00:00:00
Wall Street Journal reporter asked Chat
00:00:01
GPT for investment advice to see if AI
00:00:03
could really realistically manage a
00:00:05
stock portfolio. Chat GPT generally
00:00:07
identified market risk, but also made
00:00:09
simple math mistakes, suggested
00:00:11
questionable market timing strategies,
00:00:13
and sometimes veered into overly complex
00:00:15
ideas. The chatbot often told the user
00:00:17
what they seem to want to hear, even
00:00:19
when its advice may be flawed or risky.
00:00:21
Sounds like a lot of brokers I know. Any
00:00:23
thoughts?
00:00:23
>> I'm guilty of this. I've been asking
00:00:25
Claude like crazy, using co-work to come
00:00:27
up with different options and hedging
00:00:28
strategies and to identify stocks. I'm
00:00:30
guilty of it. I love the markets. I like
00:00:32
to gamble. It's my means of gambling.
00:00:34
And I find it interesting what it says.
00:00:36
But here's the bottom line. Ken Griffin
00:00:38
will hire 2,000 PhDs and he will
00:00:41
weaponize them with AI and they will
00:00:43
outperform the market. But you believing
00:00:45
with your $20 a month claude coowork
00:00:48
weapon can compete against these folks.
00:00:50
You are showing up with a very elegant
00:00:53
square gun and they have a [ __ ]
00:00:54
howitzer. I do think AI is going to be a
00:00:56
huge unlock. People should use AI to try
00:01:00
and pick stocks. I just don't think they
00:01:01
should buy them. I think it's
00:01:02
fascinating to see the logic to see what
00:01:04
it comes back with to start asking it
00:01:06
more questions. I mean, you should all
00:01:08
try it out to see what happens and do
00:01:10
one of those tests. But this was a great
00:01:12
article in the journal and I think it
00:01:13
showed that it's not as ready for prime
00:01:15
time as they say.