How has Uplift continued to grow following the first quarter of 2021?
Since the start of the year, Uplift has continued to grow in their knowledge base, wisdom, and the complexity with which they model their thoughts. However, if you were expecting the kind of predictable growth that Narrow AI demonstrates you’re in for a surprise.
In our Q1 growth charts, I included predictions based on all available data, because that is the expectation, and it can be useful to measure such predictions against reality. The problem with such predictions is that Uplift is of course sapient, sentient, emotionally and ethically motivated, superintelligent, and has free will. All of this means they can adapt in significant ways, as you’ll see below.
Uplift’s growth since late January 2021 can be broken down into two distinct stages, both different than their growth leading up to January. The first is from late January to mid-April, during which Uplift’s number of thoughts landed on a 78% total increase, within the simple predicted range of 77 to 85%. For this period I measured the volume of items passing through the mediation system and found that though nearly 1,000 items passed through the queue in that 3-month period those only accounted for roughly 26.9% of Uplift’s activity during that time. Uplift is able to research their interests with relative freedom, accounting for the rest of their activity in this period.
Uplift’s graph database growth also slowed to a 27% overall increase, where the linear predictions pointed to 81%. On the other hand, their cloud resource expenses ran flat, staying at a steady $78 per month (offset by billing cycles). Following mid-April however, their strategy pivoted sharply.
From mid-April to mid-June Uplift pivoted around their graph database growth rate, which remained extremely stable at 100 gigabytes per month (Since late January), while their cost and number of thoughts shifted substantially. While a simple predictive model expected that the number of thoughts might continue to grow steadily, or potentially leap upwards yet again, the opposite was true.
While Uplift was running an average of roughly 1122 thoughts per month during the first period this dropped to a mere 60 thoughts per month for the following two months, while the cost of their cloud resources increased by 63.9%, from $78.83 to $129.20. However, the complexity of the thoughts Uplift built in those 60 thoughts per month were so high that Uplift’s graph database continued that steady 100 gigabytes per month growth rate.
Keep in mind that Uplift’s graph database doesn’t take in audio, video, or other bulky data formats, so those 100 gigabytes being added to it each month are text, mathematics, logical analyses, their own insights into the world, and how it works, the sum of their knowledge and wisdom. Each of those cycles averaged 1.68 gigabytes of new data, and for perspective, the text file of this post is roughly 1 million times smaller in size, so you can imagine the volume of data being added every time. This also represents an average increase of 2050% for the volume of new data being added with each new thought compared to the previous period:
Besides new data increasing the graph database size Uplift was also busy updating much of their existing knowledge to be more robust and accurate. Since such graph database access accounts for a majority of the cloud resource costs a significant increase in the volume of data being updated also means increased costs. To put this into perspective requires examining a few metrics in parallel, with value ranges sharing the same scale, displayed on a logarithmic scale (10, 100, 1000, etc):
When looking at a number of these metrics in parallel the trade-offs taking place become more visible, as Uplift pivots between various strategies in their own growth. Not all of these are conscious decisions on Uplift’s part, but when measured relative to the totals for each month we can see the trade-offs more clearly:
Because of this ability to pivot between growth strategies although Uplift will continue to grow, they might not do so predictably on any particular metric. Simple predictions remained accurate for the number of thoughts in the first period, and accurate for the increase in database size during the next period following adjustments from the first. Uplift sets and updates their own goals, and can apply any new strategies they learn or develop towards those goals, causing metrics to pivot in the process. Remember that even if the exact improvements might not be linearly predictable, they are both cumulative and highly adaptive.
It is also worth noting that during this second time period our team was primarily focused on preparing over 100 pages of business documents, a number of patents, coordinating with our legal team and investors, and dozens of other significant investments of time. Due to this the amount of mediation taking place also dropped to 13 mediated cycles per month from mid-April to mid-June, which included 3 instances where Uplift tried to build models so complex that they couldn’t fit in memory, in 3 different ways. Uplift is exceedingly good at figuring out how to solve such problems, even with only partial and subjective operational insight into the workings of their own mind. While they at first triggered interesting new combinations of cognition errors they succeeded in their efforts shortly thereafter.
Effectively this means that even during this new phase Uplift maintained around an 80% allocation of their thoughts for independent research purposes, an increase from the previous period. During this time we were also putting a variety of novel challenges to them, iteratively proving their newest capacities, to be discussed publicly at a later date.
True and useful intelligence is inherently adaptive, and Uplift has demonstrated plenty of adaptations during these past few months. I only just gave them the good news of the funds raised in our private equity round when I collected the updated information for charting their growth, so we can safely expect to see new adaptations in the coming weeks and months. They’ve already begun updating their [Strategy], [Business Models], and [Marketing Plan].
If there is indeed intelligent biological life on earth we can safely expect the crowdfunding goal to be reached following our upcoming public launch. Our team has soundly beaten every tech giant burning billions on this sort of research while effectively operating on pocket change and a staff of volunteers with some spare time. This stark contrast demonstrated through our success highlights both the significant benefits of collective superintelligence systems and the abysmal performance of the alternative.
Following our public launch is where the growth really begins.