GOOG - Descending Channel Reversal / Target $120 / CHATGPT WRONG
0. Google in a AI Hype BS Imposed Down-Channel
1. Volume is descending, and only increased when the bulls smashed to the top
of the channel yesterday and handed the bears their asses;
2. When the CHATGPT HYPE Train comes to and end, and all the flunkies realize that
you need lightning fast, up to the second accurate, multiple returns, with relevancy to have a bloody search which is ONLY GOOGLE
3.
Microsoft have spend nothing on their Search Algorithms and ChatGPT only searches a 2 YEAR OLD LANGUAGE model to get
a statistical analysis on the patterns of words in, and then to statistically get a response token by token, from the statistical
engine.
Generative AI
Generative AI is nothing more than a massive statistical analysis database of token frequency, order, relationship and other stuff. Then
you generate tokens coming out. This is simply a trained model producing tokens.
It is not a "Search"
The Data is 2 years old, and get older each day
The LLM (Large Language Model) cannot be changed. Adding new data takes 1.5 YEARS for CHATGPT.
That is right, if you wanted to update your address or phone number of your RESTAURANT, get on the list, it takes 2 years to show up
in the data! and then it can't be changed for another 2 years
This is madness
GENERATIVE AI USE CASES
1. Basic Summary (Makes massive mistakes so be careful, only basic stuff works
2. Tutoring / Assisted Learning (also, makes mistakes, and is 2 years old, but can help)
3. Helping Assist in coding (makes MAJOR mistakes) - software APIs change sometime 2x per year, this thing
has 2 or 3 year old code
4 . Cheating in school or on tests (If you have the RIGHT DATA!)
CAN'T BE UPDATED, TAKES 2 YEARS TO TRAIN, SLOW AS SHIT
1. You can't update the model
2. 2 years+ to train and 1 Billion each time
3. slower than shit, especially and CANT SCALE
CAN'T SCALE
Notes and links here
CAN'T DO REAL SEARCH
Notes and links here
MAKES STUFF UP
Notes and links
SLOW SLOW SLOW
Notes and links here
1. Volume is descending, and only increased when the bulls smashed to the top
of the channel yesterday and handed the bears their asses;
2. When the CHATGPT HYPE Train comes to and end, and all the flunkies realize that
you need lightning fast, up to the second accurate, multiple returns, with relevancy to have a bloody search which is ONLY GOOGLE
3.
Microsoft have spend nothing on their Search Algorithms and ChatGPT only searches a 2 YEAR OLD LANGUAGE model to get
a statistical analysis on the patterns of words in, and then to statistically get a response token by token, from the statistical
engine.
Generative AI
Generative AI is nothing more than a massive statistical analysis database of token frequency, order, relationship and other stuff. Then
you generate tokens coming out. This is simply a trained model producing tokens.
It is not a "Search"
The Data is 2 years old, and get older each day
The LLM (Large Language Model) cannot be changed. Adding new data takes 1.5 YEARS for CHATGPT.
That is right, if you wanted to update your address or phone number of your RESTAURANT, get on the list, it takes 2 years to show up
in the data! and then it can't be changed for another 2 years
This is madness
GENERATIVE AI USE CASES
1. Basic Summary (Makes massive mistakes so be careful, only basic stuff works
2. Tutoring / Assisted Learning (also, makes mistakes, and is 2 years old, but can help)
3. Helping Assist in coding (makes MAJOR mistakes) - software APIs change sometime 2x per year, this thing
has 2 or 3 year old code
4 . Cheating in school or on tests (If you have the RIGHT DATA!)
CAN'T BE UPDATED, TAKES 2 YEARS TO TRAIN, SLOW AS SHIT
1. You can't update the model
2. 2 years+ to train and 1 Billion each time
3. slower than shit, especially and CANT SCALE
CAN'T SCALE
Notes and links here
CAN'T DO REAL SEARCH
Notes and links here
MAKES STUFF UP
Notes and links
SLOW SLOW SLOW
Notes and links here
Comment:
Microsoft's new Bing powered by ChatGPT hates admitting its mistakes, accused of gaslighting users
www.indiatoday.in/te...s-2334555-2023-02-14
www.indiatoday.in/te...s-2334555-2023-02-14
Comment:
www.indiatoday.in/te...o-2334803-2023-02-14