Your video is exactly what I needed, Goda Go! I'm an experienced Network Automation Engineer transitioning into AI, specifically as an AI Prompt Engineer. I've been struggling to find the right course and community to guide my journey-most just push their own agenda without offering real direction. Your insights on industry trends are invaluable. I’d love any advice or resources you can share. Thanks for the inspiration!
I'm the guy who writes the prompt that makes the prompt that gets the prompt for producing the test for the prompt produced by the prompt output from the prompt made by the other prompt I wrote. Not really, I'm just dreaming of the salary.
I am struggling with how to best position myself for these roles. Each time they come up, I'm continuing to feel behind the curve and struggle with how to present my prompring skills to fit their requirments. Thoughts?
I feel you.. well, truthfully, people get paid for the value, it should not be expected that making 500k will be easy. Doing this level prompt engineering and research is being 10-12h deep in the rabbit hole. I gave up at some point trying to keep up, because I have a business, and youtube. If I wanted to be full time prompt engineer it would be just that to keep up. To answer your question: perhaps showing innovative new ways you come up with prompting process and LLMs.
As an American, what I find funny about this whole thing is how our western society quickly marketed and discredited prompt engineering as a whole. But the companies are, in fact, very much seeking these new talents. 😂
they did because of people that were just spamming "this is the best prompt you can use" without showing research or other scientific method, they were marketer with a basic understanding of what an llm can be used for
haha :D On this episode we have software engineers getting laid offs, while prompt engineers who known how to type are becoming millionaires. AI world looks more crazy as each week pass by :D
There is more complexity once you go down the rabbit hole. But in essence, using LLMs to generate insane amounts of prompts is the way forward. Microsoft used gpt 3.5 to generate over 20mil prompts for generating synthetic data,. Of course you have humans managing that, but at scale it is LLMs that do the 80% work.
It's all about thoughts, AI is a thought processing machine, we use language (words, code, science) to trigger the process. Is the same reason is much better in english, it has many specific words to trigger certain concepts. Also hallucinations happen becusse user doesn't really ask everything properly. When I work on something firts i ask it to think how to do it from an extremely simple prompt. Once it outputs that, i make some correction go back edit my comment and replace with it's prompt, the results are amazing, AI is better at everything than any human, it amplifies human thinking
It's actually possible to check. GPT detector is still good at identifying GPT-4 content with good certitude even though the model was written to detect GPT-2. There are certainly other detectors for other language model architectures.
There are already Engineers that have written code for AI to be able to do Metaprompting on AI models as well, so this too isn't going to last long at all
I must try this out! I have completely immersed myself with Anthropic prompt engineering cookbook. You are absolutely the best women leader in AI field Goda Go and I’ve subscribed to support your work on RUclips!
Is AI a hype and its job will be dead in the future? Can you answer realistically. I am pursuing a bachelor's in Ai. And I don't know what to think of? I'll appreciate a detailed answer.
I said it in my videos, I think as long as we use LLMs where communication happens via text input (prompt) the job will be relevant. Most in demand, however, is prompt hacking (think what red teams do). So if you are already studying AI, knowing prompting is a must. I also said that it is the game for 1% of people, the rest 99% will use tools and interfaces that are optimised by the 1%. So if you up for a challange to be the 1% you can have a great career till LLMs evolve. But if we prompt with out thougts, lets say 10-20 years from now, in my view it would be still valuable to pursue.
"Safety" is such a kindergarten level hoopla. The AI isnt going to jump up and smack the taste out of you. We should be pushing the models to the limits.
LLM models without any safety measures are raw, representing the data input, which is internet, which is...not always the safest place. So the idea of safety is to install measurements and align the model outputs with our human values. But training practises are also part of it.
This is often the question and is not particularly transparent. Often it’s done through something called Reinforcement Learning from Human Feedback(RLHF) which is a fancy way of saying humans give responses a thumbs up or down. Anthropic is more transparent and provide quite a bit of information on what they call “constitutional AI”. You can google to read more but uses things like the UN human rights as a guide.
Hey Goda , really nice video ! I was wondering if I could help you with more Quality Editing in your videos and also make a highly engaging Thumbnail and also help you with the overall youtube strategy and growth ! Pls let me know what do you think ?
👉 To try everything Brilliant has to offer for free for a full 30 days, visit brilliant.org/godago + You'll also get 20% off an annual premium subscription.
Cant you just ask the model, "What is the best way to phrase this prompt?" to be a prompt engineering guru? Oh duh. I should have watched more of the video.
The title prompt engineering is a joke. There's not a better prompt engineer than asking AI for the prompt in said genre. To be a good prompt engineer is simply asking AI for the prompt. Please do explain all the hype about prompt engineering? 😮
Honestly, I like that they changed the title to research engineer focused on prompting, instead of prompting engineer. And the hype, from my observations, have been because such position feels accessible to anyone. For the companies, why they offer this much, it's because people who know how to work LLMs to the edge cases, red teaming etc, is super valuable.
If you are developing with this technology I think prompt engineer is an apt title. Since this technology is inherently probabilistic it actually takes quite a bit of “engineering” to get your prompts to output what you want and in a consistent way. If you’re just using it as an individual user you can get away with no engineering because you can just chat with it until you get what you want. But if you’re creating a program that has multiple prompts chains that all need to work in concert to consistently get a specific output…different story. Meta prompting like most generative AI will get you most of the way there but pushing it over the finish line requires quite a bit of knowledge, experience, and time.
![Seungmin "그렇게, 천천히, 우리(As we are)" | [Stray Kids : SKZ-PLAYER]](http://i.ytimg.com/vi/kAzmhLHePqU/mqdefault.jpg)
I'm submitting job application #10,384. Wish me luck.
"Good luck. We’re all counting on you." ~Airplane!
all the best!
probably crossed a million applications by now
😂👍 Good luck
Dont worry the large language model will pick you. ❤
Your video is exactly what I needed, Goda Go! I'm an experienced Network Automation Engineer transitioning into AI, specifically as an AI Prompt Engineer. I've been struggling to find the right course and community to guide my journey-most just push their own agenda without offering real direction. Your insights on industry trends are invaluable. I’d love any advice or resources you can share. Thanks for the inspiration!
Hey, they should hire this guy!
I'm the guy who writes the prompt that makes the prompt that gets the prompt for producing the test for the prompt produced by the prompt output from the prompt made by the other prompt I wrote. Not really, I'm just dreaming of the salary.
Changes resume to: Professional Metaprompter
Experience: Since Gpt-3 came out
I am struggling with how to best position myself for these roles. Each time they come up, I'm continuing to feel behind the curve and struggle with how to present my prompring skills to fit their requirments. Thoughts?
I feel you.. well, truthfully, people get paid for the value, it should not be expected that making 500k will be easy. Doing this level prompt engineering and research is being 10-12h deep in the rabbit hole. I gave up at some point trying to keep up, because I have a business, and youtube. If I wanted to be full time prompt engineer it would be just that to keep up.
To answer your question: perhaps showing innovative new ways you come up with prompting process and LLMs.
Devry: Adds Metaprompting bachelors and puts avg salary potential of 500k
As an American, what I find funny about this whole thing is how our western society quickly marketed and discredited prompt engineering as a whole. But the companies are, in fact, very much seeking these new talents. 😂
they did because of people that were just spamming "this is the best prompt you can use" without showing research or other scientific method, they were marketer with a basic understanding of what an llm can be used for
"Here I am, watching another episode of The AI Enthusiast-or should I say, 'the crème de la crème of AI news.'"
haha :D On this episode we have software engineers getting laid offs, while prompt engineers who known how to type are becoming millionaires. AI world looks more crazy as each week pass by :D
@@godagoI M strong 2. U sud try me some time
Is that all it is? I've been doing that from pretty much day 1
There is more complexity once you go down the rabbit hole. But in essence, using LLMs to generate insane amounts of prompts is the way forward. Microsoft used gpt 3.5 to generate over 20mil prompts for generating synthetic data,. Of course you have humans managing that, but at scale it is LLMs that do the 80% work.
Besides the advertisement, did this video say something?
I mean there are like 10 jobs or so for such a position, it‘s no really feasible for most or even half of people^^
Can a non-engineer become a prompt engineer?
yes but you probably wont be hired against the 100s of real engineers applying for these roles
It's all about thoughts, AI is a thought processing machine, we use language (words, code, science) to trigger the process. Is the same reason is much better in english, it has many specific words to trigger certain concepts. Also hallucinations happen becusse user doesn't really ask everything properly. When I work on something firts i ask it to think how to do it from an extremely simple prompt. Once it outputs that, i make some correction go back edit my comment and replace with it's prompt, the results are amazing, AI is better at everything than any human, it amplifies human thinking
Thats the annual rate. Gpt-5 getting released in 3 months. So you'll only get 1/4 of that salary.
It's actually possible to check. GPT detector is still good at identifying GPT-4 content with good certitude even though the model was written to detect GPT-2. There are certainly other detectors for other language model architectures.
This is truly insane!
Interview: "Ever since my mom prompted me into existence, I've been..."
There are already Engineers that have written code for AI to be able to do Metaprompting on AI models as well, so this too isn't going to last long at all
I must try this out! I have completely immersed myself with Anthropic prompt engineering cookbook. You are absolutely the best women leader in AI field Goda Go and I’ve subscribed to support your work on RUclips!
Is AI a hype and its job will be dead in the future? Can you answer realistically. I am pursuing a bachelor's in Ai. And I don't know what to think of?
I'll appreciate a detailed answer.
I said it in my videos, I think as long as we use LLMs where communication happens via text input (prompt) the job will be relevant. Most in demand, however, is prompt hacking (think what red teams do). So if you are already studying AI, knowing prompting is a must. I also said that it is the game for 1% of people, the rest 99% will use tools and interfaces that are optimised by the 1%. So if you up for a challange to be the 1% you can have a great career till LLMs evolve. But if we prompt with out thougts, lets say 10-20 years from now, in my view it would be still valuable to pursue.
AI is literally the future of humanity long term, it is a market that will never stop growing as AI continues to get smarter.
Yeah, its 3 people that get hired and thats it.
Hey man, what about when...
"Safety" is such a kindergarten level hoopla. The AI isnt going to jump up and smack the taste out of you.
We should be pushing the models to the limits.
Next Demand - Meta Prompting based on Generation (Bloomers to Gen Z, Alpha).
If all employers paid employees what they were worth…
Or
What inflation adjusted compensation will look like as US Debt moves towards Quadrillion.
Shit I didn't know that it was called meta prompting. I'm doing it all the time🤘
Gampang gampang sudah kau ku dekati kukejar kau berlari kudiam kau hampiri begitu teori simpel jadi rileks saja bila rejeki tak akan kemana thank you
What does safety first mean?
lobotomizing models, censoring
LLM models without any safety measures are raw, representing the data input, which is internet, which is...not always the safest place. So the idea of safety is to install measurements and align the model outputs with our human values. But training practises are also part of it.
@@godago who's values exactly?
@@godago who's values exactly?
This is often the question and is not particularly transparent. Often it’s done through something called Reinforcement Learning from Human Feedback(RLHF) which is a fancy way of saying humans give responses a thumbs up or down.
Anthropic is more transparent and provide quite a bit of information on what they call “constitutional AI”. You can google to read more but uses things like the UN human rights as a guide.
it said you need engineering experience in video u just shared can you open your eyes?
What highlighting tool do you use?
Readwise! which is Readwise Reader :)
Too bad I don't have my post-doc in metaprompting, just my PhD. I don't like my chances.
Can i apply from india
Hey Goda , really nice video ! I was wondering if I could help you with more Quality Editing in your videos and also make a highly engaging Thumbnail and also help you with the overall youtube strategy and growth ! Pls let me know what do you think ?
👉 To try everything Brilliant has to offer for free for a full 30 days, visit brilliant.org/godago + You'll also get 20% off an annual premium subscription.
The age of Promptception.
Cant you just ask the model, "What is the best way to phrase this prompt?" to be a prompt engineering guru? Oh duh. I should have watched more of the video.
Do you always believe what is written?
No
@@godago So, while companies continue to pretend that they pay well, we continue to pretend that we are doing the work better.
The title prompt engineering is a joke. There's not a better prompt engineer than asking AI for the prompt in said genre. To be a good prompt engineer is simply asking AI for the prompt. Please do explain all the hype about prompt engineering? 😮
Honestly, I like that they changed the title to research engineer focused on prompting, instead of prompting engineer. And the hype, from my observations, have been because such position feels accessible to anyone. For the companies, why they offer this much, it's because people who know how to work LLMs to the edge cases, red teaming etc, is super valuable.
If you are developing with this technology I think prompt engineer is an apt title. Since this technology is inherently probabilistic it actually takes quite a bit of “engineering” to get your prompts to output what you want and in a consistent way. If you’re just using it as an individual user you can get away with no engineering because you can just chat with it until you get what you want. But if you’re creating a program that has multiple prompts chains that all need to work in concert to consistently get a specific output…different story.
Meta prompting like most generative AI will get you most of the way there but pushing it over the finish line requires quite a bit of knowledge, experience, and time.
Nonsense, all hypes
Love your brilliant Videos 👀
Thank you so much!!