undefined

points

[-]

Using AI to actually learn is indeed very helpful I've found... I just don't think most people use it for that. I recently wanted to build and train my own (albeit, small and simple 10M param model) and used AI extensively to explain concepts for me, explain lines of code, and generate in depth visuals I can use to get a visual intuitive understanding of what is happening under the hood. I think people who have a natural curiosity to understand the why and how of things benefit immensely from it, but I do admit it is very easy to just offload all thinking to it instead of asking it why something is happening. I could have just asked it to implement the entire model using PyTorch and just ran the CLI command to begin training on some auto-generated dataset, but intentionally struggled through it. Actually learning from it requires intentionality for sure.

by warshinder5 hours ago|

prev|

[-]

It’s just so sycophantic though.

by zloy885 hours ago|

parent|

[-]

Lol, wdym. I like to prompt it that I want evidence, so when there is any new topic where I currently don't know how to tackle it myself, I ask that it does a research and give me evidence. So i can check the sources myself. Like I said, I let it spoon feed me.

When I know upfront how to do anything, I just give all the instructions. But the OPs point was If we offload thinking too much, so that's why I was just thinking about this example when I need thinking - that's usually when I need to learn something new.

by al_borland4 hours ago|

parent|

[-]

I've found it present wrong, skewed, biased, or old information. In other cases, it seems to ignore the elephant in the room. I find all of these things problematic when trying to learn something new. The less I know about a subject, the more susceptible I am to being mislead.

When I bring these things up, it will apologize, tell me I'm right, and adjust. But what if I didn't know enough to question it? Approaching from the other angle, maybe I'm actually wrong and it's a sycophant, as mentioned, trying to please the user?

On the topics where I'm having to correct it, I probably shouldn't bother asking in the first place. On topics where I'm not correcting it, is it because the AI was right, or I just don't know enough to call it out? This kind of thing worries me about AI being leaned on more and more as a teacher.

by zloy8838 minutes ago|

parent|

[-]

Never rely on the trained data. I always ask to do a research which means, it pulls fresh data from the web. Another thing I enforce all the time is simply "don't make assumptions, give evidence" - that basically means it has to prove from where it takes it's statements. I do pretty fine like that.

by Jyaif5 hours ago|

prev|

[-]

If you build stuff with AI it's different. It's very tempting to defer many (too many?) decisions to AI.

by zloy885 hours ago|

parent|

[-]

I do daily. So far I am doing fine, not sure what exactly your point is, sorry.

by user439285 hours ago|

parent|

[-]

Same here.

If anything, I think most here outsource too little thinking to AI.

What am I supposed to be afraid of? Losing skills I no longer need to get the job done?

by zloy8835 minutes ago|

parent|

[-]

Yea, kinda agree. I feel like my job just changed into a solution architect and quality assurance inspector.

by Jyaif1 hours ago|

parent|

prev|

[-]

For one, as of today SOTA LLMs sometimes take the wrong decisions

by zloy8837 minutes ago|

parent|

[-]

For me thinking and making decisions are two different things. I rarely let AI decide, unless it's something i do in my private time. But at work, I am the expert and I maximum let AI give me options, but I decide.

by brazzy5 hours ago|

prev|

[-]

This is also now the first thing that I find it truly useful for in a non-coding role: researching how to do things in Azure, which I have not used before.