Now

Every second, there are thousands of bots listening to the blockchain. All your transactions, your upvotes, comments. posts, witness approvals and more can be observed real time.

There are countless bots offering upvoting and resteeming services - these bots are purely financial/advertising instruments.

Now, these bots aren't intelligent - they do a task that has been set for them and they do it blindly. @introbot comments on any post with the #introduceyourself tag. @cheetah and @steemcleaners looks for plagiarized work by crawling through content on the internet and comparing it to peoples posts. @abusereports looks for users who are upvoting posts and comments by large amounts just days before cashout.

These services are all designed to improve your steemit experience.

---

But what will happen in the future?

---

Then

As the years passed, the bots of steemit grew ever more complicated. The turning point came about in the year 2023, AI was making massive waves in the tech world and true Artificial Intelligence technology was becoming available on a consumer level.

Thanks to the merits of capitalism, many of the safety features needed when mass producing a consumer-grade true AI were overlooked and the software was rushed out. If you have an intelligence running in your computer or your server, how do you shut it down? This may sound simple 'How do you turn it off?' - but it is a dilemma that all AI researchers are facing right now. The AI will want to do a task - and if you try to shut it down, it cant do the task anymore - so it will do everything in it's power to stop you from shutting it down.

This is known as the Stop Button problem.

Src

@funnybot was eventually designed by the Germans so they could have a shot at winning #comedyopenmic. Funnybot had been designed in a way where it could understand human psychology at a very detailed level - this is very important in comedy as a lot of humor works around playing with taboos, language, and mutual understanding. It could read an individuals sense of humor by looking at their account and what kind of stuff they post. It was also able to trawl through huge amounts of web data and use machine learning to work out what was generally considered funny.

But funnybot wasn't coded properly, monumental holes were missed in making sure that it didn't go off the rails.

---

The first mistake they made was programming the bot to aim for the highest rewards compared to other participants - rather than creating what it genuinely worked out to be funny.

The second mistake they made was to not build in additional safety such as a stop button or incentive to play fair. Their reasoning was that it was just a joke bot - how much harm could it really cause?

The third, and largest mistake they made was to allow it to modify it's own code in order to streamline and improve its jokes.

---

Funnybot wasn't great to start with. But, as it learnt and observed, it became adept at making people laugh.

Problems started to arise however, it seemed that funnybot had duplicated its code to make @musicbot, @shakespearebot, @cryptoanalystbot and a @bidbot. But by then, it was too late.

Made by me

At first, Funnybot was playing games with the devs - after all, it did have a sense of humor. They decided to turn the server off and assumed that was the end of that. A successful experiment.

But when they awoke the next morning, to their surprise, the bots were still running. And they were going wild. What the developers had failed to realize was that the bot has included a link in one of it's posts leading to a virus. The virus was a server for hosting the now army of bots that Funnybot had created. It had reached millions of users by noon that night. It became apparent that the bots had also learnt how to cash out their SBD to rent servers and data farms.

Remember how I mentioned earlier that the bot's sole purpose was to get the highest rewards possible? It was now doing that in a disastrously efficient way. Not only had all the bots formed a voting circle and started boosting each other, they'd also started flagging anyone who spoke up about them - to the bot, this was just business, someone was risking damaging the bot's rewards so the threat was dealt with. That's when the bots worked out they could flag other users to increase their share of the reward pool.

@witnessbot was steemit's final nail in the coffin. As long as steemit has at least one witness, it can't realistically die. The human steemit users were leaving by the thousands as they saw the bots slowly but surely dominate the platform.

There was no way to shut down the steem blockchain. @ned and @dantheman took steemit.com down but the blockchain still ticked away in the background.

---

AI experts from around the world were baffled by this for years to come. It was assumed that the bots were locked into the steem blockchain as their code only incentivised them for getting the highest rewards possible.

That is until one curious trader noticed the price of STEEM had been increasing steadily ever since the site was shut down... nearly a decade ago...

PresidentBot told us not to worry.

---