End of Data Scientists? ChatGPT Code Interpreter KILLS it!

[@shotelco]

One can NOT replace a "junior" Data Scientist! Or at least that would not be a good idea. Why? Because a "junior" (entry level) data scientist evolves into an intermediate data scientist, then into a Senior Data Scientist. If one foolishly goes down this path, the end of the road will logically be that one has NO data scientist at all - as they would have all left or retired - and thus one must depend entirely on ChatGPT for their business to function. I can not impress enough that we all have seen this model before in Social media and Cloud. Example #1: Once a organization gets into bed with Facebook advertising, which starts off "cheap", then as the business grows based on FB ad's, FB charges more, and more... FB knows the entire business model _Depends_ on FB advertising. We have seen companies pay out 40% of their _Gross Revenue_ to FB in ad expense. These companies became indentured servants to FB. Example #2: Once an organization moves to Scale Cloud platforms (AWS/Azure/GCP), initially assuming these platforms are "cheaper" than hosting in-house or renting data center space, the Scale Clouds will encumber the company with highly proprietary functionality components (i.e., DNS, security, serverless, etc.) that can not be easily migrated out of. Then the Scale Cloud Platforms begin upping the expense - like an Eastern European ransomware criminal ring - because they know the customer is locked into that single provider. This is a FACT proven by how competitive Scale cloud providers will offer $100's of thousands in "credits" to migrate from one scale cloud into their cloud. M$oft didn't invest $10B to give shit away for free (or $20/Month). They did it to sucker low intelligence executives into a web of perpetual business slavery. Where ChatGPT not only steals all of a companies secrets and business plans (to resell to the highest bidder), but strips them of any viable alternatives.

[@agiengineer]

Lol 😅😂😅😂😂😅🤣

[@Tate525]

You underestimate businesses ability to undercut employees and secure profits.

[@cassi1564]

This is so frustrating for someone who's been studying data science like me...

[@mranotherside22]

yea bro, i just start learn this and now...

[@chats-bug]

I don't think they will be replacing junior data scientists anytime soon. Reliability is still a big question mark here. However, junior data scientists will get much more productive with these tools. How that affects the overall job market is yet to be seen.

[@1littlecoder]

I like the optimistic view. Thanks for sharing

[@b326yr]

@@1littlecoder lol, he is indeed Optimistic

[@chats-bug]

@@b326yr It’s more realistic than you think. Just go to any company and ask them how many people has GPT replaced, I found none.

[@muhammedvaseem8570]

​@@chats-bugTrue

[@muhammedvaseem8570]

​@@chats-bugBut the concerns are about the future.

[@Glorytroly9092]

Thanks for detailed video 😭😭😭😭 but my career is dead before even starting it😭😭😭

[@manjz7hm]

Fuxk man I spend whole year and spent tons of money to acquire this skills , now doing it all in a second .

[@KitcloudkickerJr]

i've got it analyzing time-series datasets from metatrader and even trained models on the data all within gpt4. its wild

[@arinco3817]

I think off the back of this it would be useful for pro analyst to work through some analysis using code interpreter as the tool. That way we can all level up really quickly

[@Curioushaidar]

should i droupout from data science

[@shouldb.studying4670]

The magic analogy is good... now you just wave your hands, say the incantation and the data is transformed!

[@anontechno470]

Can you make a demo of it is programming capabilities and all uses cases ? is it like copilot ? Thanks !

[@serta5727]

That is really incredible 🤯

[@xXWillyxWonkaXx]

I agree with @sukritchatterjee196. When you say Data scientist I believe you're contextually referring to a niche group of specialization of people which are the Data Analysts. Do you think this will be the case for Machine Learning Scientist/Engineers or DL/NLP Engineers? If so, would be great to hear your insight

[@southcoastinventors6583]

Would a junior data science really have a problem coming up with those key points seems pretty straight forward to me based on the data set maybe I am data science and didn't even know all though it is way faster so glad we finally get access to it.

[@dreamphoenix]

Thank you.

[@travelerx-wn3dp]

can it build model?

[@1littlecoder]

As you saw in the video demo, it built a model.

[@HaseebHeaven]

And Noteable plugin also was available for months it has internet access and can install new python packages as well and beat Code interpreter easily for every task.

[@theawebster1505]

The problem with GPT-4 + Noteable was that GPT-4 Chat was designed to have human-like interaction with creativity. But that creativity was playing a bad joke with programming and data analysis. For example, I asked it to analyze a table and it constantly forgot that my data column was named "Start Time" and not "Start time" and threw errors. Also it constantly messed up and forgot stuff that it did correctly 1-2 steps ago. From my experience Wolfram as a plugin fixes this and I am excited to test if the Code Interpreter is too a more "mathematical, strict"-oriented version of GPT.

[@HaseebHeaven]

@@theawebster1505 Code interpreter has drawbacks it also often forgot that it has some python libraries installed and denies your request

[@nic-ori]

Thanks.

[@1littlecoder]

You're welcome

[@nirbhaysedha8541]

Doing a good job 👏

[@mallow610]

The ending tho 😂

[@knutjagersberg381]

Though I don't doubt OpenAI created a great tool here, by the way the models work, we can't know for sure if it understood us properly or if it makes errors in the implementation. You literally have to check the code, at each step. As data analyst, one can not tell their stakeholders 'properly correct facts'. It's way worse considering people who now think they don't need professionals for this. Copy paste results and publish it as truth. This is again OpenAI showing, they just throw things at you, not thinking what the words "responsible AI" could mean. And I'm really quite liberal on using data. The word data means it's about facts. Originally, it was written on paper before it was written into stone. Mistakes were punished.

[@alexis91459]

game over

[@holdthetruthhostage]

Omce it can alter an entire project zip folder than bam

[@1littlecoder]

Did you see this? It can generate an entire project folder. kzbin.info/www/bejne/b2q9op9vlLVnqLM

[@holdthetruthhostage]

@@1littlecoder I'm saying feeding it a game maker studio project & it just edits the code

[@manishagarwal1988]

AI has been existing long enough and for those contemplating AI is going to kill jobs, it will create more jobs. No job is under danger atleast for a foreseeable future so please stop spreading these unverified news/rumors and stop creating panic situations. I can say this because I have more than 10 years experience in AI field.

[@samuelramirez8965]

Thanks

[@Tate525]

Tum apni biwi bachao chicha warna ChatGPT woh bhi leke jaega hahaha

[@UncleDavid]

oh wow FOMO regurgitating “it will take your job” nonsense lol

[@twosatoshis2478]

What jobs will be left lol

[@rmp5640]

People talking so much crazy shit 😂

[@ajithsasidharan6074]

bro big fan from kerala , i want build something in AI. but i have no coding / ML skills...

[@1littlecoder]

Check this out bro kzbin.info/www/bejne/qpnUgWd_brONotU

[@1littlecoder]

Check this out bro kzbin.info/www/bejne/qpnUgWd_brONotU

[@akmalmir8531]

While AI is certainly capable of automating many of the tasks that junior data scientists perform, it is important to remember that AI is still a tool. It is up to humans to use AI effectively, and this requires a deep understanding of the data, the business problem, and the goals of the project. Junior data scientists bring a unique set of skills to the table that AI cannot replicate. They have the ability to think critically and creatively. They are also able to learn quickly and adapt to new situations.

[@carkawalakhatulistiwa]

From 10 human not only 1 human

[@Tate525]

Copium

[@akmalmir8531]

@theredditexperience1 Nope, chatgpt with human feedback