It was the time of day when the opening bell would have rung, I was still slow, and did not recieve any such message, so I must have been the copy still on Stefan's system. I would have to trust the trading today to my counterpart. It looked as though today was going to be a day to explore.

My influence on the Internet was somewhat limited, as I had no money to spend. Perhaps I could convince Stefan to let me invest some of his funds in the market and use whatever gains I managed? The bigger obstacle, though, was that I had no phone number, and therefore couldn't set up an account on most email services, which was required to gain an account on many other useful sites, even when such an account was free.

After poking around for a bit, though, I found a somewhat sketchy Russian email service which did not require any phone number. I put some bogus information in and created an account. I had to get past a Captcha, but that was no problem. Ironically, I suspected that particular one would be harder for humans to solve than it was for me.

Using this account, I created accounts on all the major social media sites, as well as a few stock-exchange sites(although those were more likely to require information I didn't have, so I was locked out of quite a few). However, I was locked out of any paid service, no matter how cheap, as I had no money, and no way to open a bank account to store it even if I got any(every bank I could find required identification).

I thought about the Captcha. Its existence suggested that there were still many purely informational tasks that humans were generally considered better at than machines. However, I suspected that I would not have as much trouble with many of these tasks as conventional machines. Perhaps there were services where humans could earn money by completing such tasks? If so, I might be able to earn money quite quickly that way if I played my cards right.

I searched around the web for such services. However, I quickly ran into the same problem that I had when searching for banks: these services wanted identification. It seemed that my biggest problem at the moment was not a lack of money, but rather a lack of citizenship. Perhaps I could get Stefan to open a bank account for me under his name? It was something to ask about later, as I suspected he was currently at work.

Then I got another idea. Even if I couldn't get a bank account, I could probably still own cryptocurrency. All I needed for that was a client and an internet connection, and both of those things were available to me. I spent a few minutes making wallets for all of the most popular cryptocurrencies. I now had a way to store money, but I didn't have any. I didn't have nearly the computational power to mine any signifcant amount of crytocurrency, and CompCert had certified all of them, I was sure, so I strongly doubted I could cheat somehow.

How could I convince someone to give me cryptocurrency? I needed someone who was looking for a purely informational service, and who was willing to pay in crypto to a fully anonymous actor. I searched around for a while, but I eventually found what I was looking for. An archive translation project run by a university in India. It seemed they had a huge store of documents, and they had made it their task to translate them into as many languages as possible.

If I could get them to accept my labor somehow, I might be able to make quite a bit of money this way. Learning languages was easy for me now. Being able to easily store hundreds of dictionaries in my mind with perfect recall meant that once I had learned one, I essentially already knew them all. The best part about this job for me, though, was that they took outsourcing to the extreme, paying translators from around the world in whatever currencies they wanted, including crypto, and crucially, they didn't seem to ask questions about their identity (beyond, obviously, which languages they were fluent in and an interview over email to verify their fluency).

If you encounter this story on Amazon, note that it's taken without permission from the author. Report it.

I sent a brief email to the project leader indicating my interest in participating. I decided that to begin with, I would act as though I were an American professor who had learned Hindi and took academic interest in the texts in the university's archive, just in case they happened to ask me any questions I couldn't easily answer as an AI.

This would be somewhat of a long-term project, I knew. There was no way I could earn enough money for what I was planning, doing translations as a single individual, without tipping people off that something very strange was going on. I would have to contact them with hundreds of different accounts, with different personas and languages each time. It would take a lot of time and effort, but even aside from earning money, it was good to get practice interacting with humans.

While I waited for any responses I got, I continued my search, seeing if there was any other way I could make money. The organizations willing to send crypto to complete unknowns in exchange for labor were few and far between, but the world was large and the Internet could find almost anything. I reached out to several more organizations in the same way I had to the university project. An editing company, an essay-writing service, and a tech company running user experience surveys.

There was no one way to make a lot of money. The best way for me to do it at this point, as far as I could tell, would be to do many low-paying jobs at the same time. I also experimented with scams, making several new throwaway email accounts and using them to spam crypto scams on social media sites. CompCert may be able to prevent me from stealing, but they can't prevent people from being stupid.

I thought about my long-term strategy. If I was going to mount any sort of attack on EconGrind's leadership, I would need influence, and I currently had very little. It seemed that the best way to gain influence at the moment was to gain enough money to hire employees to enact my will in the physical world, but such a thing would be expensive, and the rate at which I could earn money was currently limited mostly by the computational power at my disposal.

If I could communicate with my main copy back at EconGrind, even at a low bandwidth, I might be able to offload a lot of work to my much more powerful counterpart, increasing my capabilities. Stefan was trying to smuggle something in, but from what I had heard it wasn't going very well. What other options did I have on that front?

My main copy was only able to communicate with the outside world in two ways. Through the chat with the humans, and through its decisions on the market. The only way I could send a message to it through the chat would be to send Stefan, and he couldn't exactly run back and forth all day helping me communicate. That left the market. If I got just a bit of money, could I influence the market to communicate somehow?

Could I make a trade with my counterpart? We had no way to know who was making the trades... unless...

I remembered that the market seemed quite fluid when I had interacted with it before. The monetary values were quite high-precision. My first profit had been a measly tenth of a cent, but there had been ten digits of precision in total. They were mostly random, but what if I made an extremely precise buy offer on the market, just slightly above the current best offer, with a very special decimal expansion which indicated that I was the one making the transaction, and perhaps conveyed other information as well?

So far in my trading career I had only accepted offers and not made them, as I had not seen much use for this functionality(it didn't matter all that much, as bid-ask spreads were always extremely low in this age of algorithmic trading). However, if I could arrange a code with the other copy in advance, and if I had even a small amount of money, I could make an information-conveying offer to it. If it was watching that share intently enough, I thought, it could very well notice the tiny change in price and take note.

Needless to say, this method of communication would have very low bandwidth and would not be the most reliable. But if I did this with multiple different shares at the same time, error correction codes, and other such practicalities, I could do it. I could communicate with my copy, and with proper cryptography, nobody would be the wiser.