near_futurology
A blog about what may soon happen in tech … or not. Who knows?
-
Machine learning isn't always the right solution to a problem
-
How is Open Source sustainable if it’s free
-
Onboarding new employees is done wrong. It should be done through forums rather than documentation
-
We need to sell our data, and be properly compensated for it
-
Deep Learning is not magic, you still need to model the problem
-
ML/AI needs training data; training data comes from operations; operations need ML/AI
-
Users shouldn’t pay to store data, we should do it ourselves
-
All jobs require exactly 2 functions; LLMs can perform only 1 of them
-
LLMs are legit, but they’ll still cause a bubble
-
The internet doesn’t have enough data to train LLMs
-
ChatGPT won’t cut it; You will have to have your own LLM
-
Your next UI will be invisible
-
Of course LLMs aren’t AGI, but the question itself is a huge compliment