Like the video? Support my content by checking out Interactive Brokers using the link below: www.interactivebrokers.com/mkt/?src=ptlPY1&url=%2Fen%2Findex.php%3Ff%3D1338 Follow Along With Written Version: hackingthemarkets.com/gpt-3-prompt-engineering-for-financial-data/ 0:35 Data Extraction 2:39 Named Entity Recognition 5:02 Keyword Extraction 6:30 Extracting Financial Numbers and Data 10:39 Date Extraction 11:38 Prompting For Structured Output (CSV, JSON) 13:50 Prompting with Examples For Output Formatting 17:38 Prompting Database Design 18:18 Generating SQL Statements From Extracted Data 20:53 Sentiment Analysis of Fed Speech Transcript 23:57 Text Classification as Hawkish, Neutral, Dovish
@aiautoglasscrm
Жыл бұрын
Your videos are filled with awesome ideas. Learned that I can request json formatted data and so much more.
@RainyDays4079
Жыл бұрын
Thank you so much for the information cause I was a bit worried about this prompt engineering career. you made it easier for me to understand. God,bless please share more video’s.
@eleefant3873
Жыл бұрын
this dude can't be stopped!
@GeorgeVenturi
Жыл бұрын
Wow, this is incredible. >>> print(“thank you”) thank you
@NeferLopez
Жыл бұрын
You are the MAN! Thanks so much for your brilliant videos! 🤘
@quantontheway
Жыл бұрын
Thanks Larry. Looking forward for the upcoming videos.
@sriramkrishna6853
Жыл бұрын
Yayyyyyy! More Larry OpenAI stuff. Letssss gooooo!
@jserpapinto
Жыл бұрын
amazing, thanks!
@mohamedalitoure9696
Жыл бұрын
thanks Larry 🙏🙏
@prabhacar
Жыл бұрын
Great demo/presentation! Thank you
@kingtrippy5006
Жыл бұрын
Insanely helpful sir
@eTaupe
Жыл бұрын
Really love your channel Exquisite Keep the up trend
@csuliyu
Жыл бұрын
love your channel
@nodesignlaws
Жыл бұрын
amazing, thanks
@bonadio60
Жыл бұрын
Fantastic, the json output is great!
@mr67475
Жыл бұрын
bro this stuff is amazing, thank you so much for sharing and explaining
@peralser
Жыл бұрын
Thanks for sharing this very great and value content.
@bakari007
Жыл бұрын
Very interesting ! Great video ! Impatient to see the next one ! Thank you ! 🥰
@groundedinfirstprinciples383
Жыл бұрын
Absolutely what I've been looking for.
@lifsys
Жыл бұрын
Awesome as always, thanks for the content!
@jerrywang3225
Жыл бұрын
These openai episodes are getting better and better, just keep them going. Thanks
@mohsensafi7778
Жыл бұрын
Great video as always😍👍
@KuriousBat
Жыл бұрын
Great content... Keep up the great work!
@wayneqwele8847
Жыл бұрын
Thats crazy, that is way better than most popular sentiment analysis techniques available. And way easier too.
@linghaoyi
Жыл бұрын
Thank you for your excellent tutorials. Merry Christmas and Happy New Year!
@bocobox
Жыл бұрын
Thanks for the video. it's very clear and informative. really got to how to use ChatGPT with prompt.
@jazzorcazz
Жыл бұрын
very useful channel, love the ideas and how you present them
@hectorniyan
Жыл бұрын
This is awesome, Larry. You are a genius. Thank you. Always learning.
@jeffreylabranche3437
Жыл бұрын
Thank you Larry for the amazing content! 👍
@user-qx8wv2zi1x
Жыл бұрын
🔥🔥🔥🔥🔥🔥🔥🔥 TYVM Larry.
@CuriousityExpress
Жыл бұрын
You just shattered what I thought were the limits of gpt-3's responses w/o training! Great job man keep it up :D
@TheJustinjet
Жыл бұрын
Excellent work!
@tak68tak
Жыл бұрын
amazing as always. can't wait for the next.
@WorldWideSk8boarding
Жыл бұрын
This dude might be the best on youtube More 👏 finance-related 👏 python 👏 stuff 👏 please This was an awesome playlist
@RickHunter-fz7oh
Жыл бұрын
Just discovered your channel and suscribed, exactly what I was looking for. A channel that discusses OpenAI in english. I appreciate the fact that you take time to explain, every single line of command, it really helps. Some time ago, I read that some big Real Estates companies were using AI to automatically scan and filter house oppportunities. It seems that OpenAI has leveled the playing field for individuals now.
@sportfoliokings
Жыл бұрын
what's frustrating with so many newer NLP instructionals on yt is everyone just uses expensive models to get great outputs as if that is the optimal solution for actual practical, repeated use. most youtubers only make a token small reference to cost. sometimes there are slightly more difficult approaches to accomplish a task and get a slightly less robust output, but running the program is free! i hope you will take this more into consideration in future videos.
@dservais1
Жыл бұрын
3rd to like. Larry is back 😀😀😀
@ozorg
Жыл бұрын
great stuff!
@keridince
Жыл бұрын
YOU ARE A FUCKING GENIUS LARRY
@r2rmi11y7
Жыл бұрын
Love the finAI content!
@taijingshen4503
Ай бұрын
Awesome video! I love how you explain it in a very easily digestible manner. One question: under what kind of circumstances from your examples, we will consider using Chain of Thought (CoT) Prompting techniques over ICL?
@J2897Tutorials
Жыл бұрын
You could probably use virtual audio cables for the live analysis. That might work cross-platform. I used them years ago to feed Google Translate my PC's audio instead of my Mic to translate what foreign live streamers were saying. It was kind of cool at first, until I realised how poor Google Translate actually was. 'Very awkward via the website too.
@dejohnny2
Жыл бұрын
You are on to something.
@techsavvy9258
Жыл бұрын
Amazing. Any chance of handling multi classification for reviews?
@yellowboat8773
Жыл бұрын
No idea about python or coding for that matter, but man this was interesting. Looks like learning to code today you are at a big disadvantage
@Shiffo
Жыл бұрын
So, Theoretically you could use GPT-3 to scan in invoices and use named recognition to recognize the supplier name. Based on history on the supplier name you could prepare a journal entry and automate a journal entry. Just thinking about ways to scale purchase invoice entry.
@edrodven
Жыл бұрын
Excellent video have you heard of the XBRL and used to extract the information for analytical purposes?
@almostworthy2973
Жыл бұрын
Hi Larry, the video is amazing. What I was looking for is something different. Is there way where I can ask GPT3 to summarize a article via an API and will it have the memory to remember that earlier was the first article sent for summarization and the second prompt(without actually passing the article) is asking to extract some specific information from the same article. Just to summarize the flow: API Call 1: Sends a prompt with command "summarize the article below "+ attachedArticle. API Call 2: Sends a prompt asking to extract financial numbers from the previous article. (Note this time I won't send the article with the prompt.) Please do let me know if such scenario is possible?
@RaoVenu
Жыл бұрын
I find it interesting that the results from the classification results are not predictable / consistent across multiple executions. I tried the exact query as you did in 4:44 with text-davinci-003 model and first run I got people names under job titles and no People classifcation. Retried and got same results. I changed to a different model and completely out of whack results. Then switched back to text-davinci-003 and got the same results as you. So why didn't I get the same results during the first 2 runs? 🤷🏻♂
@user-sv7bx6gn8n9
Жыл бұрын
Hi, I have a question about real-time trading: is that possible to scan whole US stock market data (price and volume) every minute (or even every 5 sec)? If possible, how to achieve this?
@parttimelarry
Жыл бұрын
I have many videos on real-time stock data, check some of my broker API tutorials - Interactive Brokers, Alpaca, Ameritrade, etc.
@Shiffo
Жыл бұрын
Is extracting the keywords from your KZitem script a good way to add Tags to your KZitem video? (I know that Tags are not so revelant to KZitem anymore, but seems like a cool way to implement GPT for KZitem)
@Andersson584
Жыл бұрын
Where can I find the colab nb?
@parttimelarry
Жыл бұрын
This video didn't have a colab notebook, I linked the writeup with all of the prompts on the website, both in the description and the pinned comment.
@disservice
Жыл бұрын
Dude you're awesome! Btw I am looking at your webside hackingthemarkets and want to make it my parttime/latenight project. And hints? Maybe some high level keypoints so I can make my own curriculum and get grinding.
@rantg
Жыл бұрын
may I ask what you do for a living?
@parttimelarry
Жыл бұрын
I am currently unemployed
@rantg
Жыл бұрын
@@parttimelarry do you trade? i follow you for a while and you seems to know a lot by now.
Пікірлер: 59