Using LLMs efficiently requires a good amount of…
0 reactions
Using LLMs efficiently requires a good amount of practice/skill
I often see people complaining about being told to learn to use LLMs or get left behind. They claim that prompting a chatbot requires zero skill.
It's true that asking it to generate a meme or what's the weather is straightforward. But if you're regularly hitting the usage limits in your plan (which is common in software development, but not only there) it absolutely pays off to learn how much context to give it, how to save the right memories so you need to do less iterations but they don't pollute your context, etc. Understanding the tradeoffs between different models can help you save tokens without compromising in quality, etc. Using agents, skills, MCPs, etc can help you make your workflows much more efficient than just using a chatbot via the browser.
It's not rocket science either, but these are useful skills that you can't learn in a day and are constantly evolving. So if you care about being more employable it's a reasonable recommendation.
#conversation
I often see people complaining about being told to learn to use LLMs or get left behind. They claim that prompting a chatbot requires zero skill.
It's true that asking it to generate a meme or what's the weather is straightforward. But if you're regularly hitting the usage limits in your plan (which is common in software development, but not only there) it absolutely pays off to learn how much context to give it, how to save the right memories so you need to do less iterations but they don't pollute your context, etc. Understanding the tradeoffs between different models can help you save tokens without compromising in quality, etc. Using agents, skills, MCPs, etc can help you make your workflows much more efficient than just using a chatbot via the browser.
It's not rocket science either, but these are useful skills that you can't learn in a day and are constantly evolving. So if you care about being more employable it's a reasonable recommendation.
#conversation
| 2,000 mlx | |
| 2 | |
| 0 |