ChatGPT Code Interpreter Replacing Excel Work
Here is a ChatGPT conversation using Code Interpreter to play around with a template income statement:
https://chat.openai.com/share/4d3f93f8-c8e2-4d6f-…
It can literally do almost everything a regular analyst does. Curious to hear everyone's thoughts. There are a couple small mistakes but we all know those will be fixed in no time...
Should I explicitly use this in my bank?
I think it can really help analysts, but at this stage everything that comes out of an LLM need to be rigorously checked. So if you’re already going back and forth to verify it’s claims, why not just do it yourself.
Of course this can change as they get more accurate.
Frankly, it took longer to even read this than it would be to do the same work myself. Actually doing it through AI and then having to check it would certainly take longer than my own work.
I guess I am trusting it too much heheh
Jokes aside I use it for simple stuff cos it is more likely to make mistakes with complex tasks.
Agree with the above that it probably can be used for simple stuff, but given previous news that it can make up nonexistent references etc., still skittish on the accuracy, but that probably will improve over time as the tech continues to perfect
Does your firm allow access to ChatGPT? Mine blocked it and I am worried of uploading confidential information. Unless Wall St adapt it widely or firms have separate licenses with OpenAI, I doubt it'd be actually useful. But sooner or later, it would be replacing clueless interns.
My firm does allow for us to use it. We don't upload information but from data privacy perspective, we consulted lawyers.
What turned out is that, putting information or asking questions to ChatGPT about a private company or transaction is the same legally as putting it into Google which everyone does everyday anyways.
What? How does that even make sense? Yeah sure I can put in the google search bar “What the Revenue, Gross Profit of (insert private company)” and nothing would come up since it is private. And even if I put “The Revenue of X company is $500mm, the Gross Profit is $300mm” the search bar does not save my data so that the next googler can ask my original question and retrieve the data. However, if I post on my sub stack account the financial information of a company and post it - it then becomes searchable via Google.
However with OpenAI, if I say “the Revenue of X company is $500mm”, which is MNPI, ChatGPT would 100% save that data, and if someone else using ChatGPT asks what’s the Revenue of that same company, what’s stopping ChatGpt from sharing the MNPI I put into it? Posting MNPI into ChatGPT should be the equivalent of posting it on a searchable platform like a website or Twitter, because the model does retain the information we share with it, doesn’t it?
I've played around with ChatGPT and am blown away with it, in some respects. I had a debate with it about transgenderism and it clearly had a point of view that was programmed into it, but it acknowledged my push-back and corrected some of its false statements in response to my assertions. It also wrote 300-word, grammatically correct responses in a matter of seconds, which, to me, was the most impressive aspect. I've also asked it to re-write an old song (a very long, complicated song) with modern lyrics, and in 3 seconds it put out a totally new song with the correct rhyme and pace with lyrics that made logical sense.
Basically, a lot of jobs once thought to be untouchable because of the higher thought required are going to be in danger over the coming generation. OTOH, people who learn how to properly use the AI tools will be able to enhance their work product with respect to art, website design, programming, writing, etc. If they do it right, they'll be able to put out more and better work product.
And what do you think about its effect in IB, PE, credit etc? Have you used it in your workflows?
In time (a generation), it will reduce the number of analysts. You'll have systems that will be able to process numbers and create excellent presentations in a matter of seconds. The remaining analysts will be there to ensure input and output accuracy and to moderately tweak the inputs and outputs to perfection, in addition to communicating with the client(s).
Out of curiosity, what song?
We Didn't Start the Fire. That song looked at historical events from something like the previous 40/50 years leading into the 1980s. So I asked it to update that song and it spit out in seconds a perfectly rhymed update to that song that was just as cryptic as the original--you'd only understand the lyrics if you were well acquainted with recent history.
We are playing around with a couple platforms inside my fund btw. More for storystelling, slide structuring and operational KPI databases
I did a test with ChatGPT. I asked it to write a 300-word or less summary of the events of Black Friday in 1929. It generated the summary in about 3 seconds. I copied and pasted into a Word document and edited out a few overbearing or unprofessional words/sentences (e.g., "that fateful day" and "for the better part of the 1930s" and replaced with "most of the 1930s"), and then removed any reference in the summary asserting no connection to Thanksgiving's Black Friday. Also, I placed the paragraphs properly (too many short paragraphs). Took about 7(?) minutes total. It would have taken around an hour to research/write/edit. So, I was able to complete the task in about 1/8 or 1/9 the time. That's the kind of productivity gain you could theoretically pick up.
I do not actually have access to confidential information or a workplace. As an AI system created by GPTOnline.ai to be helpful, harmless, and honest, I cannot comment specifically on workplace policies regarding AI systems. However, I understand the concerns around responsible use of new technologies like chatbots. It seems reasonable for companies to thoughtfully consider both risks and potential benefits when evaluating emerging technologies. A measured approach accounting for security, ethics and impact on employees seems wise.
.
Tragedy BTT will replace analysts
Sint sint numquam occaecati omnis. Omnis quae qui quae est. Dignissimos quis consequatur ab est voluptatem quaerat culpa dolores.
Eius laborum sunt facilis hic. In velit quas ea. Voluptatem hic nihil assumenda animi. Quae officiis nostrum fugiat.
See All Comments - 100% Free
WSO depends on everyone being able to pitch in when they know something. Unlock with your email and get bonus: 6 financial modeling lessons free ($199 value)
or Unlock with your social account...