AI, Software, Coding, Internet Security Thread
- Thread starter Indos
- Start date
Hamartia Antidote
Elite Member
Hamartia Antidote
Elite Member
Interesting...I noticed this early in the video
This is an example of a local "Americanism" that seems to have gone global.
So the "Nutrition Facts" label was a US Government requirement food companies fought for YEARS and YEARS not to be put on their product labels. Sort of like how cigarette makers don't want to alert people of any health dangers.
This was an insane fight over such minutia as the minimum font size, text bolding, and bare minimum requirements. It finally appeared in 1990.
Nutrition Labeling and Education Act of 1990 - Wikipedia
FDA's Role in Improving Nutrition: Labeling and Other Authorities
The History of the Nutrition Facts Label | Unsung History
Last edited:
Hamartia Antidote
Elite Member
Yeah but I don't think those arcade machines were using modern "English" programming languages. These were the days of cryptic assembly which not many people had the patience to do...so the Japanese had a niche.
note the .asm files
Hamartia Antidote
Elite Member
So like people who claim Tesla Full Self Driving "simply doesn't work" because it sometimes can't handle a particular situation...you have developers who say "AI can't write code" because once in a while it is wrong. So they refuse to use it.
The fact is I can ask ChatGPT to write me a method that does a particular task and it generates it in 3 seconds. I can then copy/paste it and determine if it works. That saves me time. I can even ask it if people typically use X or Z and it will come back and say Z. I can then ask "why?" and it will come back with a justification.
Indos
INT'L MOD
- Thread starter
- #66
AI is good to gather knowledge, make and debugging coding. For content writer it is also helpful
AI in my opinion is still less capable to make good strategic move. In my business I dont follow Chat GPT suggestion when it comes to make strategic decision. They are good for technical things and anilizing limited data set, when it comes to make strategic decision when it needs large data set they will be logically less clever than human
Indos
INT'L MOD
- Thread starter
- #67
With slower global population growth, I would say AI technology will fit more into this type of society. Minus population growth is bad, what I mean is 2.0 and 2.1 fertility rate.
Agriculture, restaurants, foodsellers, armed force, police, e - commerce, minimarket chain, manufacturing, customer service, construction (positive population growth demand more housing for example) etc still requires human workers, I dont like programing customer service tool/device, human will be still preferable.
The boom of tech engineers in computer programming during boom time where the demand is high of course will face AI era, this is where adjustmen will happen in the jobs market, but I think AI could generate more entrepreneur as the most benefit of AI in my opinion goes to people with knowledge, programming skills, but with selling capability.
Less capital is now required to make start up company, particularly in tech industry. People with brain, good mentality, coming from middle class families have now better leverage to compete with people coming from wealthy families when it comes to entrepreneurship in tech industry (digital economy).
Agriculture, restaurants, foodsellers, armed force, police, e - commerce, minimarket chain, manufacturing, customer service, construction (positive population growth demand more housing for example) etc still requires human workers, I dont like programing customer service tool/device, human will be still preferable.
The boom of tech engineers in computer programming during boom time where the demand is high of course will face AI era, this is where adjustmen will happen in the jobs market, but I think AI could generate more entrepreneur as the most benefit of AI in my opinion goes to people with knowledge, programming skills, but with selling capability.
Less capital is now required to make start up company, particularly in tech industry. People with brain, good mentality, coming from middle class families have now better leverage to compete with people coming from wealthy families when it comes to entrepreneurship in tech industry (digital economy).
Last edited:
AjayGhatak
Trusted Member
This is a fallacy. Sega Saturn used C mostly. Sega Dreamcast used C/C++. Arcade boards since mid 90s used C and C++ to speed up game development. Assembly was the development laguage for surprisingly many embedded systems as late as 2010s. Infact, for DSPs it was practically the only way.
Games written in early 90s for SNES, Mega-drive were assembly language ones. N64, Playstation etc all used C/C++. Mid-90s most of the game development had moved to C/C++ in japan.
AjayGhatak
Trusted Member
Its very early to tell. Surprising amount of work in development of software is not really programming but decisioning. Thats the crux of engineering. Making decisions. This is even more true for web based development.
Right decisions can save you millions.
Here is one video that I found very intriguing because the way they approaching hosting a major website / e-commerce portal. They make a lot of counterintutive choices. LLM, as they are right now will be hard pressed to present it. Because these choices are not really something you can read a book and learn.
ALL of LLMs I tried asking suggested me cloud based setups. When you really want to run your business frugally, you have to actually know and understand the underlying internet.
There is a difference between running code and a solution that can support a million customers for pennies. Unless you are exposed to these things, you will not even be able to ask the right questions or prompts to LLMs.
And this is very important because these questions have no right answers actually. Most of them is "it depends upon cost and scale".
For a lot of them its not even "This or that" choice (see the video). Its "mostly this" but "when it fails or sometimes that".
"But Sometimes" is a very interesting phenomenon in engineering of any type including software. Because engineering is never one shoe fits all kind of business. You have to take care of "But Sometimes" situations or you will end up chosing less optimal solutions. Because "Sometimes" is well... sometimes and not every times.
i guess this guy explains it very well.
Last edited:
Hamartia Antidote
Elite Member
I was thinking more along the lines of Arcade machine games not home consoles.
AjayGhatak
Trusted Member
Sega's entire line of home machines actually came from their arcade boards scaled down to serve home markets.
AjayGhatak
Trusted Member
Here is my recent experience with ChatGPT in a REAL world scenario. The ChatGPT involved here is o1-preview.
So... I bought a rental property built in 80s. I got basic inspection done before purchase. I just wanted to be sure before I rent it out so I asked ChatGPT what things I should do (I will post exact of its responses later, they are HUGE). It suggested a lot of things and I used it as a to do list. This was my first major rental property.
One of the things it suggested was hazardous material inspection. Well, for 600 dollars I got it done. Results? House exterior has HUGE amount of lead... That means.... 10s of 1000s of dollars in remediation. I asked ChatGPT as well to ensure all my options are explored. It gave:
Encapsulation, Enclosure, Paint Removal, Componen replacment. All solid choices from least expensive to most expensive.
I asked the HazMat Remediation folks. They said, just encapsulate it with an acrylic paint coat. I got the quote for that too. 15K in material and labour.
I was about to bite the bullet but then I ran accros my old handyman. He said it is all BS. I have stucco as my exterior and no way in the hell I can have lead in stucco. He told me that most likely something got mixed up somewhere and the folks who did the testing did not understand your reasons behind it. I did not even know about stucco as exterior... good call! So I went back and talked to HazMat folks and they said they always mix all samples before testing because thats what is required in demolition and almost all of these sample tests are for demolition. Face Palm!
So I asked them to do it again this time each part separately. And Lo and behold! All the lead was in the window siding... Just 300 dollars to replace all of them. The most expensive option was the least expensive one.
This is why one must be REALLY careful when applying chatGPT to real world. Real world is VERY complex and LLMs are not there yet. Whats more, if you do not know something, it is likely you will NEVER prompt LLM properly.
So... I bought a rental property built in 80s. I got basic inspection done before purchase. I just wanted to be sure before I rent it out so I asked ChatGPT what things I should do (I will post exact of its responses later, they are HUGE). It suggested a lot of things and I used it as a to do list. This was my first major rental property.
One of the things it suggested was hazardous material inspection. Well, for 600 dollars I got it done. Results? House exterior has HUGE amount of lead... That means.... 10s of 1000s of dollars in remediation. I asked ChatGPT as well to ensure all my options are explored. It gave:
Encapsulation, Enclosure, Paint Removal, Componen replacment. All solid choices from least expensive to most expensive.
I asked the HazMat Remediation folks. They said, just encapsulate it with an acrylic paint coat. I got the quote for that too. 15K in material and labour.
I was about to bite the bullet but then I ran accros my old handyman. He said it is all BS. I have stucco as my exterior and no way in the hell I can have lead in stucco. He told me that most likely something got mixed up somewhere and the folks who did the testing did not understand your reasons behind it. I did not even know about stucco as exterior... good call! So I went back and talked to HazMat folks and they said they always mix all samples before testing because thats what is required in demolition and almost all of these sample tests are for demolition. Face Palm!
So I asked them to do it again this time each part separately. And Lo and behold! All the lead was in the window siding... Just 300 dollars to replace all of them. The most expensive option was the least expensive one.
This is why one must be REALLY careful when applying chatGPT to real world. Real world is VERY complex and LLMs are not there yet. Whats more, if you do not know something, it is likely you will NEVER prompt LLM properly.
Users who are viewing this thread
Pakistan Defence Latest
Country Watch Latest
-
TF-X / KAAN / Hürjet Turkish Fighter & Trainer Aircrafts News & Discussions (6 Viewers)- Latest: MMM-E
-
-
-
-
Latest Posts
-
TF-X / KAAN / Hürjet Turkish Fighter & Trainer Aircrafts News & Discussions (6 Viewers)
- Latest: MMM-E
-
-
-
-
