I hear ya man, that white space type of control-freak-y stuff really gets me. These are the other control-freak-y things I don't like:
1-Being required to have an opening curly bracket.
That one looks like a sideways mustache and that just creeps me out.
But the closing curly bracket looks like a sideways happy face, I want lots of that in my code.

2-Having to use the exclamation point for negation.
It's SO negative.
Instead of a friendly symbol that says "hey it's ok whatever value you want to be, but if it just so happens you're not the value we want here, we're going to kind of go-ahead and do a little bit of stuff"

The exclamation point comes across as saying "If you are NOT the value we want THEN DO NOT ENTER this special little club that gets to do some special stuff, go away and do not come back (unless you are in a loop and might change into one of the good values, then we will reconsider)"

3-Having to use ones
They look very poke-y, like you could lose an eye or something.
Zeroes look just like donuts, which taste great and wouldn't hurt anybody.

Just because their press release says they accomplished this task doesn't mean they actually did. ERP systems are large and complex due to managing so many aspects of a business. Most companies don't employ the type of experienced people to build something like this (vendor apps like SAP or Netsuite have grown over time and with staff dedicated to this space, last count I heard was SAP was >100million lines of code).

I suspect reality is significantly different than the press release.

Compaq created the PC clone (by reverse engineering IBM's BIOS and hardware setup), which opened the floodgates for the industry.

Just to play devils advocate:
Assuming we are talking about C, the C compiler can implement the storage of variables in different ways as long as the required language conditions are met, like for int minimum size based on hardware details.

The C compiler could store variables with a type byte (for example) and it could reference that value for various purposes during execution.

In addition, because of C's ability to access memory with few constraints, the do_something() function could alter the type byte at runtime, which could alter the execution of that loop.

So that's an example of a type changing at runtime that the compiler did not detect at compile time. With knowledge of the specific compiler that produces this type of variable storage, a person could intentionally take advantage of it to introduce dynamic types within their program.

You would need to have strict rules for the compiler to prevent this type of hole.

Sure, and it will also blindly write you a program to calculate the 4th side of a triangle. You don't know when you are getting valid results and when you are not, and neither does ChatGPT.

Apparently some of it was released already:
* * Twitter Recommendation Engine *
* Written by Jim "Tex" Leet
*
IDENTIFICATION DIVISION.
PROGRAM-ID. COBCALC.
ENVIRONMENT DIVISION.
DATA DIVISION.
WORKING-STORAGE SECTION.
01 PARM-1.
05 CALL-FEEDBACK PIC XX.
01 FIELDS.
05 INPUT-1 PIC X(10).
01 INPUT-BUFFER-FIELDS.
05 BUFFER-PTR PIC 9.
05 BUFFER-DATA.
10 FILLER PIC X(10) VALUE "TWEET_ID".
10 FILLER PIC X(10) VALUE "USER_ID".
10 FILLER PIC X(10) VALUE "DATA_PTR".
10 FILLER PIC X(10) VALUE "END".
05 BUFFER-ARRAY REDEFINES BUFFER-DATA
OCCURS 4 TIMES
PIC X(10).

PROCEDURE DIVISION.
DISPLAY "CALC Begins." UPON CONSOLE.
MOVE 1 TO BUFFER-PTR.
MOVE SPACES TO INPUT-1.
* Keep processing data until END requested
PERFORM ACCEPT-INPUT UNTIL INPUT-1 EQUAL TO "END".
* END requested
DISPLAY "CALC Ends." UPON CONSOLE.
GOBACK.
* End of program.

I see people pushing ideas and products and state that approach ABC results in decoupled systems. The funny thing is that viewed from axis X they are correct, but they fail to understand or point out that from axis Y they've just increased coupling the same amount they decreased it on the other axis.

There is no free lunch.

He made a tweet about "can't feel my legs" with a picture of a wheelchair guy.

I think our neural nets go beyond that to create our "understanding". It seems like a key aspect is having an internal model of something, learned and built up over time, and being able to use that model to test the validity of questions or future state.

It's not just word correlation or image correlation, it's a model for the substantial attributes, functions and rules of operation of the thing being modeled.

How does taking crypto benefit Amazon and Walmart?

Vaccine reduces severity of infection (as does getting the infection itself due to antibodies that remain present in your system for some duration). What are you thinking the poster would re-evaluate?

"IBM just discovered ChatGPT and asked it to create a PR piece about something IBM might do"

Remember, she said it would be the "summer of love"

Providing correct answers is NOT what ChatGPT and Bard are about. They are statistical models about the relationship between words and phrases. They will both always get wrong answers because the method has inherent limits.

Despite the limits, they are very impressive.

You're right you can't trust the output. ChatGPT is not designed to provide accurate answers, it's a large language model that is designed to use statistics to return the most likely words given the context and it's learning of the relationship of words and phrases. It's pretty much guaranteed to produce inaccurate results because the method has limits.

But it is pretty impressive what it produces, even if not always correct.