Unless you are a hermit and have been living off the grid for the last few years you have heard at least something about artificial intelligence. The probability that this “something” included the mentions of either OpenAI or Sam Altman is also very high.
The man has been in the centre of what some call a major tectonic plate movement in tech or even have compared to the discovery of electricity. Who is better placed than him to share his thoughts about artificial intelligence?
Here are my 3 takeaways after watching Sam’s meetings with students at the University of Michigan and Harvard University on YouTube:
Lesson #1
After the latest version – the GPT-4.0 – the main goal has shifted towards teaching models to reason. In OpenAI speak that would mean the process of moving from Level 1 to Level 2.
Sam didn’t offer a more detailed explanation, so I had to look elsewhere to find out
about the 5 levels of intelligence used by OpenAI:
- Conversational AI – where computers can interact in conversational language with people;
- Reasoning AI – where systems can perform basic problem-solving tasks (equivalent to a PhD level of complexity) without access to external tools;
- Autonomous AI – where systems (or agents) can operate autonomously on a user’s behalf for some time (say, for several days);
- Innovating AI – where systems are capable of developing innovations independently; not just running the processes; but improving them;
- Organisational AI – where artificial intelligence is at the level that it is capable of performing the work of an entire organisation.
We are on the right path to building general intelligence but we are still a long way from getting to the self-improvement loop.
Lesson #2
The 2 key commodities of the future are going to be artificial intelligence and energy. The history of energy demonstrates that massive improvements in quality of life have been achieved with the energy prices coming down.
Sam is a firm believer that fusion will work and that it will become the cheapest and cleanest source of energy. If it takes longer to develop, the solar energy plus efficient and sufficient storage capacity solution could create a similarly affordable energy supply. The longer-term objective would be to reach the levels where the price of intelligence approximates the price of energy.
Pushing the price of computing and the price of energy to levels as low as possible is also an instrument to ensure lower levels of inequality. The other would be to make access to computing a fundamental human right, similar to the universal base income concept.
Lesson #3
Those who are contemplating starting AI businesses or becoming investors in such businesses today essentially have a choice between 2 strategies:
- to bet that the technology has reached a level as good as it is ever going to be, or
- to bet that the technology is going to become significantly better.
With OpenAI and others working daily to make the technology better, in Sam’s opinion, the decision should be quite obvious. The main misconception prevailing in society still today is to seriously underestimate how good the new models really are and what the speed of technology development will be going forward.
And, as a final note, a piece of life advice, based on his own story – that it is entirely possible to simply get out and do stuff in the world. Which is what Sam has been and continues doing.
For those willing to try and pick top lessons of their own here are the links to Michigan and Harvard meeting videos.
* from anything that you are reading, watching or hearing you can realistically expect to remember only a limited number of things. My solution is to pick just 3 items or ideas from any material. This number is non-negotiable. Even the most extraordinary experience gets compressed into 3 things to remember. This approach has worked well for me.
The note was first published on Medium.com on 22 October 2024.
