-
Popular LLMs training data, what do they use ?
Popular LLMs training data seems to be universal and generic. This is why such models are so popular, they more or less know an answer to everything. But how do they come about to those answers ? What is the source of that ? Where do they get the data from ? Let`s search the web the old fashioned way and find out. Popular LLMs training data types The training data for these models come from all around the world. We humans are the ones that provide it. It is our work that is pushed into a model. LLMs training data reflects carefully curated huge datasets designed to provide high…
-
AI Slop Is Destroying The Internet – Kurzgesagt – In a Nutshell
AI Slop Is Destroying The Internet Just watch the video and remember on your next ‘work’.
-
Before:2023 google hack – get results before the AI crap
Before:2023 – add this before the search query and get results before the AI crappy era. In addition it should also render a lot more organic results without pesky ads. Since the algorith changes it feels like everything thing is either sponsored or just plain in ad. Call me a tin foil hat but the dead internet theory isn`t that dead anymore 😉 I miss the internet where people blogged and wrote their thoughts like on myspace. Too bad it didn`t work out, everything got monetized. Try to add “+blog” to hopefully find a personal blog of a person that really dug into your problem and has a nicely documented…
-
Commonly used sentence in AI NLP speech recognition ?
The commonly used sentence in AI NLP speech recognition is The quick brown fox jumps over the lazy dog AI for everyone The curiousity about that particular sentence is that it containes every letter in the alphabet, from A to Z. It is also called a pangram or a holoalphabetic sentence. Uses every letter of a given alphabet at least once. Pangrams serve various purposes as showcasing typefaces, testing equipment, and honing skills in handwriting, calligraphy, and typing.
