Recommendations For Noob In Data Degree

Interesting. Thx for info!

My degree is in data science. But some of the courses seem to be ‘related’ to data analysis. Ex: A big data class I’m taking rn have Business Analysis majors in it too, but my pure data sci classes are only data sci people.

Right now AI can probably replace interns, you know, those "I am a student who is willing to work for free for experience" kind of people. Yeah, I would rather use AI for that.

It can't even fully replace junior level employee atm.

And even if AI replaces all the junior position jobs one day, imo it is a good thing anyways. Way too many junior level jobs are just repetitive easy brainless tasks because someone has to do it. And it is more efficient to delegate easy jobs to the cheapest employee.

This kind of easy job is wasting smart young people's time and talent imo. They are forced to do easy tasks because they are young and cheap, not because they have no talent.

I wish AI can eradicate all easy low level jobs and let everyone start as mid-level regardless of age. So smart young people get to use their talent on higher level tasks. Imo that's a better use of people's talent anyways.

I'm not vouching for synethetic users. Today I wouldn't advise using them, but as I've said.. people said AI [insert anything] was ***x years/months ago. The worst mentality to be in is "it'll never happen". Ignorance like this has caught millions of people out.

Also FYI, I've corresponded with Don Norman (of Nielson Norman) over the years, and he added me on LinkedIn (this would be like, idk, Bill Gates adding you if you're in programming).

He's ok, but he's known for pushing his preferences as objective and everyone should follow it.

I see people claiming this, and I agree that LLMs can do intern-tier work, but interns aren't actually that useful and coming up with projects for them is a PITA. At the end of a successful intern project, you have someone who has learned something and you can maybe develop further into a useful employee. At the end of an LLM intern project, you have an LLM and still need to figure out how to idiot-proof its next task.

Also, LLMs cannot replace even mid-tier devs right now (or upskill low tier devs) because the fundamentals that those devs lack make them incapable of validating the LLM's answers. I have some junior devs at work right now who keep opening PRs mostly made by LLMs and it is painful AF.

I don't think people realize that intern programs are really just ways to filter out prospective new hires for junior positions. If the person is useless you can just opt not to hire them and not worry about employee termination and potential discrimination lawsuits. Contract to FTE positions are similar but for more mid range or senior positions.

My personal philosophy is, prediction without timeframe is pretty useless realistically.

I can easily say "AI will eradicate all jobs in the future" and I'll have a good chance to be correct. Because I didn't include time in the prediction. Is it happening 5 years later? 30 years later? Or 500 years later?

If humanity still exists after 500 years the society is probably going to be very different from today's. A world with AGI is very likely to happen in the future too. Except none of that will affect me 500 years later.

Even if AI replaces 80% of jobs after 30 years, it still doesn't affect me. I don't have plans to apply for jobs at age 70 lol.

What matters to me is problems that I face immediately. My deadline next month, or business revenue growth next year. Those are far more important than technology 30 years later.

So personally, I don't see a reason to worry about things that far away. Life is too short for that.


As to OP, their immediate problem is probably graduate first. They can always worry about jobs later. Your degree don't really matter for jobs anyways. Plenty of people do jobs that's irrelevant to their degree. So why worry about jobs and AI this early?

Because it'll cost a lot of money and take a lot of time? idk

Its so refreshing being free from data science, finally.

This morning I restarted my tutorials on Angular, started a book to learn C++, and during the day I will review Git and Github plus do some progress in that UX course I mentioned.

Life is so good

I almost enrolled in a data science 2 years course, just didnt because they screwed up with my payment, but now Im grateful it didnt work.

Now I can learn something better like german next year.

So what's your solution? Because AI will take away jobs after 30 yeaes, so you sit at home and do nothing because the investment isn't worth it in the long run?

We may as well say all investment isn't worth it because we'll eventually die and take nothing away with us lol.

It's ultimately a pessimistic pov that does nothing productive imo.

I've given my advice multiple times in this thread. You only don't care to see it because you want pointless arguments on the internet because you are Afania. Try rereading the thread all over again and if you're still struggling to see where it is I'll quote it for you.

And now you are taking this way too personal.

I already read your "advice", I personally don't see value in it....yet, not because I "want" pointless arguments. As plenty of people pretty much said the same thing I did. Saev/Thorny/Nif/Panta already made the exact same point that I want to make towards your "advice" on jobs and AI issues. Do they ALL want pointless arguments? Why would they even do that?

If so many people disagree with your "advice" and your pov on this issue, then maybe it has nothing to do with people wanting an argument, but because of the problems within the "advice" itself.

Don't taking things personal. Look at the value of an opinion objectively and discuss that.

Calling someone Afania, he didnt have to offend that much.

There should be a limit of how low we can get

I'm not taking it personally. You asked what my solution was. That either meant you didn't read my posts, cannot comprehend my posts, or were ignoring the fact that I had suggested multiple times that I think he's be better off developing in the direction he already had a background in to misrepresent my position, which you did. This is called creating a strawman argument.

You disagree with my advice - that's fine. You are free to do that. Congratulations?

Thorny agreed with me, but ok? Even if he didn't that's nothing more than argument ad populum.

Please don't waste any more of my time. I will talk to the adults.

Huh? When Panta said they are learning UX your response was a dismissive "UX design will all be 90% automated with AI in a few years and the bottom end of UX designers completely wiped out. Not a hugely safe bet either IMO."

Then when Panta pointed out UX still has values in the industry for tech (which I 100% agree, Ive used way too many tools with terrible user interface that I wish every programmer knows UX), your response was an irrelevant "synethetic users" and "human aren't unique".

Then I made a point that synthetic users does nothing for UX atm, your response was "Today I wouldn't advise using them....but The worst mentality to be in is "it'll never happen".

Which follows up with a even more dismissive "it cost a lot of time and money"!

I am sorry but, you are the only one who continue to lead the discussion to an irrelevant dismissive direction, dismiss every post with "AI good, jobs aren't safe" while offer nothing to solve the actual problem currently existed.

You know, real problems like programmers made things with terrible UX and made my life hard, which needs to be solved now. Why did you dismiss tech people learning UX to improve user's quality of life if you can't even solve this problem with your "advice"?

You are right about one thing though, this conversation is a waste of time because it goes nowhere.

I thought we were talking about the OP, not the derail about UX design. If we are talking about the derail then none of the people you refer to have opined on UX, so you still need to learn how to read and formulate rational arguments.

I also teach on the design course from which our students get all of the best internship and graduate jobs in UX design in the UK, including teaching user research, personas, user journey maps, and UX theory and practice (including Figma), etc. I appreciate you trying to educate me on UX design by posting one person's position whose entire career is based on telling people to work with people, but I think I know enough to have an opinion of my own.

I didn't "educate" you, none of the post that I made is about how to make do UX correctly. I only referenced one post about the usefulness of synthetic users in UX(which is barely educating), and even you agreed with the post.

I only said your posts are dismissive by nature and I challenged such pov, that's the only point that I made.

If you post "best practice to learn UX" when somebody said "I am learning UX", it would have been way more helpful than "UX will be replaced by AI in the future, it's a waste of time" kind of dismissive attitude.

You are an educator, no? The role of educator is to help people learn and encourage/help them to learn better. not telling them don't learn.

It seemed like you were trying to say "ur wrong, because of this". We're not in agreement about what the long term implications are though.

I don't mean to, but I'm autistic.

The thread is about future proofing by making a career switch, I'm only answering along the lines of that subject. I didn't say learning UX was a waste of time, just making an informed suggestion.

I haven't told people not to learn. I have given my opinion on learning specific things in a specific context (making a career switch at middle age). As I've said, I think learning about people would be an area I'd promote in the specific context of the discussion. Be that pscyhology, sociology, STS, etc.

I also teach on Graphic Design degrees. I would not advise someone to get in £50,000 of debt to get a degree in Graphic Design for career reasons. If someone wants to do that to pursue a passion or to spend a few years having a great time, I would fully support that.

Huh? In jobs/AI issues my stance is exactly the same as Thorny/Saev/Nif's. But I didn't participate in the discussion.

I referenced their posts because I disagree with some of your pov on this matter, while agreeing with them. But I am not interested in repeating what Saev/Thorny already said.

You accused me for creating strawman arguments, I thought you meant my posts in the UX discussion is strawman argument. That's the only discussion that I participated in.

Well this was my original comment, for reference.

I was being provocative but I genuinely don't think anyone starting from scratch today will be able to catch up with AI in Python coding in the relevant time frame. Even with 2 years of training I don't think most people will be as good as how even open source LLMs will be in 2 years time.

Any human? Maybe not, but I wouldn't say anything is certain now.

I see.


I do agree with this. One should never expect a degree is future proof, with or without AI.

Ehh...

Good sir, I do not think you know how this works...

Still waiting for you to tell us how you went from mechanical engineer to being an expert developer in a couple of years.

Yeah, we all know you're a clown, no need to remind us.

And problem solved.

Man, I can not wait for the "AI" bubble to pop. *** Sam Altman with Roko's Basilisk's ***.

LLMs with carefully constructed datasets given clearly defined, limited-scope tasks can be and are very useful in a number of applications. If you're doing anything more than that with it, your output is all but guaranteed to be trash, and you aren't even learning anything so you can get better the next time.

Oh yeah, ML is extremely good at finding patterns and anomalies inside curated datasets. Having is parse over everything in your documents folder would let you use natural language to ask it to either retrieve information in there rather then me hunting through the hundreds of technical documents I've already written. We use DataDog's Watchdog AI to monitor and detect performance issues within our production environment and it's predicted several outages and allowed us to remediate it before there was an issue.

It's stochastic data analytics on steroids. I think the current novelty is that applying that stoichiometry to our language allowed someone to build a chatbot that was a few orders of magnitude above the rest.

You can literally find thousands of such posts from artists 2-3 years ago, replacing LLM for CLIP or GAN in that statement. Oh how that did not age well.