The Art of Investing Message Board

STOCKTALK

We've detected that you're using an ad content blocking browser plug-in or feature. Ads provide a critical source of revenue to the continued operation of Silicon Investor. We ask that you disable ad blocking while on Silicon Investor in the best interests of our community. If you are not using an ad blocker but are still receiving this message, make sure your browser's tracking protection is set to the 'standard' level.

Strategies & Market Trends : The Art of Investing

PICK 49.91

+1.0%

4:00 PM EST

Public Reply Prvt Reply Mark as Last Read File

Previous 10 Next 10 Previous Next

To: Sun Tzu who wrote (10616)	11/23/2025 4:51:56 PM
From: Johnny Canuck	of 10701

I think the point made in the previous video I posted did a good job of summarizing the pros and cons of each AI. Gemini is making strides in being able to process visual data and turn it into more structured descriptions for processing. Anthropic has built a business by being good at some enterprise tasks. I think they mentioned that Claude was the best at coding. Though there are other videos the suggest Claude or related AI will never be able to completely replace experienced programmers. It is just too hard to debug complex code bases without a human in the loop to pull it all together and decipher it all if something goes wrong. At my day job we use AI to create module/functions which we test extensive before using as opposed to getting it to write complete programs. We use it more as a fancy interface to an super large library of functions essentially or use it to prototype something we have not done before so we can pick it apart to understand what it does. I think it depends on what you need. My concern about depending on any one AI is that it does not take much to significant change the characteristic of the AI. I watched a presentation at B-Sides in Las Vegas, a hacker convention, as few years ago about corrupting AI engines but feeding it biased datasets. The concern being that at the time, META updated their version of an AI every 4 hours. By feeding it certain data you can slowly bias the AI in specific directions and generate blind spots or make it miss categorize information. For now I have figure out what performs the best for specific tasks and switch depending on what i need. I also use independent AIs to identify hallucinations by asking it the same question and comparing answers. It is still not perfect but if you are experienced and can spot hallucinations or flaws in the logic it can speed up your work flow.

Report TOU Violation

Share This Post

Public Reply Prvt Reply Mark as Last Read File

Previous 10 Next 10 Previous Next