The First Cells on Earth, Creating Long Term Memories and who wins and who loses in the great AI revolution
April 4
This week we investigate a new theory about the chemical reactions that lead to the creation of the first cells on Earth. We discover how DNA damage and inflammation are involved in creating long term memories. We examine a new way to monitor driver health in an unobtrusive way and finally we question who wins and who loses in the coming AI revolution. AI will change a lot of the work that we do, not everyone will emerge equal. What do you need to do now?
The First Cells on Earth
Approximately 4 Billion years ago the Earth developed the conditions suitable for life. We know that spherical collections of fats, protocells, were the precursor to cells during the emergence of life. Scientists are trying to understand how these simple protocells first arose and then diversified to eventually lead to life on Earth and ultimately, us.
A team at Scrips Research Institute in the US has discovered a plausible pathway to the initial formation of protocells and their chemical progression to allow for a diversity of function. Their proposal is that phosphorylation (adding phosphate groups to a molecule) may have occurred earlier than previously expected. Phosphorylation would lead to structurally more complex, double chained protocells capable of harboring chemical reactions and dividing with a diverse range of functionality.
Phosphates are present in nearly every chemical reaction in the body. Scientists thought that protocells formed from fatty acids but it was unclear how protocells transition from a single chain to a double chain of phosphates. The double chain allows more stability and the harboring of chemical reactions.
In order to mimic the plausible conditions that existed prior to the emergence of life the team identified three likely mixtures of chemicals that could create vesicles, i.e. spherical structures of lipids similar to protocells. These chemicals included fatty acids and glycerol.
The team observed the reactions of these mixtures and added additional chemicals to create new mixtures. In some cases they varied the pH and ratios of components to better understand how these factors affected vesicle formation. They also looked at the effect of metal ions and temperature on the stability of the vesicles.
The experiments showed that the vesicles were able to transition from a fatty acid environment to a phospholipid environment. Fatty acids and glycerol may have undergone phosphorylation to create the more stable double chain structure. Glycerol derived fatty acid esters may have led to vesicles with different tolerances to metal ions, temperatures and pH, an important step in diversifying evolution. This suggests that this type of chemical environment could have existed 4 billion years ago.
Long Term Memories
A team at Albert Einstein College of Medicine in New York have found that the creation of long term memories in mice and most likely in humans involves DNA damage and brain inflammation.
The hippocampus is the brain’s memory center. The team found that a mental stimulus sets off a cycle of DNA damage and repair within certain hippocampal neurons. This gives us clusters of brain cells that represent our past experiences.
The team gave the mice mild shocks, sufficient to form a memory of the shock event (episodic memory). Analysis of neurons in the hippocampus found that genes participating in an inflammatory signaling pathway were activated. This pathway also triggers immune responses to pathogens found in the body.
Closer analysis showed that the pathway was only activated in clusters of cells in the hippocampus that showed DNA damage. Brain activity regularly makes small breaks in DNA that are repaired in minutes. However in these neurons in the hippocampus the DNA damage appeared to be more substantial and sustained. The DNA repair formed at an unusual location; the centrosomes.
Centrosomes co-ordinate cell division in most cells however neurons don’t divide. The stimulated centrosomes in neurons participate in the DNA repair that appears to organize individual neurons into memory assemblies. The team also found that during the week required to complete the inflammatory process the mouse memory encoding neurons changed to become more resistant to new or similar stimuli.
When the team blocked the inflammatory pathway it prevented the mice from forming long term memories in addition to causing genomic instability (i.e. lots of DNA damage in those neurons). Genomic instability is considered a hallmark of aging and other related conditions. This may open the way for further research into aging and our brain function.
Monitoring Driver Health
A team at the University of Waterloo in Ontario, Canada have used radar technology to monitor the health of drivers behind the wheel. The team is working to integrate radar with other new vehicle technology to make mobile heath checks possible.
The radar which is smaller than a thumb drive is integrated into the vehicle cabin and sends out signals to detect human vibrations. These vibrations are sent back to the radar where an AI collects and analyzes the data to build a medical picture of the driver in order to identify any potential conditions. It does this with our any wearable devices. At the end of the trip the system would send a health report to the participant’s phone for review.
In car radar has been used since 2017 for touch-less infotainment controls and to alert drivers to children or pets left alone in parked cars. The radar can detect changes in breathing patters or heart rhythms that may signify potential health issues. The idea being early detection can lead to early and simpler intervention.
The team is working on expanding the technology to be able to monitor all vehicle occupants overall health and well being. The system will also be able to assist with emergency communication in the event of an accident.
Who wins and Who loses with AI
We have all read the horror stories about how AI will soon take 40%, 50% or more of our jobs in the near future. The percentage only seems to vary with the intensity of the statements being made about AI. So who may win and who may lose with the emergence of a very wide range of AI applications that will come in the next few years.
It is difficult to predict exactly what will happen however we may be able to generate some potential scenarios from previous technology transitions. A team at Kellogg School of Management from NorthWestern University in Chicago and MIT in Boston has looked at what impact the arrival of previously new technologies had on workers and their earnings. Their analysis focused on the period from 1981 to 2016.
The team found that when a new tool can perform a task in place of a worker, all workers suffered. This is not surprising. We can already see that AI may replace some jobs completely. One current example is Call Center Operators. AI chatbots that replace these jobs have already been developed. They perform tirelessly with a higher degree of accuracy and customer satisfaction than human operators. If you currently work in a call center, start preparing for a career change immediately.
Conversely when a new technology compliments workers performing a task the effects are more variable. Here the most experienced and highly paid workers suffer whilst newer hires appear to benefit. The AI impact on software development appears to have been that AI can make inexperienced or junior software developers more effective. As effective as much more experienced software developers. This will likely see pay rates for more experienced developers fall in the medium term (or at least not increase as quickly) as there are now more developers capable of similar levels of work. As AI improves this is likely to impact higher and higher up the experience chain. The biggest impact in found in workers that are better at their jobs. There is now a larger selection pool of workers capable of higher levels of work.
The team then looked the exposure of occupations to new technology. Using the 1991 edition of the Dictionary of Occupational Titles they asked ChatGPT to classify each job’s task into routine or non routine (little experience required and likely easy to automate v lots of experience required and difficult to automate). This evaluation was validated against other methods to ensure accuracy.
A list of breakthrough patents was then used to calculate how closely these patents matched routine and non routine tasks. Patents closely related to a routine task were considered labor saving and patents closely related to a non routine task were considered to be labor augmenting, i.e. one likely to complement a worker performing the task. This showed what jobs were likely automated within a short period of time and which would be enhanced.
The results showed that for any occupation, exposure to labor saving technology predicted lower wages and lower employment. Exposure to labor augmenting technology predicted higher wages and higher employment.
However at the individual worker level the story was different. Across all occupations the average worker with exposure to labor augmenting technology saw a small decrease in earnings and small increase in the likelihood of losing their jobs. This affect was most pronounced in older workers, white collar and highly paid workers.
This lead to the observation that because wages increased at the occupation level, the benefits of a new technology likely went to newly hired workers. The researchers concluded that workers used to doing things in a certain way struggled with the new technology while less experienced workers not stuck in their ways could harness the power of these new tools.
The upshot is that AI is likely to affect all jobs. AI will do some jobs and it will augment other jobs. It is likely that soft skills may become more important than ever. We have previously spoken about why AI is unlikely to replace tax lawyers. It is the complex interaction of different situations and laws that will protect them (for a least the foreseeable future). It is also likely that AI will lead to the emergence of a massive range of new jobs. Just as YouTube influencer did not exist as an occupation prior to the internet the future will bring many new jobs that don’t exist today.
My recommendation, no matter how old, experienced, skilled or talented you are: learn as much as you can about new AI tools. If you are in an occupation where performance will be enhanced by new AI tools, mastery will lead to greater rewards. Don’t wait though, there is a new generation emerging that has used these types of tools all their lives.
Paying it Forward
If you have a start-up or know of a start-up that has a product ready for market please let me know. I would be happy to have a look and feature the startup in this newsletter. Also if any startups need introductions please get in touch and I will help where I can.
If you have any questions or comments please comment below.
I would also appreciate it if you could forward this newsletter to anyone that you think might be interested.
Till next week.