A cautious approach to AI: the new bubble
Everyone's chasing the shiny thing again. Wait! it can sting you.
Bitcoin Block 793,392
000000000000000000030b3e6f2eb25c3dd0ba88ceca750ba770bc54098a491b
Economics is a simple subject on paper. It is concerned with the decisions humans make in the face of scarcity. But the problem is humans are the center stage of economics and humans are unpredictable creatures. Many factors come into play instead of the logical and statistical decisions that occur on paper. A logical solution that seems plausible and is bound to occur meets a proclivity that is a result of human interest and down goes your logic in the drain.
Humans fundamentally operate in cycles. Those cycles aren’t supposed to repeat but they sure do rhyme and the present economic conditions surely provide evidence of that. Where are those who considered real estate the supreme form of investment? Where are those who touted SPACS as a new way to invest in budding companies and gain high returns? Where are those who made the metaverse look like our final destination? Well, they faced a loss of $40 billion trying. Where are those who boasted the potential of bringing everything on the blockchain? Well, times have changed and the bubbles have been deflated. The steam has run out and this engine has come to a halt. But that doesn’t mean human progress stops.
The real and identifiable goal isn’t monetary prowess. Making money is a game. Many play it and few come to encapsulate all of its rules and are able to achieve a breakthrough. So measuring success in the form of profits is a crippled metric. Generating wealth must be the main objective and by that, it means building something that people want to actually use. The new shiny thing AI has sprung up to be everyone’s darling. with new AI tools coming every day and companies getting on the bandwagon. Adobe with their new Firefly, Snapchat with their new AI bot, and countless other AI-integrated tools that aim to ease the workflow of developers, graphic designers, programmers, and writers to name a few areas. The big daddy here is Meta jumping on the AI train after ditching its dream of putting VR glasses on every head. The recent Shutterstock deal with OpenAI has broadened the scope of laws for generative art and artists are getting trampled.
Now people obviously go towards the thing that catches their lazy brain’s attention. Generative art and text are something that has caught the public’s eye because of the breakthrough it provides to normal activities.
Contrasts in the AI image
Elon Musk and Sam Altman, the man behind OpenAI are cognizant of the threats such a technology provides to humanity. They deem activities that are distinct to humans such as creativity as being done better by machines. You give the machine pointers and it will generate something beyond the scope of humans. Elon is quite vocal about the threats AI poses saying blatantly in an interview with Tucker Carlson in April that AI has a "potential of civilization destruction". Now, obviously, there are optimists and pessimists and both of them present their proclivity in upright facts. Complete recognition of one side and dismissing the other is fool’s play. When it comes to Elon, considering his dominance-seeking personality, he might have some grudges against OpenAI for not making him the head of operations and leaving the company in 2018. This might have something to do with him directing some of the engineers at Tesla’s AI operations but let’s just call it a conflict of interest.
On the other side, the subtle one, Sam Altman has not been that vocal until recently. He has a knack for gaining power and his steps might be remembered in history as altruistic or devastating for the industry because of the pivotal role he plays. He wants to involve the government and regulatory bodies to assess and mitigate risks. Now, going for centralization might not be the sound approach because of data capture and a threat to privacy but these are his views. Maybe this stance has led him to make OpenAI a for-profit and conceal its work behind closed doors. The company that has believed in democratizing access to AI has fallen a prey to lack of investments and capital requirements. Microsoft has to hold hands to keep it afloat and Sam won’t go down that easily.
Competition breeds creativity
Although the development of AI started in the 1950s it is only in recent times that the public has met eyes with it and is starting to realize its potential. The initial runs on AI have been from the British logician and computer pioneer Alan Turing with his highly commemorated works on computer intelligence and his famous Turing test to differentiate a human from a machine. To quote one of his early lectures:
“What we want is a machine that can learn from experience,” and that the “possibility of letting the machine alter its own instructions provides the mechanism for this.”
The early attempts to make a computer intelligent have been focused on making it able to solve complex problems and find creative solutions to the logic involved in games like chess and checker. IBM has been a major player in developing intelligent systems in 1997, Deep Blue, a chess computer built by IBM, beat the reigning world champion, Garry Kasparov, in a six-game match. This is a breakthrough that later on has led to further development in the space and creates a foundation for the progress we see today.
The real problem with developing computer intelligence is reinforced learning. You see a baby go from the crawling stage to the walking stage, that is all reinforced learning such that the child has the ability to decipher the stimuli he receives from the environment. The problem with machines is that we don’t have time to train them like humans. Why would you want to raise a machine child and tell him step by step? He must have the ability to creatively find a solution. A solution to that can be creating a perfect simulation for all the possible situations but then again that will be a cumbersome process. OpenAI has the lead when it develops the robot hand that uses domain randomization. Basically, feeding data back to the system multiple times and gathering it through multiple inputs on the cloud to better train the model ending up with 100 years of experience.
The making of ChatGPT and releasing it to the public followed by its immense rate of adoption has waken up competitors to be more active in this field. ChatGPT has made a ground-breaking entry into the AI space and has opened the door for more investment and funding in the realm of machine learning models. The likes of companies such as Meta and Google have accelerated their machine learning programs and many new startups have emerged. Now we can bring up the failed metaverse investment Meta has done and now jumping on the AI bandwagon. But humans generically make mistakes and a company consists of humans to the core so mistakes are expected. Learning from them is what counts. We can regard this as another bubble, as well as buzzwords like deep learning, AI, machine learning, and language models, are being thrown around and no one knows the clear distinction. VC money is piling in every company with an AI tag to get recognition. Google has gone wild in the unraveling of its Ai ventures by mentioning the word AI like it were the center point. Is rushing to roll out Bard and such a massive development on Palm2 a well-planned approach or is it just an attempt to get ahead of its competitors?
AI, the right way
One thing is for certain, when a mania phase of a cycle starts, money is thrown around and research starts making progress and gains pace. Although the people getting in without knowledge get wrecked the field, if it is logical, benefits in the end. The researchers, engineers, and provocative thinkers get a ground for their voice and humanity benefits as a whole.
Most importantly, a cautious approach must be adopted. Just because people find it attractive does not mean it is beneficial to pursue. One thing is for certain, computers have gained immense potential over the years and human intelligence can be enhanced by integrating them together rather the quotidian superseding talk. Half a second is an eternity for computers to process information whereas humans can only gain a sense of consciousness at that time. Such immense power demands a responsible attitude toward its handling and effective use.
One thing that is necessary for harnessing the potential of mimicking human intelligence in machines which is decentralization. Everyone should have their own personal AI assistants to better cater to their own human ingenuity. It is like a self-help book but 2.0. Now that self-help book talks to you and gives you self-curated advice. In addition, data preservation and privacy are crucial in making that level of intelligence capable to serve every individual justly.
A spectacular example is the use of AI to point out mental illnesses at Cincinnati Children’s Hospital. The data acquired from the patients is used to decipher information about them that he may not be aware of himself. This is through analyzing the specific behaviors beforehand and deducing the possible defect. Suicide actions have been prevented because of this! This is because data is being constantly collected and stored in data centers. Then that data is fed into a machine-learning model and the outcome is indicated. Soon, health care recommendations personalized to each body might be possible.
AI and jobs
Now the burning question that arises in every mind is whether you should start packing your bags and find a good place to die as you’ll be replaced, betrayed, and left sunken by this seemingly benevolent innovation. Well, the answer is complicated. If you consider the humans that are present in this state, then yes all of you will be replaced. But stop for a second and remember the biggest advantage of being human: flexibility and creativity. None can beat us at that.
The first jobs to be replaced are the ones requiring the least creativity and are more repetitive in their operations. That is, they have the least amount of skill and technique involved. If things can be taught to an AI and it can produce the same work in less time then the workers are at risk of losing their importance. However what we fail to view is that at a broader scale, when projected out, this creates a more creative, resilient, and innovative human colossal. The word AI has been thrown around much in recent times which has left us wondering about this question. But in reality, developing a machine powerful enough to mimic the brain’s neuronal power has been a point of interest since the 1950s. Hence this will be a gradual process where inefficiency will be filtered out and the ones left will be unique in their attributes.
Brain-machine interaction
Instead of focusing on who will be replaced, the better question to ask is how machines will be integrated with human activities such that they can walk hand in hand producing something magnificent. Finding the answer to that and developing on that is what must be left after the clouds over this AI bubble are cleared. New ideas will be tested, and money lost but in the end what will be left are new questions, new ideologies, and a new path forward. Those are brain-machine interfaces that will revolutionize how humans function and their cognitive ability enhanced.