GC downtime (August 26) due to flood of Bots

[John]

All the downtime today was due to bots, specifically AI data scraper bots as far as I can tell but might also be search engine related.

They were sending far too many bots way too fast and it was making GC unusable. Basically caused an unintentional DoS (denial of service) "attack" on GC. So I turned GC off while getting that all sorted out.

The IP addresses were from a major Internet company (hundreds or maybe even thousands of IPs) and were being used by a different major Internet company.

TikTok / Bytedance apparently must be really quite interested in the conversations going on here at GreekChat!

[blueGBI]

Must be due to #bamarush!

[carnation]

Can they be blocked?

[cheerfulgreek]

Is that what happens? The more interesting the site, the more bots that show? I don’t understand what they collect and the importance of what they collect. And then where does what they collect go?

[John]

Originally Posted by carnation

Can they be blocked?

Usually, yes. But not always. Depends on how determined the bot/spider operators are.

I rerouted these particular bots in a way which should minimize their impact on GC. Similar to what I did with the bots from approx 2 weeks ago.

Originally Posted by cheerfulgreek

Is that what happens? The more interesting the site, the more bots that show?

Yes, that's pretty much how it goes as far as I know. At least with search engine type bots as search engines want to direct their users to interesting & useful sites.

GoogleBot is a good example of a beneficial bot as that bot indexes sites & helps increase listings in search which might result in more new website visitors.

Originally Posted by cheerfulgreek

I don’t understand what they collect and the importance of what they collect.

The bots from Bytedance / TikTok seem to be AI data scraper bots. But might also be search engine related as well.

They might be training AI / artificial intelligence chatbots using conversations that they scrape from forums all throughout the Internet.

Around a decade ago the world entered a new era of AI. Way back in the day one way AI was made was with programmers creating decision trees to determine what to do depending on different scenarios and that would be many levels deep. Modern AI doesn't do that. Instead, modern AI gives the engine, the neural network, as much labeled data as possible which the AI trains on in order to recognize patterns within the data. The AI is trained & tested over and over, repeatedly. Constantly refining, tweaking the neural net while continually improving accuracy. AI recognizing patterns in images is a good example of this type of training.

With GC being text based conversations, AI data scraping on that is likely to be for training AI language models. Basically very sophisticated chatbots. Similar to ChatGPT if you're familiar with that. Huge amounts of data, as much as possible, is used in training these AIs.

Originally Posted by cheerfulgreek

And then where does what they collect go?

My guess is that if used for training AI, the data is refined and is used for training AI now and maybe even in the future. Could be one giant combined dataset of many different online forum discussions that they continually add to & improve. All that data may be used for training AI, but might never actually be seen by anyone other than the AI researchers and/or data engineers working on it.

[cheerfulgreek]

Based on your post, John, it sounds like bots will be the reason for a future AI takeover. I mean, with the information they’re getting, it’s eventually going to be used for AI in everything. That’s the way it reads.

[John]

Originally Posted by cheerfulgreek

Based on your post, John, it sounds like bots will be the reason for a future AI takeover. I mean, with the information they’re getting, it’s eventually going to be used for AI in everything. That’s the way it reads.

I'd probably blame the future AI takeover more on math & programming in general. Modern AI probably boils down to highly accurate pattern matching that is enabled by complex math & statistics. Programming enables it to be automated and fast.

But it does need lots of data for training, so bots are a part of it.

Modern AI is such a revolutionary advancement that any businesses (or governments) that can benefit from it need to adopt it or eventually they will be surpassed. The disruption will be similar, probably more significant, to when the Internet started being commercialized and ecommerce was disrupting businesses in all different industries.

[cheerfulgreek]

Originally Posted by John

I'd probably blame the future AI takeover more on math & programming in general. Modern AI probably boils down to highly accurate pattern matching that is enabled by complex math & statistics. Programming enables it to be automated and fast.

But it does need lots of data for training, so bots are a part of it.

Modern AI is such a revolutionary advancement that any businesses (or governments) that can benefit from it need to adopt it or eventually they will be surpassed. The disruption will be similar, probably more significant, to when the Internet started being commercialized and ecommerce was disrupting businesses in all different industries.

Yep. I totally get that AI is going to be big business, if it isn’t already. It’s just that the possibilities for future AI systems are limited by what we currently know about the nature of intelligence. And then, I think it’s going to be a huge mess, because based on the way I see things happening with AI, it seems like all evidence suggests that human and machine intelligence are radically different. Know what I mean? Don’t you think it will be more problematic than beneficial, based on that? I mean, you’d know better than me, but I just don’t see the benefits.

[John]

Originally Posted by cheerfulgreek

AI systems are limited by what we currently know about the nature of intelligence

I think "AI" might be more correctly referred to as "APM" / "artificial pattern matching." AI systems aren't "intelligent" but they can in many ways now replicate the patterns of intelligence well enough to convince us of AI "intelligence" and that ability/feature of AI will certainly improve in the future. Future AI will be so good at doing that most people will likely believe that what they are seeing is actual intelligence when it's just pattern matching.

Originally Posted by cheerfulgreek

seems like all evidence suggests that human and machine intelligence are radically different

In regards to how the brain does what it does and how AI does what it does I would probably agree that it's radically different. But the end result I think is quite similar. Humans, as far as I understand, do much of our learning based on patterns that we perceive with new or more complicated learning often or usually based on the patterns that we previously perceived. That is very similar to AI in that AI is an extremely sophisticated pattern matching calculator, in a sense.

Originally Posted by cheerfulgreek

Don’t you think it will be more problematic than beneficial, based on that?

Great question. Probably both more problematic and more beneficial. But I'm not sure which one will be more significant.

I guess in comparing possible end results... AI could make life significantly better, more productive, more advancement, more efficient, more possibilities, etc. But future AI could also result in disaster, or maybe be used by people in order to cause a disaster scenario. So from that perspective, maybe AI will be more problematic because of the potential for serious harm.

But that cat is out of the bag, no going back in. Even if it's regulated it's probably too late. Those who are regulated from developing new AI will simply be those who lose out to the people who take advantage of the technology.

Consider Google's AI AlphaGo beating world master champions in the game of Go. AlphaGo was trained based on actual games between people. Then Google AI devs went and made AlphaGo Zero which was given the rules of Go and trained itself, basically played itself for a period of time and learned to master the game that way. AlphaGo Zero defeated AlphaGo significantly. I think at first it was just by a relatively small margin, then after some changes AlphaGo Zero won 100% against AlphaGo. I don't know the exact history of it, but it was something like that.

So that was a game that an AI mastered versus people and another AI completely defeated the first AI.

Consider that the military has war "games" for training & analyzing scenarios, etc. The US military not only should be using AI, they must be. They have to. Because if they are not and an enemy or future potential enemy develops AI that can defeat all the human military war game "players," like the AIs in the actual game Go, the side that does not have that advantage probably loses. And it might be that the government which reaches AI superiority first is the one that wins it all.

Then there are governments that will use massive data collection & AI training/development against their own people.

Lots of great benefits but also lots of very serious potential problems.

[cheerfulgreek]

^^^ Oh wow! I didn’t look at it from any of the perspectives you’re looking at it from, John. So true, and you make a lot of sense. I think greed will be the downfall of it, though. I mean, for like the other side of the benefits. The bad side. It’s just that a lot of companies now launch AI teams because they’re afraid of falling behind other companies/competitors, without fully knowing where or for what purpose they’ll use AI. And then too, a lot of companies pretend to use AI when they don’t, just to increase their chances of obtaining funding. That’s the greed part. That’s the part I think will get worse. There’s like also a fair amount of general confusion about what AI can and can’t do.

What’s interesting though is that we now use it a lot, daily, sometimes without even realizing it. Do you think it can or will get out of control? I mean, right now AI is completely under human control, but in the future, it might not be under our control anymore. Seems like eventually every single task is going to be done by AI.

[John]

Originally Posted by cheerfulgreek

Do you think it can or will get out of control?

Definitely yes to the first question and probably yes to the second.

In some sense, there have been AIs which temporarily got out of control until they were reigned in. That would be some computer viruses. Although they aren't modern AI.

Eventually there will be modern AI based computer viruses, though.

With that in mind, there will be AI trained to hack computers. I'd be really surprised if that hasn't already been accomplished. If it hasn't or at the very least if it isn't being worked on & developed then many people in the computer security side of government (and cyber warfare) need to lose their jobs. Hacking is like solving puzzles... puzzles are games, basically & AI has already mastered many games, so why not hacking. Problem is that if we don't do this what if an enemy does. Train the AI to hack, train the AI to find new vulnerabilities, train the AI to infect, to evade detection, stuff like that and we have a real problem, especially if integrated into a virus.

Then there will be AI trained to protect against AI that is trained to hack. And who knows which side will win that battle.

But it doesn't even need to be an AI designed to do bad things for it to get out of control. Could be a completely well intentioned AI but some aspect of the logic was missed or not protected and the AI determines a solution to whatever problem it is solving is to do something really terrible.

It all seems like science fiction but we probably aren't too far off from having self driving cars that are safer than people driving. AI controlling things that have the potential to do a lot of damage becomes more and more real over time.

[cheerfulgreek]

This is interesting, John. I remember when I was a flight attendant, our airplanes had auto-land, so they could land themselves. I mean, the pilots still had to make a few adjustments, but not many. Human error is what generally causes plane crashes, when they rarely happen. I’m sure the same would be true with self-driving cars and trucks.

I hope a massive cyber attack where it affects how we live never happens. There are so many things we take for granted that can be wiped out by cyber attack from an enemy.

[John]

Originally Posted by cheerfulgreek

airplanes had auto-land

Originally Posted by cheerfulgreek

Human error is what generally causes plane crashes, when they rarely happen. I’m sure the same would be true with self-driving cars and trucks.

My guess is that auto-land is more like an advanced type of cruise control. Although maybe those types of systems have been integrating moderin AI in recent years.

Self driving cars are based on modern AI tech which is quite different in how it functions.

I don't know if self driving cars are yet statistically safer than people driving. But over time I'm fairly sure that self driving AI will approach then eventually surpass the ability of people to drive safely.

Originally Posted by cheerfulgreek

I hope a massive cyber attack where it affects how we live never happens.

There have been many, but for the most part were able to be resolved, as far as I know.

One example is the Colonial Pipeline crypto ransomware attack in 2021. That cyber attack had an almost immediate significant impact, but the company was able to resolve fairly quickly. When that pipeline was shut down it impacted most of the US East Coast and several States declared States of Emergency due to it. The day after that cyber attack the company paid $5 million in ransom and luckily the hackers provided the fix after getting their ransom.

There have been many companies and different levels of government that have had to pay ransom to hackers to recover essential systems. I recall reading of a number of companies that went bankrupt after not being able to recover from cyber attacks.

Originally Posted by cheerfulgreek

There are so many things we take for granted that can be wiped out by cyber attack from an enemy.

That's for sure.

[cheerfulgreek]

Originally Posted by John

One example is the Colonial Pipeline crypto ransomware attack in 2021. That cyber attack had an almost immediate significant impact, but the company was able to resolve fairly quickly. When that pipeline was shut down it impacted most of the US East Coast and several States declared States of Emergency due to it. The day after that cyber attack the company paid $5 million in ransom and luckily the hackers provided the fix after getting their ransom.

There have been many companies and different levels of government that have had to pay ransom to hackers to recover essential systems. I recall reading of a number of companies that went bankrupt after not being able to recover from cyber attacks..

I remember that. Ok, so I read this post a few days ago and I began to wonder since it’s like nearly impossible to catch these criminals who do this, why don’t they just keep doing it and requiring ransom money? It’s not like they can get caught. I remember when that happened and I wondered the same thing when it happened.

[John]

Originally Posted by cheerfulgreek

I remember that. Ok, so I read this post a few days ago and I began to wonder since it’s like nearly impossible to catch these criminals who do this,

Around a year ago I watched an interview of a guy who was part of some FBI task force for investigating high profile / most wanted hackers, or something like that. He said that we will never catch them and the only ones who do get caught are usually because they made some sort of tiny mistake. There's a documentary about the silk road hacker who was basically caught this way, made a little mistake.

Originally Posted by cheerfulgreek

why don’t they just keep doing it and requiring ransom money? It’s not like they can get caught. I remember when that happened and I wondered the same thing when it happened.

I read something about this a while back. Seems to be that if the hackers get a reputation for not honoring their ransoms then people / companies / governments will probably stop paying future ransoms.