AI will save the world.
We’ve been calling many different technologies AI for more than a decade. Yet, this time it seems different. To answer that, two fundamental questions must be answered “Why now?” and “So what?”.
AI has a compelling “Why now” with the rise of advanced GPU chips - A100 by Nvidia (AI leader atm) and the development of large language models or LLMs. Transformers are well-suited for GPUs, which makes it practical to process enormous amounts of data and train AI models with billions and trillions of parameters. LLMs have surprised everyone with their invariable capability to generate coherent & believable language models. The ability of diffusion models (stable diffusion or firefly) to generate high-definition & aesthetically appealing images from a brief prompt (Paul Daugherty, CTO, Accenture, said "The hottest new programming platform is the napkin") has surpassed conventional assumptions about AI’s capacity.
However, what was considered AI in the 2000s is no longer considered AI. Gartner has described that over decades of technological hype cycles, the wild enthusiasm is followed by disappointment - the trough of disillusionment (coined by Konstantine Buhler). Experts estimate that LLMs accuracy, once near perfect levels (currently at ~90%), will enable AI to act as an extended mind along with search engines & smartphones rather than the elusive technology it is right now.
Why AI won’t take away your job
The pundits are out screaming that AI will take away your job. The doomsday is near, and Orwell’s dystopian world is a reality not too far from now. Well, they are neither wrong nor right in the true sense. It is important to look at this inflection point from a different lens.
AI or Gen AI will enable automation. A lot of automation. And with automation comes redundancy in many jobs. Automation impacts jobs and people. But is it the first time we’re automating? Every other inflection point in history has led to job losses in certain categories but has also created new classes of jobs. What happens is a mere shift in the job category from one location, class, or skill to another. Total number of jobs doesn’t go down. This is what the “Lump of Labour fallacy” talks about. It is a misconception that there is a fixed amount of work to be done and that if a machine does some work, there will be less work for people. But if it becomes cheaper to use a machine to make, say, a pair of shoes, then the shoes are cheaper, more people can buy shoes, and they have more money to spend on other things besides, and we discover new things we need or want, and new jobs.
All of this implies that we should expect LLMs to destroy, create and accelerate jobs just as Excel & typewriters did. It’s just more automation. The machine lets a person do 10x the work, but you need the person.
Another important factor to consider if you’re scared of ChatGPT is the error rate. It can answer ‘anything’, but the answer might be wrong. This is called Hallucination - trying to sound confident even when it doesn’t know the answer. When you ask ChatGPT or Bard a question, it isn’t replying to you with an accurate answer. It is basically responding to the most likely answer or matching a pattern. It’s not a database lookup. AI isn’t that smart (just like an astronaut can’t ride a horse in space). This isn’t AGI. If it was, then yes, you should be worried.
Written without the help of AI :)
Here are some interesting articles/books I would recommend to tickle your AI curiosity:
Articles/Reports:
领英推荐
4. https://stratechery.com/2023/ai-and-the-big-five/ (See what MANGA are up to)
Accenture is bullish on this theme (see recent report of its commitment to invest over $ 3B in AI) and has published some amazing work in this field. This report on GenAI is worth a read:
Books:
Also, if you believe this trend is worth the hype and want some skin in the game, then you can either invest individually in these companies, such as Nvidia, Microsoft, Apple and ASML or take the mutual fund route (something that I prefer because it gives me an opportunity to invest in multiple companies across USA, Netherlands, Taiwan etc.). One such fund that I know that does this amazingly well and is leveraging FOF strategy is the "DSP Global Innovation Fund", which invests in:
If you have any thoughts or ideas or simply want to interact on this topic, do reach out.