Exactly. Also states that today's SOTA models don't require specific prompt structures. Then the demoed Anthropic prompt has the typical Anthropic XML style markers such as .
True ✅️ 🤣🤣 Prompt engineering is not going anywhere atleast for next 10 years. Infact Domain Knowledge + Prompt engineering is going to be a basic requirement
Yep! That's what I got from this video. Don't prompt engineer, but engineer prompts like I'm doing here. Also, thinking clearly is a vague-ass skill to be taught in an 11 minute video.
I actually love that I came across this. I have an interview soon for prompt engineering, and frankly what I do with AI is communicate with it, to get outcomes and outputs I want. It take trial and error, but for me the conversation helps me with all the skills you mentioned; thinking clearly, asking better questions, figuring out exactly what I want/need.
Just so everyone knows, prompt engineering, whatever you want to call it, is actually very important right now. Whether you're dealing with image generation, video generation, llms etc. You have to be able to know how to break down a problem and word it to the AI. These AI systems are not to the point where they can pick up on minor little linguistic nuances that we take for granted. In a lot of cases you really need to break it down step by step in your prompt what it is You want. So yes this title is very much clickbait.
What you are saying is so clear to me now. Thinking about it, for instance, if I were to write Python by myself, the programming would have to be correct, but more importantly, clear to get the program to output exactly what I want. The best part I find with AI is that I actually enjoy the back-and-forth to get to a satisfactory conclusion.
Jordan Peterson Is the subject of many UA-cam videos in which he describes how to be unstoppable through clear writing and clear thinking. He emphasizes Unstoppable is due to focused rain and clear, strong articulation of what a person means. Don't obscure our intended meaning with jargon or "fluff." Recall the mantra of the TV show "Dragnet" -- "Just the facts; we don't need a story." Generative AI needs "Just the facts" of critical thinking to produce the content we need and want 🥇
Thanks for watching and taking time out of your day to comment! Kind of missed the point of the video, eh? Feel free to shoot over any questions, happy to simplify even further for ya. 👍
there is a difference between prompt engineering (write better understandable prompts for LLM) and improving the self ability to express information, huh
Today, Prompt Engineering is less about a hard skill to understand the machine and more about the soft skill to properly comunicate. Know what and how to ask still important, but as the tech advances, the machines got better in understand the nuances of the prompt.
I believe a hybrid method is the most useful way to go about it due to how easy it is to cause the AI to get confused in the information between sessions. Also the more back and forth the easier it is to accidentally cause AI to make incorrect associations and return less if not completely useless responses.
Is it? I look at prompt engineering as refining a prompt in a specific format (ie: chain of thought, or multi shot) and formatting as such. The point is that needing to understand these skills is becoming less and less important.
@@RickMulready I'm new to the game and trying to understand. I see your rationale, but I'm confused as to why these skills are becoming less and less important. Can you explain? Thanks for the great video.
C'mon Rick. Yes, prompt engineering is evolving as the models evolve. Just as communication is important between people, communication is important with machines.
And NVidia CEO thinks we should stop teaching kids how to program. These wide sweeping bold statements and claims are so click baity. So very click baity.
Nice one, Rick cool video. I think there's a few things you missed here like prompting frameworks that help create better output depending on the task. For me, I realise that their will always be new prompt generators but what I aim to do with what I'm building is not only create a prompt for the user but educate them on why that specific prompt is better for their use case
Appreciate it. I do hear you and right now having a basic understanding of prompt engineering is helpful, but moving forward it will be less and less of a necessary skill as we’ll just be able to ask a (specific) question and the models will be smart enough to do it. 🙌
@@RickMulready I hear what you're saying, AI is getting much better at understanding what we want. Although, even if there is some structure or format used then the AI will be able to provide closer to what we want and quicker without having to re prompt it
In my Podcast on Spotify - The Deep Dive - we've included some comments about this concept. Episode title: Do I Need to Learn How to Prompt AI? - - The show notes include 8 important questions (w/answers). I like the idea of the debate. We need more debate.
Video title is a bit misleading. The message is to use the tool, we don't have to know how to create good prompt. Thanks for info about anthropic site.
@@RickMulreadyit’s not clickbait. It’s just a very debatable point. I don’t totally agree with you but great video It’s true that for simple and maybe mid complex task you don’t have to use prompt engineering but for more complex tasks like making a model act in certain way and follow a rubric, Prompt engineering is required for good results
Yes I used prompt engineering AI tool and it was really helpful instead learning to save time but still something is wrong and I don't get the result that I am looking for. It takes a human :)
No. See, if I am able to clearly imagine and articulate my question then why would I need another AI to generate a prompt to put this prompt in other AI to get the final output? So, you are saying we need to use AI to use AI.
Definitely clickbait and quite inaccurate. If you really knew what prompt engineering was at a higher level, you would understand how powerful it is… Even with these new models. As a matter of fact, it’s even more valuable.
Actually, you are teaching back how to prompt the prompt?! Sounds like zero content here to stop learning prompt. That is why we say "prompt engineering" when using Ai tools.
Respectfully, you don't know what prompt engineering is if you think your methods are something different and novel. Literally everything you did here is prompt engineering. Listen, at a high level all prompt engineering is, is being strategic and cognisant regarding how you communicate with the LLM, so that outputs delver what you actually want. You haven't discovered fire here, and nothing you demonstrated is remotely new. Again, you're literally doing the very thing you claim not to be doing, which makes you look and sound foolish.
That's why i don't follow the crowd you all don't think like a human. Your talking to an entity humans are slow in thinking he will not progress past prompting. Think about it you can have another ai write the prompts 😅😅😅 they waisted allot of time assuming
@CsabaTothMr Clearly missing the point of the video, which aligns with my opinion and thus the title and thumbnail. I appreciate you taking time out of your day to watch and leave a comment 🙌
✅ Join my AI for online businesses Skool community: rickmulready.com/skoolp2
“Stop learning prompt engineering…”
Goes on to teach prompt engineering.
Exactly. Also states that today's SOTA models don't require specific prompt structures. Then the demoed Anthropic prompt has the typical Anthropic XML style markers such as .
Spot on!
True ✅️ 🤣🤣 Prompt engineering is not going anywhere atleast for next 10 years. Infact Domain Knowledge + Prompt engineering is going to be a basic requirement
Yep! That's what I got from this video. Don't prompt engineer, but engineer prompts like I'm doing here. Also, thinking clearly is a vague-ass skill to be taught in an 11 minute video.
Yup
Stop mastering driving, focus on being better at braking, shifting, cornering and avoiding hazards. Yeah seems like a good advice!
I actually love that I came across this. I have an interview soon for prompt engineering, and frankly what I do with AI is communicate with it, to get outcomes and outputs I want. It take trial and error, but for me the conversation helps me with all the skills you mentioned; thinking clearly, asking better questions, figuring out exactly what I want/need.
Umm, to me this sounds like strategic prompt engineering.
Click Bait
Just so everyone knows, prompt engineering, whatever you want to call it, is actually very important right now. Whether you're dealing with image generation, video generation, llms etc. You have to be able to know how to break down a problem and word it to the AI. These AI systems are not to the point where they can pick up on minor little linguistic nuances that we take for granted. In a lot of cases you really need to break it down step by step in your prompt what it is You want. So yes this title is very much clickbait.
You can't say, "stop prompt engineering," and then tell me to do the thing that prompt engineering is.
What you are saying is so clear to me now. Thinking about it, for instance, if I were to write Python by myself, the programming would have to be correct, but more importantly, clear to get the program to output exactly what I want. The best part I find with AI is that I actually enjoy the back-and-forth to get to a satisfactory conclusion.
Exactly. 🤙
So what you're saying is we need to learn prompt engineering?
Ya, I'm that what it sounds like. The same thing with a different name using another tool.
you mentioned your free community? I can only find a paid community, unless I am looking at the wrong location
so basically we need to learn prompt engineering?
yes
So we have verbal prompting? Which is the direction of interacting with our modern computer systems.
Jordan Peterson Is the subject of many UA-cam videos in which he describes how to be unstoppable through clear writing and clear thinking. He emphasizes Unstoppable is due to focused rain and clear, strong articulation of what a person means. Don't obscure our intended meaning with jargon or "fluff." Recall the mantra of the TV show "Dragnet" -- "Just the facts; we don't need a story." Generative AI needs "Just the facts" of critical thinking to produce the content we need and want 🥇
I’m about to start learning prompt engineering. Thanks .
Nice, glad the vid's helpful!
It all makes logical sense to me as a creator of an Ai Music Channel. Thanks for your time and knowledge 🙏
What’s the link for anthropics console? There’s no link in the description.
Did you endup getting the link?
@@laurj00 yes thank you!
This guy advises against prompt engineering and then literally advises us to prompt engineer.😂... You can't make this sh*t up.
Thanks for watching and taking time out of your day to comment! Kind of missed the point of the video, eh? Feel free to shoot over any questions, happy to simplify even further for ya. 👍
there is a difference between prompt engineering (write better understandable prompts for LLM) and improving the self ability to express information, huh
@@RickMulreadyyou're bad. Really bad lol. Stop talking
Your sarcastic answer you makes you look even dumber. It's like you don't actually understand the term 'prompt engineering' yourself.@@RickMulready
isn't asking better questions what prompt engineering actually is?
correct. asking better questions and giving better instructions
Today, Prompt Engineering is less about a hard skill to understand the machine and more about the soft skill to properly comunicate. Know what and how to ask still important, but as the tech advances, the machines got better in understand the nuances of the prompt.
I believe a hybrid method is the most useful way to go about it due to how easy it is to cause the AI to get confused in the information between sessions. Also the more back and forth the easier it is to accidentally cause AI to make incorrect associations and return less if not completely useless responses.
Yeah! Excellent point.
Thanks for this interesting video @RickMulready! How to get a link for anthropics console?
Ok, but isn't clearly articulating your ask and defining the context the very definition of prompt engineering?
lol yeah, I got that too. we're right don't worry lol.
Is it? I look at prompt engineering as refining a prompt in a specific format (ie: chain of thought, or multi shot) and formatting as such. The point is that needing to understand these skills is becoming less and less important.
@@RickMulready I'm new to the game and trying to understand. I see your rationale, but I'm confused as to why these skills are becoming less and less important. Can you explain? Thanks for the great video.
First here, my friend... AI my new found love
Call me old fashioned but I will rely on my prompt engineering skills 😉
Prompt engineering is hardly old fashion.
He actually straight up taught better prompt engineering in this video.
Stop prompt engineering and start prompt engineering instead :)
Riick what do you think of the new Chat model ???
Check out this video I put out over the weekend --> ua-cam.com/video/Cq1OxaoKOzc/v-deo.html
C'mon Rick. Yes, prompt engineering is evolving as the models evolve. Just as communication is important between people, communication is important with machines.
Link in description, but you will never find link in description
And NVidia CEO thinks we should stop teaching kids how to program. These wide sweeping bold statements and claims are so click baity. So very click baity.
Nice one, Rick cool video. I think there's a few things you missed here like prompting frameworks that help create better output depending on the task. For me, I realise that their will always be new prompt generators but what I aim to do with what I'm building is not only create a prompt for the user but educate them on why that specific prompt is better for their use case
Appreciate it. I do hear you and right now having a basic understanding of prompt engineering is helpful, but moving forward it will be less and less of a necessary skill as we’ll just be able to ask a (specific) question and the models will be smart enough to do it. 🙌
@@RickMulready I hear what you're saying, AI is getting much better at understanding what we want. Although, even if there is some structure or format used then the AI will be able to provide closer to what we want and quicker without having to re prompt it
Very true!!!
In my Podcast on Spotify - The Deep Dive - we've included some comments about this concept. Episode title: Do I Need to Learn How to Prompt AI? - - The show notes include 8 important questions (w/answers). I like the idea of the debate. We need more debate.
We mention your video in the podcast.
Video title is a bit misleading. The message is to use the tool, we don't have to know how to create good prompt.
Thanks for info about anthropic site.
Is there an alternative to anthropic for prompt generation? It is a paid tool
I don't know what you trying to say
The title is clickbait
It’s exactly what the video is about. Appreciate you stopping by!
@@RickMulreadyit’s not clickbait.
It’s just a very debatable point.
I don’t totally agree with you but great video
It’s true that for simple and maybe mid complex task you don’t have to use prompt engineering but for more complex tasks like making a model act in certain way and follow a rubric, Prompt engineering is required for good results
@@RickMulready it's totally click. The models are smarter and our instructions are simpler, it's still prompt engineering 😂
Your comment is clickbait
Why is it “clickbait?”
Yes I used prompt engineering AI tool and it was really helpful instead learning to save time but still something is wrong and I don't get the result that I am looking for. It takes a human :)
No. See, if I am able to clearly imagine and articulate my question then why would I need another AI to generate a prompt to put this prompt in other AI to get the final output?
So, you are saying we need to use AI to use AI.
Share the link
What skills , please help comment
PE's will always be able to extract superior data or convince the model to take an action vs someone without PE techniques.
Over that last year I've noticed that LLMs want us to think more like kids, and that's okay for me.
Agree! Thanks!
"Stop Watching This Video"
So stop prompt engineering and instead better engineer your prompts.
Semantics 101
Stop means Start in this video ...and title ...
what you are describing is essentially iterative prompt engineering.
Definitely clickbait and quite inaccurate. If you really knew what prompt engineering was at a higher level, you would understand how powerful it is… Even with these new models.
As a matter of fact, it’s even more valuable.
...we should have paid attention to those English classes
So, you're saying forget PE in favor of PE, because what you're suggesting is PE.
awful
It's what you call 'opinionated charlatanism'
Be wise as serpents.
Actually, you are teaching back how to prompt the prompt?! Sounds like zero content here to stop learning prompt. That is why we say "prompt engineering" when using Ai tools.
Respectfully, you don't know what prompt engineering is if you think your methods are something different and novel. Literally everything you did here is prompt engineering.
Listen, at a high level all prompt engineering is, is being strategic and cognisant regarding how you communicate with the LLM, so that outputs delver what you actually want. You haven't discovered fire here, and nothing you demonstrated is remotely new.
Again, you're literally doing the very thing you claim not to be doing, which makes you look and sound foolish.
In short you think humanitarians will be mostly useful to the task😊 because thry know how to explain themselves.
Lol "How to be more intelligent" is a hilarious video topic.
That's why i don't follow the crowd you all don't think like a human. Your talking to an entity humans are slow in thinking he will not progress past prompting. Think about it you can have another ai write the prompts 😅😅😅 they waisted allot of time assuming
I dont understand how this is a skill...lol
This video is Ai generated.
Mmmmmm😅
Exactly some of the beginner advice I give in my consulting business when introducing AI.
You are one of the rare influences mentioning this at all. 🎉
Appreciate you watching. 🙌
Click bait 😂
Those who can, do. Those who can’t, make stupid clickbait UA-cam videos.
Stop giving clickbait and sensaniolistic titles... Do This Instead: admit that you are actually doing prompt engineering
@CsabaTothMr Clearly missing the point of the video, which aligns with my opinion and thus the title and thumbnail. I appreciate you taking time out of your day to watch and leave a comment 🙌