Would you trust ChatGBT

[Anonymous]

Chat tells you what you want to hear. Anytime I hear someone mentioning "asking chat" I laugh.

[Anonymous]

Anonymous wrote:

Anonymous wrote:

Anonymous wrote:

Anonymous wrote:If you gave Chat GPT honest and detailed numbers as a gut check and got this response would you trust it?

“Final bottom line

You are in the top tier of retirement security:
• Lifestyle is fully protected
• Market risk is non-threatening
• Taxes are the primary enemy — not spending
• A seven-figure inheritance per child is the base case, not the upside case”

This wouldn't tell me anything I didn't already know, or should know, by looking at the numbers.

Duh, but you have to come up with the numbers first, and that’s what it did

So this is meaningless. Like a pat on the head or something. I don't get the point.

Huh? I put my numbers in and asked how they would work in retirement and it gave me an answer. I’m not sure what you’re getting at. It’s no different than giving your numbers to a human advisor and asking.

[Anonymous]

It just told me I have more money than most private school families in the DC area, but don’t spend as much as many of them on flashy cars, housing and vacations. Meanwhile it says I can buy a vacation home and fly business, and still grow money for my kids. When I retire early, people will be surprised but will respect me for it and consider me quietly wealthy. It also said I can’t afford to go on an “Hermes Journey” unless I really budget for it, but other shopping is fine. Haha good to know.

[Anonymous]

It depends quite a lot on which model version you are using. Free models go way off the rails at arbitrary moments, forget recent information, and are very very good at making up fake explanation to justify contradictions.
Some models use calculators and write computer programs to get math correct.

[Anonymous]

Anonymous wrote:

Anonymous wrote:

Anonymous wrote:

Anonymous wrote:

Anonymous wrote:If you gave Chat GPT honest and detailed numbers as a gut check and got this response would you trust it?

“Final bottom line

You are in the top tier of retirement security:
• Lifestyle is fully protected
• Market risk is non-threatening
• Taxes are the primary enemy — not spending
• A seven-figure inheritance per child is the base case, not the upside case”

This wouldn't tell me anything I didn't already know, or should know, by looking at the numbers.

Duh, but you have to come up with the numbers first, and that’s what it did

So this is meaningless. Like a pat on the head or something. I don't get the point.

Huh? I put my numbers in and asked how they would work in retirement and it gave me an answer. I’m not sure what you’re getting at. It’s no different than giving your numbers to a human advisor and asking.

Oh it's quite different.

[Anonymous]

The problem is you have to know enough to check its work and at that point what was the benefit of asking it?

It is just looking for similar text, so if your situation is similar to others’ then it probably can rip off the work of others and approximate the answer.

[Anonymous]

Anonymous wrote:

Anonymous wrote:

Anonymous wrote:

Anonymous wrote:

Anonymous wrote:

Anonymous wrote:If you gave Chat GPT honest and detailed numbers as a gut check and got this response would you trust it?

“Final bottom line

You are in the top tier of retirement security:
• Lifestyle is fully protected
• Market risk is non-threatening
• Taxes are the primary enemy — not spending
• A seven-figure inheritance per child is the base case, not the upside case”

This wouldn't tell me anything I didn't already know, or should know, by looking at the numbers.

Duh, but you have to come up with the numbers first, and that’s what it did

So this is meaningless. Like a pat on the head or something. I don't get the point.

Huh? I put my numbers in and asked how they would work in retirement and it gave me an answer. I’m not sure what you’re getting at. It’s no different than giving your numbers to a human advisor and asking.

Oh it's quite different.

Except it wasn’t. I did both. Same answers. That’s my point.

[Anonymous]

Anonymous wrote:Absolutely not

Trump is burning this country to the ground.

We will be lucky if the $ makes it through

Whatever monies you currently have will be worth 1/4 of what they are now.

Below ten million will be poverty level. Yes, it will. Healthcare plan "coming in two weeks" ....... idiots.

You voted for this garbage.

ignore this TDS. OP, seek out your financial advisor.

[Anonymous]

Anonymous wrote:

Anonymous wrote:

Anonymous wrote:OP here. We have actually had a real meeting with a real financial advisor. ChatGPT’s analysis and recommendations are basically the same. I can understand why so many are concerned that AI will put everyone out of work . . .

AI is a joke when it comes to personal finance. If you think it's a replacement for a person with experience and knowledge, go on bogleheads and look at the crazy responses AI came up with when passed with basic questions. The only people I know using AI in this way are crypto bro types.

Yea well I disagree. If you know nothing about the subject then yes you’re taking a chance. But if you approach the discussion with ChatGPT with some knowledge it can be extremely useful and accurate.

To me, it’s like using an Internet translator. If you know nothing about the language you’re asking to have translated it can make big mistakes. But if you know the language a little bit, you’ll catch them easily. Same with Chat GPT.

If your knowledgeable about the subject, what's the point of using it if you can't rely on it with any degree of certainty? I used it the other day to play around with with Roth conversion numbers and it was kind of entertaining. Maybe some day, but at this point it's kind of a glorified search engine.

[Anonymous]

Chatgpt is accurate until it isn't - all of these llm's are the same and can make profound wrong statements with full assurances that they are right.

[Anonymous]

Anonymous wrote:Chatgpt is accurate until it isn't - all of these llm's are the same and can make profound wrong statements with full assurances that they are right.

As can humans. And those human professionals are increasingly using AI to inform/augment their work in the first place.

[Anonymous]

Anonymous wrote:The problem is you have to know enough to check its work and at that point what was the benefit of asking it?

It is just looking for similar text, so if your situation is similar to others’ then it probably can rip off the work of others and approximate the answer.

Exactly

[Anonymous]

Anonymous wrote:

Anonymous wrote:Chatgpt is accurate until it isn't - all of these llm's are the same and can make profound wrong statements with full assurances that they are right.

As can humans. And those human professionals are increasingly using AI to inform/augment their work in the first place.

These why you either learn the subject yourself or use the services of someone who is competent. Comparing a dumbass to chatgpt isn't the best argument.

[Anonymous]

not from my experience of working with chatgpt. it made simple silly mistakes, more than once. but you do you.

[Anonymous]

I am using it currently to plan detailed retirement strategies. It is far improved over the last year. Always double check any math calculations of course, but it makes far fewer dumb mistakes than it used to.