This is ChatGPT: 1 million registered users in just 5 days after launch. Less than 3 months later, ChatGPT is crashing continuously because of over-usage. The US market will soon be offered a paid subscription that is expected to be another hit.
No other Tech adoption has happened so fast. The 1 million users milestone, was reached in 2.5 months by Instagram; the Spotify platform reached it after 7 months; Twitter had to wait for 24 months.
The 100 million users milestone has already been reached by ChatGPT in less than 3 months, compared to 2 years for Instagram and 1 yr for TikTok.
But the nature of the ChatGPT Tech Diffusion is even more remarkable and there is no better example than the one that Satya Nadella, the CEO of Microsoft, shared during a WSJ panel in Davos during the WEF. [1]
His example speaks to the power of democratization of natural language processing models. It speaks to the potential of deploying GPT models to start closing the digital divide that is plaguing our society as we speak and will only get worse.
In today’s society, as advanced technologies penetrate every aspect of our lives more and more, many people feel intimidated and are becoming increasingly “digitally illiterate”.
I am not talking about the large percentage of the global population that remains offline due to poverty and no access to the internet and hardware (computers, smartphones). [2] I am referring to the part of the population that does have access online but doesn’t have the training to leverage its capabilities and is dangerously left behind because of all the novelties introduced.
People used to know how to pay with cards but what about P2P payment apps. People used to know how to buy an airplane ticket online but didn’t know how to complete PLF forms and show vaccine certificates to travel during the pandemic.
There is an ever-increasing need for digital self-education.
As jobs will be lost because of technological advancements, there is a growing need to educate adults in their 30s and 50s and 60s; to work and perform the newly created tasks.
This is exactly where the potential of deploying GPT models, is immensely important as it can offer huge Social impact.
Digital inclusion can accelerate if the public sector deploys these natural language processing models and helps society adapt to the new realities. With one stone this kind of deployment of GPTs can hit two birds. One is educating adults to become part of the digital workforce (only if done at scale with the public sector commitment) and second alleviating the mental health crises and its social implications from the growing Digital Divide.
Again, the potential social impact of this type of Tech Diffusion is underestimated. It goes beyond, the currently popular discussions about the implications of these advanced Natural language processing models in content creation (text, image, video, music, arts,…).
In Financial Services, we can expect 2023 to be the Chatbot year in Banking (as Ron Shevlin titles his latest Forbes article) thanks To ChatGPT. But the social impact potential in financial services lies in the collaboration of public services and private `Fintech` (native Fintechs or transformed incumbents) to leverage this tech and amplify financial inclusiveness at levels never seen before.
Imagine, training these models to answer and implement the actions needed, on questions like:
- `I don’t know what I am supposed to do to apply for this subsidy?`
- Ì don’t know what to do to improve my credit score?`
- `I am planning XYZ in my investment portfolio, but I am not sure of the tax implications?`
- `Based on the Fed`s current monetary policy, what can I consider as I have an ABC mortgage?`
All these, of course, are collaborations of the public & private sector.
I must confess that as I write all this, I bring to mind Warren, the ChatGPT like focused on analyzing stocks, markets and investing, launched by Fintech Kensho. It had a search capability a la Google and established a collaboration with CNBC and a paid subscription offering.
I had played around with the paid subscription that allowed questions like:
- Which Apple supplier’s share price goes up the most when the company releases a new product?
- What happens to the currencies of oil-producing countries when oil trades below $45 and the USD is strengthening?
- Which part of the yield curve is more affected 3 months before the first rate hike?
- How do pegged currencies to the USD react, when the dollar strengthens in a strong trendy fashion against most major reserve currencies?
In 2015, I wrote on Daily Fintech Kensho: Warren is like Watson and Siri, for analysts, investors and traders. Later that year Kensho was acquired by S&P Global (Techcrunch wrote, S&P Global snares Kensho for $550 million)
Kensho is functioning independently of S&P Global but leverages the data capabilities of S&P Global. It offers several diverse solutions. Some are laser focused in investing and other are broader. Here are 3 examples:
- The S&P Kensho New Economy Indices provide the most comprehensive framework with which to capture the industries and innovations of the Fourth Industrial Revolution.
- Kensho NERD (a knowledge graph approach) unlocks new connections and deeper insights by systematically uncovering the companies, people, places, events and more in your text.
- Kensho Scribe is a transcription solution optimized for tricky, real-world audio.
The point is that we have a choice in deciding where these advanced natural language models are deployed. We can decide to establish collaborations that have huge potential of social impact network effects (like the one of the Indian Ministry Ministry of Electronics and Information Technology (MeitY) and Microsoft). Big tech companies and governments can make this happen soon. The tech is here now and will be getting better and better every day. It is not a choice that requires a long-term vision and patience. We can start reaping benefits very soon.
The alternative is to use these models in very specific business or scientific areas. The case of Tesla`s AI lead, Andrej Karpathy (mentioned by Satya Nadella) or the cases that Kensho tackles, of course, have benefits and returns but these types of deployments are not about Social impact and taking care of the growing Digital Divide.
We actually want both to happen. I am not worried about whether Techies will leverage these advancements or whether the race between the BigTechs around this type of AI deployments will bear fruits.
The risk is that we don’t see enough of the public/private deployments, and we don’t leverage these models toward tackling the growing Digital Divide.
APPENDIX
ChatGPT`s answers to the Kensho type questions:
ChatGPT`s answer to ` Which Apple supplier’s share price goes up the most when the company releases a new product?` as you see is not actionable for an investment analyst, investor, or trader. Although correct it is too broad.
- It is difficult to determine which specific Apple supplier’s share price will go up the most in response to the release of a new product, as the impact can vary depending on the supplier’s specific role in producing the product and their overall financial performance.
However, in general, suppliers that are heavily involved in the production of new Apple products, such as those that supply components like display panels, cameras, and batteries, may see a rise in their share price. This can be due to increased demand for their products as a result of the new product release.
It’s worth noting that stock market reactions to new product releases are not always predictable, and other factors such as the overall market conditions and the financial performance of the supplier can also influence the stock price.
ChatGPT`s answer to `Which part of the yield curve is more affected 3 months before the first rate hike?` does not contain any facts from previous hikes to size the effect.
- The short end of the yield curve (3-month to 2-year maturities) is typically more affected in the lead-up to the first rate hike. As the central bank signals its intention to raise rates, short-term interest rates tend to increase in anticipation of the actual rate hike. This can cause the yield curve to steepen, with a larger spread between short- and long-term interest rates.
ChatGPT`s answer to `How do pegged currencies to the USD react, when the dollar strengthens in a strong trendy fashion against most major reserve currencies?`
- When the US dollar strengthens against most major reserve currencies in a sustained manner, currencies that are pegged to the US dollar are likely to follow suit and appreciate relative to those currencies. This occurs because the value of a pegged currency is directly tied to the value of the currency it is pegged to. In this case, as the US dollar strengthens, the value of the pegged currency also rises, maintaining the fixed exchange rate between the two currencies.
However, maintaining a fixed exchange rate can put pressure on a pegged currency’s central bank, as the rising value of the currency can make exports less competitive and hurt economic growth. In such cases, the central bank may intervene in the foreign exchange market by selling its own currency and buying the US dollar to maintain the peg.
All Comments