. Meat space is excellent at detecting and breaking these logic problems and infinite loops. We're safe from evil alien signals. Perhaps a decompression bomb within the signal could cause lots of decoding grief. Now that I think about it, I can imagine us being stupid enough to dedicate all our resources into decompressing or cracking an alien signal to up to the point of extinction. And maybe that's when they step in and call it. We were just part of a great galactic prank.

Re: AI Joins Hunt for ET: Study Finds 8 Potential Alien Signals

#66

Post by haupt »

readyplayeremma wrote: post #32
zh3 wrote: post #22

>The AI identified 20,515 signals of interest, which the researchers had to inspect manually. Of those, eight had the characteristics of technosignatures and couldn’t be attributed to radio interference. The obvious question being, if the AI is so smart why was it necessary to use humans to check 20,515 signals to find the eight with the "characteristics of technosignatures"?

AI is not a well-defined term. This is more likely a specific machine learning technique that is designed to identify all the boring "normal stuff" and ignore it, and then do that at very large scale. By doing that, you can find the interesting parts so that humans with more limited resources can determine if those newly flagged things need to be added to the boring list, or if they do represent something truly inter…

>The term is clearly used here for marketing purposes.

I think most uses are for sensationalistic purposes. I'd wager almost nobody in the general public really understands what AI is or can pin it down. It doesn't help when so many different media outlets abuse the terminology by using it to refer to different things. What's even worse is that whatever ideas people have about AI tend to come from Hollywood.

Re: AI Joins Hunt for ET: Study Finds 8 Potential Alien Signals

#67

Post by flangola7 »

tshaddox wrote: post #58
zh3 wrote: post #22

>The AI identified 20,515 signals of interest, which the researchers had to inspect manually. Of those, eight had the characteristics of technosignatures and couldn’t be attributed to radio interference. The obvious question being, if the AI is so smart why was it necessary to use humans to check 20,515 signals to find the eight with the "characteristics of technosignatures"?

How do you judge whether 20,515 is a large number or a small number? If the AI provided 2 candidates is that still too high?

SETI@home sorted through I think trillions of data points.

Re: AI Joins Hunt for ET: Study Finds 8 Potential Alien Signals

#68

Post by detrites »

m3kw9 wrote: post #24
zh3 wrote: post #22

>The AI identified 20,515 signals of interest, which the researchers had to inspect manually. Of those, eight had the characteristics of technosignatures and couldn’t be attributed to radio interference. The obvious question being, if the AI is so smart why was it necessary to use humans to check 20,515 signals to find the eight with the "characteristics of technosignatures"?

Because it’s not so smart, it just try to recognize a set of patterns that was trained to look at

Which is what we do. The rest of what we believe makes us special is simply emergent, and arbitrarily defined in terms of just being further useful pattern recognition, rather than any fundamental property of "intelligence".

Re: AI Joins Hunt for ET: Study Finds 8 Potential Alien Signals

#69

Post by dylan604 »

jacobsenscott wrote: post #48
O__________O wrote: post #39

Earlier quoted context omitted.

Assuming AGI existed, detected it, and understood the signal’s message — things might get interesting depending on what the message said and how the AGI responded.

Needing at least 3 round trips to establish an ssl connection to the alien internet - things won't get interesting for a few thousand years.

i'm sure an AI SSL version would be much more efficient than 3 round trips. after all, it's supposed to be smarter than what we could make, right?

Re: AI Joins Hunt for ET: Study Finds 8 Potential Alien Signals

#70

Post by varjag »

kayo_20211030 wrote: post #16
pr337h4m wrote: post #5

Earlier quoted context omitted.

"To be sure, because they don’t have real signals from an extraterrestrial civilization, the researchers had to rely on simulated signals to train their models. The researchers note that this could lead to the AI system learning artifacts that aren’t there. Still, Cherry Ng, one of the paper’s co-authors, points out the team has a good idea of what to look for. “A classic example of human-generated technology from sp…

Sounds like a case of assuming ones priors. Of course, should we wish to find human space travelers whose SatNav(tm) has broken, I'm sure it will prove invaluable.

A singleton set beats an empty one.
Post reply on HN