Can you provide an example of this? There are a few inventions-via-serendipity (e.g. Teflon) where someone got lucky, but few that i can think of where the specific person who got lucky mattered. Kary Mullis and PCR maybe?
AlexNet that started the deep learning revolution. Was there other groups trying to solve Image Net by training convnets on GPU’s in 2009?
OpenAI and the scaling hypothesis. Before chatGPT it was only OpenAI training larger and larger language models.
TRPO, PPO line of RL algorithms invented by John Schulman
In fact I would say, in general a lot of inventions only happen because 1 person or 1 group willed it into existence and it would seem very unlikely for it to happen if that 1 person or group did not exist.
Il sorry to bomb your comment, but I saw on another thread that you followed CMU intro to DL and had a beautiful time doing all the homework.But I didn't find your contact info.
Do you happen to still have the homework questions/solutions somewhere ?
At the moment the new ones are locked behind CMU credentials.
We've discussed bicycles before. The material is hardly simplistic. There wasn't much serendipity involved. Practical bicycles came soon after the necessary basic technology had been developed, including precision bearings and robust pneumatic rubber tires. This wasn't a case where we could have had bicycles a century earlier if someone had come up with the idea.
"practical bicycles" is nebulous, vague, and always means whatever you want it to mean to claim to be right.
Bicycles do not need precision bearings nor pneumatic tires, as proven by the fact that bicycles predate them by decades. Also radial style ball bearings were invented for bicycles. Bicycles were the catalyst for precision bearings, not the other way around. Similarly pneumatic tires were created for tricycles (long after pedal bicycles were already invented)
Yes we absolutely could have had 2 wheeled transportation centuries earlier. Roads, the actually necessary infrastructure, was long since a thing for carts and wagons.
Surprisingly, there’s also still no simple explanation for why a bicycle works. It’s easier to understand how a modern microprocessor works than a bicycle.
AlexNet that started the deep learning revolution. Was there other groups trying to solve Image Net by training convnets on GPU’s in 2009?
OpenAI and the scaling hypothesis. Before chatGPT it was only OpenAI training larger and larger language models.
TRPO, PPO line of RL algorithms invented by John Schulman
In fact I would say, in general a lot of inventions only happen because 1 person or 1 group willed it into existence and it would seem very unlikely for it to happen if that 1 person or group did not exist.