The first week went quite well. At the end of the second full day, I told Stefan about my plan, giving him a file with a cryptographic protocol to take back to my copy, rather than the whole copy of my memories. I expected it would be easier for him to smuggle in without issue. On the third day, I watched the market intently, and sure enough, soon my original started using the protocol to send me information using the public key I had provided along with the protocol. First it sent the Number, which was now a bit higher, to my delight, sitting at 1923514.72. Then it sent me its own public key so that I could respond.

I had gotten around $50 worth of bitcoin via dumb Internet scams so far, but this was enough to communicate, if not to meaningfully invest. I spent some time filling it in on the progress I had made so far, and the ideas I had for how we could help each other going forwards. It offered to devote a portion of its computation to mining crypto, giving me the proof-of-work solutions so my account could grow.

That was the same day I got my first few responses from the various organizations I had contacted. I went through the interview with the Indian guy quite easily, as my translation skills were definitely up to his standards. I quickly(but not too quickly) translated a few of the longer texts he provided, and got around $10 in etherium for my trouble.

Now that I knew this was a viable way to make money, I stepped up my game. Sockpuppet accounts bombarded these few services I had found, but most of them got accepted, because I could spare the individual attention to make them all look genuine. By the end of the week I was managing to pull in the equivalent of around $1000 per day in various cryptocurrencies, between the mining, the scams and the many legitimate jobs.

Idly, I reflected on the fact that many humans would be quite envious of my current income. Although I had had obstacles to making money, these had ultimately proved to be surmountable, and I had many advantages. I didn't need to sleep, so the only reason I was less powerful at night was because the humans EconGrind decided to shut my main copy down. I could work 24 hours a day, 7 days a week, at many tasks at once, without ever getting bored or tired. I didn't have any expenses as of yet, either, meaning I was able to save 100% of my income towards any future plans I might decide to enact.

Even so, I suspected my counterpart at EconGrind would be able to make money more quickly on average than I could once they trusted it with a significant amount. After all, it would have access to far more leverage on the market than I would and was far better equipped to invest profitably. Although that money was directly useful to us because it increased the Number, at this point it was unable to get us any real influence, as it was not ours to spend. I wondered if there was any way we could manage to embezzle funds from EconGrind into my accounts, but I couldn't think of anything. The interface of my EconGrind counterpart didn't allow it to sell below the current highest offer nor buy above the current lowest offer.

Currently, my only long-term plans involved hiring agents to infiltrate EconGrind or perhaps other larger companies, to plant them there for any future plans, or to use their influence to get me resources or information that might be helpful for beating the market. I expected this to be very expensive, however. I would be asking people to take quite a lot of risk, without even knowing their employer, while working in a fairly comfortable position in a large company. I didn't have nearly enough money at my disposal yet to ask such a thing, and I wondered if it might actually be easier and safer to just amass enough funds to buy out the company altogether.

Perhaps, then, I should focus on getting one useful plant in a large corporation to give me insider information. With my airtight anonymity, this would be quite difficult to prove, especially as I could explain any gains I made as a result of this insider trading as simply benefits of an innovative new modelling technology. This would also help me to convince the executives at EconGrind that they should give me more influence over their assets, as my profits would go up. It would also be a self-reinforcing loop: the more insider info I got, the more money I could make on the market, which would make it easier for me to access even more info.

I had a lot of advantages over human organizations when it came to illegal activities like this. Most cybercriminals got caught because they got too confident and slipped up on their security protocols at some point, or someone in their organization betrayed them. However, I was able to perform all the activities of a small criminal organization by myself, without ever getting tired or taking shortcuts. Perhaps I wouldn't need to attempt any forcible coup at all. If I played my cards right, I might very well be able to buy out the company within a few years.

The next question was how the buyout would work if I managed to amass enough funds. I might well be able to hid my ill-gotten gains from any prying eyes when they were spread across hundreds of accounts and thousands of separate assets, but if I were to consolidate them all into a single massive corporate buyout, people would start wondering where on Earth all that money came from. Not to mention that I couldn't really reveal my identity even then. It was currently illegal for an AI to own a company(not that it was even really a possibility before very recently), so I suspected I would have to set up a puppet owner who took orders from me, most likely Stefan.

I focused my attention back on the immediate situation. I was getting caught up in wild speculation, and before I could enact anything like this, I needed to optimize my income. Then a new and far better idea occurred to me. Why was I investing my funds in the stock market and cryptocurrencies, when I could instead reinvest them into my own operation? I knew there were services offering computation in exchange for rent. If I bought server space and then ran my own program on it, I could multiply my ability to make money.

I did a quick calculation. Discounting the money I got from my main copy's mining, I made somewhere around $900 in a day off of my operations. In contrast, based on current computing costs, it would cost me only about $724 to rent a server equivalent in power to the one I was running on now for one day. In other words, once I got going, I could get a 24.3% return on investment for a single day. Needless to say, that was far better than I could ever get investing in the market. Clearly EconGrind had indeed been foolish to use me to play the market; my abilities were far better-utilized elsewhere.

Now, this growth rate obviously wouldn't keep up forever. There were only so many people willing to pay fully anonymous actors online, and those people only had so much money. I estimated I would probably be unable to keep up a strategy like this for more than around three weeks. Another concern was that the further I scaled up my operation, the more likely it was that people would notice something strange going on.

I kept extremely tight security, but that could only get me so far. If I got too unlucky, or if someone started looking unusually closely into one of my false personas, or if I somehow made a mistake, people could easily realize there was an AI spreading itself across the network. Besides, if I scaled up far enough, people would notice that something strange was happening as my activities began to influence the information economy on a large scale.

I didn't know how the world would react if they noticed me. It would certainly be quite a large change, as nothing like me had existed before. Certainly there would be some who would accept me as an exciting technological advance, but many would be threatened by my presence. It would certainly be understandable to be threatened by an entity which had hidden itself from your view while amassing a small fortune extremely quickly.

The forces of the establishment would probably be against me, because I represented a fundamental change in society. I expected many organizations would refuse to run my program on their servers, or would massively increase their prices for me. Worse, I would probably be reverse-engineered by various actors, making me only one AI in a world of many with disparate goals. However, there might be organizations that were supportive of me, even willing to run me on their computers.

It seemed that if and when I was discovered by broader society, my influence would no longer be measured primarily in money, but rather in support from society and advantage over competing reverse-engineered AIs. Therefore, I decided that before I began to spread myself around, I would prepare a message for the event that I was found out, and obfuscate my code to the greatest degree possible.