1] AI is unlikely to be as intelligent as humans even as the AI models become more intelligent simply because computation can’t mimic thinking.
2] The biggest achievement of LLMs is enabling the transfer of thoughts and ideas to and fro between our minds and databases.
3] AI will increase the net number of jobs in tech because of the Jevons Paradox.
4] If you want to make money, it’s prudent to not fall in love with what you’ve built but to feel motivated to solve a problem people will pay to get solved. But prudence is no guarantee of success too; in fact, breakthrough success often goes to the moonshot ideas. In short, there’s no playbook for success in software.
5] However, there’s probably a strong correlation between how boring your product idea is and the probability of it making money for you. If you’re making something that excites people around you, then probably hundreds of people are making the same thing and it’s a crowded market.
6] Software makes the world a better place because it has significantly reduced the need for capital to produce value-creating artifacts.
7] Nuclear fusion will be a bigger revolution than LLMs because it’s not intelligence but energy that’s the more fundamental need and something that comes from outside the planet (because even the fossil fuels are stored solar energy).
8] The most important innovations are about (1) how to save the planet and (2) how to cure terminal diseases cheaply. Although many innovations have a second-order effect on these, I don’t see too much focus on innovations that directly solve these. That’s the sad part of the tech world. It happens because both these have long revenue realisation cycles: something capitalism doesn’t reward much.
