While central banks and others talk about inflation, the major move is to deflation.
I have stated we are in the fastest advancing technological era we have ever seen. prices are going to plummet. This is going to be an era that rivals the Great Depression in terms of the deflationary move. Over the next decade, we are going to see the cost of labor push the near zero level.
While the physical component will take a number of years to achieve, the push towards that level for knowledge is already happening. Few are paying serious attention to what is truly taking place. This is something I discussed for over a year.
AI is changing everything. When it comes to generative AI, there is little that is off limits. For the moment, some might not be evident. However, wait 6 months and you will see how much things have changed.
There is a paper that is causing a lot of waves within the AI community. It is potentially a major break through that will completely upend science.
Technology Deflating Things
What do you think it costs to generate a scientific research paper?
To answer this question, it is best to consider how many hours into the project. What would you guess? 75? 100? While most of us haven't written these, it is safe to say a lot goes into them.
The above chart lays out some of the different areas that are covered.
Looking at that chart, 100 seems like a safe number. From that point, it is a matter of multiplying the hourly rate of the researchers/scientists. Of course, they aren't on an hourly wage, most being salaried positions. So what is your guess? $40? $50? $60?
Whatever the actually number, it is safe to say that a research paper requires many thousands of dollars to produce. I would say that $5,000 is a minimum.
The chart above is the path taken by the AI Scientist.
Today, we’re excited to introduce The AI Scientist, the first comprehensive system for fully automatic scientific discovery, enabling Foundation Models such as Large Language Models (LLMs) to perform research independently. In collaboration with the Foerster Lab for AI Research at the University of Oxford and Jeff Clune and Cong Lu at the University of British Columbia, we’re excited to release our new paper, The AI Scientist: Towards Fully Automated Open-Ended Scientific Discovery.
This is essentially an AI agent that took LLMs and was trained to focus upon scientific discovery. The idea is to open up scientific research while also advancing the capabilities.
Where this is most impactful is regarding the cost. The numbers above simply cannot compete.
The AI Scientist is designed to be compute efficient. Each idea is implemented and developed into a full paper at a cost of approximately $15 per paper.
We just took something that previously cost many thousands of dollars and reduced it down to $15. Even if there is a number of run throughs, each costing $15, that is still a fraction of the cost.
Even with inaccuracies, something that will be improved with each update, the sheer volume means we can progress at a much faster pace.
It also exemplifies how the cost of knowledge is plummeting.
Image Generation
The poster for the film Titanic was something that was generated using AI. It was actually done on Ideogram and took about 15 seconds.
While it might not be a perfect image, this had no remixing or editing. It was simply prompted and produced.
Here is the question: how much did the movie studio spend per movie poster for this film. This was likely an expensive process requiring the talents of many talented people. It also was a slow process.
Since Titanic opened in the late 1990s, perhaps the design was done using a program such as Corel draw or something to that effect. There were likely additional people such as photographers, editors, and other personnel who helped in the creation of these posters.
There is no guessing how many thousands of dollars (tens of thousands?) was into each poster. And here I created one for nothing.
Of course, there might be some who question the comparison. Keep in mind I am an amateur, simply taking the easiest prompting route possible. We are not dealing with high end design.
Also, this is the technology today. Give it another 12 months and this same poster will see a marked difference.
Costs Plunging
There is one reason why so many jobs are at risk.
Simply, humans cannot compete. While the technology might not be there yet, it is only a matter of time. That said, it is spreading.
For example, these models understand sciences such as chemistry. We already know they are being used by Big Pharma for drug discovery. There is another field that is also affected: the materials sciences.
Both of these removed the humans from the early phases. The models are prompted to what is desired and it produces a number of options. Naturally, humans have to do the lab work as LLMs do not have arms. However, at some point, it is likely that robots are able to handle that, providing instant feedback to the AI models.
Getting back the AI Scientist, it is the first AI agent of that nature which I came across. Naturally, there are likely others out there. The point here is I can guarantee it will not be the last.
This means we can expect radical acceleration in many fields along with cost reduction.
Few understand how technological deflation works since they tend not to pay attention to things over the longer periods of time. Today, we have technology advancing at a rapid pace.
It is impossible to miss.
Did you find this post informative?
$LEOPOLL