Finally someone going beyond the basics. Make people smarter and more efficient. Chat-GPT is so much more than a grammar checker!!
@KCM25NJL6 күн бұрын
Good video bro.... if I were to kindly request that you stick a pop filter in front of your HyperX? I'm not sure if it applies to many, but for those of us with higher end headphones, there's a really strong sub-bass response from your recording, which try as I might, I've struggled to EQ out. If I were to describe the sound .... it's like a loud bassy heartbeat in my head throughout your video..... and it's nothing more than the wispy breaths on your condenser as you narrate. Anyway.... keep em coming :)
@AppyDave5 күн бұрын
Thanks for the feedback on the hyperX, and glad you enjoyed the video
@TheDeliveryDoctor15 күн бұрын
Leaps and bounds, David! Thanks for including us in your journey. What you are developing here is light years beyond what many of us thought possible. I can’t wait, not only for AWB to go live, but for your course that will guide us to greater heights using your tool(s)! Congrats 🎉
@alec.johnson17 күн бұрын
A really well structured demo of your process and awesome software! Fantastic work David 😊
@vitalis20 күн бұрын
What you are doing is really novel and I do think that eventually we might have a coding language for LLMs however you should think about changing the design of your website. It’s the typical one that clickbaiters and course sellers use, and basically associates your work with spammers. Just my two cents as someone who works in the digital space.
@TheDeliveryDoctor29 күн бұрын
Brilliant, David! I’m at a loss for words to say more, and you knowing me, you know that says a lot! 😆
@alec.johnsonАй бұрын
Fantastic demo… really shows how one can use ai to enhance and accelerate workflows!
@AppyDave29 күн бұрын
Thanks, glad you liked it
@cinoss5Ай бұрын
Thank you
@vidIQАй бұрын
That's very cool! 👏👏👏
@AppyDaveАй бұрын
The original video from your channel was very cool. I just added my own personal perspective.
@swannschilling474Ай бұрын
I think automating KZbin videos is a pretty good shortcut to be shadowbanned! 😅
@AppyDaveАй бұрын
I'm actually Automating the processes. Anyone that has used a tool like VidIQ or a video editing software is Automating processes as well, this is not any different from that point of view. I checked out your channel and your videos look great, thought it seems like a couple of years since you last uploaded, I'm curious why?
@swannschilling474Ай бұрын
@@AppyDave I just did not find the time to do more! I am always putting ways too much time into this stuff! I am having a lot that I would like to share in KZbin but its too time consuming, so I mainly post on Instagram! Shorts did not seem to work for me! I also build a couple of bots to boost my reach on Instagram but refrained on it later because social media will punish a channel so the plattform will not be overrun by this type of content! Unless you are on TikTok probably! 😁
@proflux547Ай бұрын
I'm looking forward to the next video
@AppyDaveАй бұрын
Great, I'll be doing a few around my various KZbin Automation workflows
@ahmadbakour369Ай бұрын
Is it ok and safe ti upload this games on Google play. Or it's just testing or temporary?
@DanielAfonso-IT_ConsultantАй бұрын
It's pretty good. I ported the published Collab notebook to use AWS Bedrock and have instantly improved prompts for some workflow automation things I'm working on. Thanks for adding your experience and workflow to the mix.
@johnnoren7244Ай бұрын
Awesome video. And using a DSL is really clever, I will steal that.
@AppyDaveАй бұрын
:) glad you liked, the next video I put out will go more into the DSL
@OKraft28Ай бұрын
Awesome! This makes work soo much easier. 😁
@AppyDaveАй бұрын
Glad it helped!
@KennethChiarelliАй бұрын
Interesting
@AppyDaveАй бұрын
I have more videos coming out on this topic
@onlygoatsfarm2 ай бұрын
And you’re on your way to monetization in Adsense! Thanks for this! #vra
@AppyDave2 ай бұрын
Getting there 😁
@onlygoatsfarm2 ай бұрын
@@AppyDave same! Just gotta get those watch hours! Lol
@sizzleserveyum2 ай бұрын
Awesome tutorial!
@AppyDave2 ай бұрын
Glad you liked it!
@emilyseilhamerart2 ай бұрын
Thank you for this! I've been seeing "Buy me a Coffee" a lot recently and was curious what it was. I was looking into Patreon, but this looks like another great option!
@AppyDave2 ай бұрын
Glad it was helpful! It's an easy set up.
@christiansforrevival3162 ай бұрын
very good
@AppyDave2 ай бұрын
Thanks
@TeachYHWH2 ай бұрын
This is REALLY GOOD
@AppyDave2 ай бұрын
:) cool, glad you liked
@TeachYHWH2 ай бұрын
I am going to start this on my channel. I was JUST asking myself the other day how does this work
@AppyDave2 ай бұрын
Yeah, it is an easy setup
@arc-system-for-dads2 ай бұрын
Awesome video, Dave! Keep up the great work 💪🏻
@AppyDave2 ай бұрын
Thanks Luke :)
@arc-system-for-dads2 ай бұрын
@@AppyDave 👊🏻❤️
@bloodgourd2 ай бұрын
I got banned last week after supporting the company since version 3. It hurts. I've generated something like 20K images and twice now have paid for a year's subscription (meaning they're going to keep the rest of my money, I guess). My first warning was when I generated an image of Donald Trump with an alien toupee. Stupid as hell, but no biggie. it got me banned for like a day. This time, it was permanent, and I have no clue what it was. I have pushed my luck with gore in the past, but come on. A lifetime ban? That's insane. I'd love to know a decent alternative, but I was really enjoying MJ and it sounds like they're getting ready to implement some amazing new things. Assholes...
@hakicapivaraАй бұрын
I tried generating funny images with Donald Trump randomly because I wanted to create some humorous work that I thought Americans might find amusing from my point of view. However, I don't take sides in politics at all. I just find Trump to be a funny guy in terms of memes and thought of testing it out. All the prompts were blocked, and weeks later, I was poorly received by MJ support on Discord. The next morning, I generated some images involving the American flag. No mockery, just beautiful images involving America. In summary, I didn't do anything against the rules, but I believe the reason for my ban was that I was perceived as a supporter of one political side or the other. The support team gave me an empty and nonsensical response, and I also lost money because there will be no refund, and I won't be able to use the remaining credits from my plan. If you have any alternative suggestions, I'm open to them. I don't intend to use MJ anymore, and I'll do everything I can to oppose this company! And I’m not American.
@bloodgourdАй бұрын
@@hakicapivara American politics are a joke. It's all Coke vs Pepsi nonsense that serves the interests of rich assholes. It's frightening that a corporation decides what's acceptable and what's not to generate. Sure it's AI art now, but how long until you won't be allowed to criticize people in power with any medium? I've been messing around with Flux Pro through Glif and Civitai. I've already made some cool stuff. Have fun!
@AppyDaveАй бұрын
I too am doing everything to appose the company, I feel they are have no integrity and today decided to start using Leonardo.AI as my main tool, I will do some videos on it when I have some spare time.
@NeverhappenedAI2 ай бұрын
Yeah, i got banned recently also, two days after i paid for a new month. Sort of similar problems, i used to live in Turkey, then i moved to Spain and now live here. Their support never explain the ban nor ready for any sort of communication, i think that's a BS really. Well, decided to fininsh with them and use other tools, glad i didnt pay for a yearly plan... I sympathize with you and can feel the upset at this company's behavior.
@AppyDaveАй бұрын
Thank you for sharing your experience! It’s always helpful to hear from others who have faced similar challenges. Wishing you the best of luck in finding a more reliable service!
@Murshid6ixty2 ай бұрын
I have 2nd ban in my midjurny what to do please someone help me
@AppyDaveАй бұрын
I have decided to move over to Leonardo.AI
@kevinzhang34002 ай бұрын
I sympathize with AppyDave. I'm in Australia and haven't been to other countries, so I'm not sure if this is due to cross-region login issues. Since I started using Midjourney, I've already had two accounts taken down (today marks the second one). On August 7th, on a day when I didn't log in, the MidjourneyBot informed me, "You have been blocked from Midjourney due to accessing our services via automation, scripting, or third party tools. Ban ID: AA080724." I saw this message today (August 11th). This is the second account of mine that has been banned. I initially suspect I may have been hacked). My second account hasn't been logged into for this month. Today, for the first time in August, I logged in and immediately saw the account ban notification. I have no idea what happened. I tried contacting them again, but I couldn’t get any help. So my account had brutal, uncommunicative, blocked. Since my first account had the same issue, so I was extremely careful with the second one, to the point where I wouldn't log in unless absolutely necessary. Yet, it still didn't last a year... Each time they ban an account, they also block the associated IP, credit card payments, and community services. I believe this could be one of the reasons for MidJourney's potential downfall in the future. I can no longer tolerate using it, no matter how useful it is. It’s truly a disappointing experience.😮💨
@jonhobbs51132 ай бұрын
I'm cutting off Midjourney too now. Companies should not get away with doing people wrong. It's sickening!
@AppyDave2 ай бұрын
I agree, I have had a few issues with paid services causing me grief , they should be accountable and thats why I created this video, I can't make much of a difference, but every little bit helps
@jonhobbs51132 ай бұрын
@@AppyDave Thank you for continuing to make content that informs us all on the latest and greatest AI apps out there. We all need to know which service(s) is the best and most fair and your perspective really helps!
@x-processes2 ай бұрын
@@jonhobbs5113 Thanks :) its hard work getting these videos out, but it is a passion project
@zenkaiproductions38042 ай бұрын
Do you know how to use the api?
@AppyDave2 ай бұрын
I have not tried it out, but I will put it on a list to look into to
@bernardbell19042 ай бұрын
How to get Source code of the website created or how to export it?
@AppyDave2 ай бұрын
There is a download button in the top right hand corner
@winkletter3 ай бұрын
Great tutorial. I like to add "dev tool" buttons that display the code, a changelog, and a playtesting menu. I usually make the playtesting menu so I can change game variables, but I hadn't thought to have it show a live debug screen.
@AppyDave3 ай бұрын
Glad you enjoyed, the play testing variables is a good idea
@lidjitsu3 ай бұрын
Wow, Thank you. You are creating great content.
@AppyDave3 ай бұрын
I appreciate that!, thanks for the feedback, is there a type of video that you are interested in seeing more off?
@sajeebmekdad43283 ай бұрын
I also got banned. But I do not use any automation, script or third-party tules as they said. I did not contact any support from them. 😢😥
@AppyDave3 ай бұрын
Its a shitty feeling and costly
@sajeebmekdad43282 ай бұрын
@@AppyDave they say that they will refund me
@AppyDave2 ай бұрын
@@sajeebmekdad4328 great to hear Sajeeb, I'm about to make a 2nd request to them as well, will let you know how I go
@zazerb7503 ай бұрын
Hey mate, thanks a ton for the video. Was exactly what I was looking for, a succinct run through of realistic use cases.
@AppyDave3 ай бұрын
Thanks for the comment, and glad it hellped
@alec.johnson3 ай бұрын
This was such a great video and really interesting demonstration David! Top notch!
@AppyDave3 ай бұрын
Thanks Alec, appreciate the complement.
@ADXART8883 ай бұрын
We were talking about the potential competition a year ago, here it is.
@AppyDave3 ай бұрын
True
@ethanlewis14534 ай бұрын
If you purchased three weeks earlier from today, then you can still dispute the charge on your credit card. The cut-off is usually about 60 days. A technical solution for those who don't yet have a banned account could be to use a whitelist residential IP using a specialty proxy service like IP Royal. It does take some level of computer skills to set up. Since the company is headquartered in San Francisco, your legal options are limited because it would be extremely expensive. If someone were in the USA they could try a DIY lawsuit in federal court using ChatGPT for help.
@DavidHuntPHG4 ай бұрын
Thanks Dave! That helped me a lot - even with the tool I have been using. Do they all have to match the reference graphic dimensions with the output graphic dimensions? The way it looks many of these all seem to have the same core engine? You seem to have gotten younger since I meet you in 2012!
@AppyDave4 ай бұрын
Hi David, Graphics dimensions can be controlled in some tools and others can use capabilities like aspect ration, eg. 16:9 Core engine is usually Stable Diffusion or some based on DALE3. I don't feel younger, but love the compliment. :)
@DavidHuntPHG3 ай бұрын
@@AppyDave thank you
@GaelleBretinTokpo4 ай бұрын
Thanks 🙏, it was very interesting👌
@AppyDave4 ай бұрын
Glad you liked it!, let me know if there are any topics in Prompt Engineering you want to see covered
@pungutan_liar4 ай бұрын
Thank you for the detail
@AppyDave4 ай бұрын
Glad it was helpful, let me know if there are any prompt engineering topics you are interested in
@JeffreyMacapinlac4 ай бұрын
Thank you for this very timely information!
@AppyDave4 ай бұрын
You are so welcome!, glad it was useful, let me know if you want specific prompt engineering videos.
@importyntai4 ай бұрын
Truly informative video David. Just followed you in LinkedIn. Cant wait to try your GPTS
@SimonBakerCo4 ай бұрын
I see Diffusion Bee being great for MEME's😆
@AppyDave4 ай бұрын
Yes, there is the use case that makes sense
@lauranolan73674 ай бұрын
Looking good Appy D 👌🏽
@AppyDave4 ай бұрын
Thank you :)
@susanhansonart4 ай бұрын
thanks for the tips
@AppyDave4 ай бұрын
Glad it was helpful Susan
@ShelleyMoses-ru7rh4 ай бұрын
U should always leave a link to repo and and install commands
@AppyDave4 ай бұрын
Good advice, I dropped a link 🔗 too the repo, still getting the hang of KZbin videos and the hundred little details with each upload
@tonyzola85504 ай бұрын
Fantastic. Perfect. You provided a good service. Thank you.
@AppyDave4 ай бұрын
Thanks , it's not my normal content style but I thought it was needed
@xspydazx5 ай бұрын
It's actually using Jupiter notebook ... In the back ground to execute app it is running a cell in notebook , or executing it's code in a cell in Jupiter as it uses the Jupiter client library .... So to create your own function calling script , you can use instructor (to create the correct json function calling to Jupiter client ) .. IE the template to follow for the call .. as well as your script needs to contain the instructions to follow , IE to return the code in segments etc , to resolve problems returned . On the next turn , as well as examining the problem and solving first before resubmitting the task ,.. as well as asking it to create any function it needs to perform a task , given these progrsmming langs , and gathering info about the os , so that bash can be used where possible , on knowing the system structure , as it should default to scripting first . Until it has learned enough to use the bash . As well as confirming all tasks until the task is fully error free. Enabling for tasks to be auto run .... The prompt used is quite in-depth .. Perhaps even to create the "summarizer response* always ... IE collect all past messages and summarize , then use this context to respond , hence the context history should be summarised every as well as calculated as when it goes out of the sliding window context it will need to be concatenated and summarised . As well as stacked if necessary in the past history ... To enable rezumarizing even further back in the history .. giving the model a high briefing of the past. IE optimizing the history shuffling the important context to the front messages and relating the irrelevant ... Once you have the prompt set all you need to do is intercept the messaging completions and query's and return the responses . Etc . Verbose no verbose (it stores anyways in json message formates so it's easy to update the model during training with these messages ... So you can discuss information with the model first with truth fil and factual data and it will be stored in it's conversations ....then when you upload these and the relevanr document data. In training .. you will have truly taught the model new knowledge as well as had good discussion which it can refer too also ! .,
@AppyDave2 ай бұрын
cool
@xspydazx2 ай бұрын
@@AppyDave i since updated this method : As you only need the two functions : Method : 1 one to executecode on the system... and one to save/load a file from your tools folder : So you can allow the model to generate any function for itself :( best way ) and it can add the tool to its tool collection : then it can execute the code : this way is a beautiful way , as i used the stop and think, then act technique .. if it needs to act it will execute one of its tools or create a tools : this method really does work well : as it can generate bash code and python scripts etc: you can just create tools and add them to the tools folder ! < it will load these tools into its tool box ! > Method 2: This is simular to the above but it can only execute a tool , you supply the tool names in your code and it searches the docstrings for each function and adds it to its toolbox : ( i found that the model does not like the pydantic SLOW!! ) << the docstring method makes your overal code lovely and strong : the model knows by your docstring how to operate its functions and tools : Personally i have the model making its tool call behind the scenes so it can make the call and return only the final output to the use as the map it creates ( the list of tool names and , funciton names ) is enough to parse to see if the funciton is i the box then is can be called by the model its output returned and it can continue with its response generation , so if it needs a multi function it should pla and execute in the correct order storing the values in its Response history ( messages and extract them for its final response ( ie self conversation )... This method waorks also but the above method does incorperate this: Method 3: Yes the Notebook Method: so this method does not store tools it just generates functions and scripts and you can supply a list of pydantic functions: again the main internal fuciton is to execute a code : here we use e2b sandboxes to create a session with a notebook and an environment allowing you to execute code in this surface hence work on a project : Sadly i did not alow it to save these books ( maybe it does in appdatatemp files ... but you can reload a session if you know the session id: and the final output is saved to your working folder : the main thing with funciton calling despute the methodology you use : its important to let the model Tell you in the response it is executing a funciton : so you can intercept the respose and return the answer: the answer shoud containn the result and aknowedgement that this is the response for this funciton call :: then iterate through your function list ( for error checking and execute the funciton from the list with the parameters given : For me method one is super good working well as an agent : Once you can execute code on your system then you have a open interpretor : So why give the model functions ?? .... API CALLS !< these contain your api Keys !!!> Only! If you need to search a website etc you do not need an api you need a llm !!<< it knows how to do this without paying money ! I recently spent time training my model as a function caller with many different datasts and aso i trained for being a planner... so it can be a planner only not solving tasks but providing recipes for tasks: Then i retrained it to the messaging format template ! this enabled for the model to function correctly : the most important is the model is using a messaging format template as oter output types will not work : so if a mpodel has been trained for alpaca or normal mistral you need to change the chat template for the model to be the Chatml ... OpenAI format ! then it will perform exactly as you require even if you convert if to gguf: i was having problems with out put when i realized that the chat templates are deterimental to a model ! : so for normal use ( non messaging ) then you should remove that template and put a Q/A template then make the gguf and it will perform in GPT4all etc correxctly ! << the messaging model will not perform in these apps ... they will only perform on setrver setups... so if your havinbg problems making OpenInterpretor and other OpenAI (open source stuff ) then it is this !! <<
@AppyDave2 ай бұрын
@@xspydazx Wow, you've really delved deep into this! It sounds like you've found some effective methods for using Open Interpreter. Thanks for sharing your insights!
@xspydazx2 ай бұрын
@@AppyDave yes it was quite hard to figure the whole thingy out as the main this was just one function !! << the execute code and return it to the model >> the rest was logistical in fact ( like why did it not work with MY model ) << on MY tiny system ... So after this ChatTemplate was solved then the model begun to work with all my ollama/lmstudiio projects : So the chatML template for output on fie tuing was the also most crucial aspect < hence only these models work on funciton calling the chat models must use the pydantic method ... (which was too slow) i als realised that my Opus model (Manager) should have stable diffsuion and other specialist model as tools : And infact we also should make the RAG be a tool and ot be an input stage , so if the model needs to use rag to get a richer response it can do a rag ! , if the model feels the rag is generating most of the responses it can generare a dataset from the rag for fine tuning ... if the model has dataset for fine tuning it can do a quickiy train of the small rag dataset!.... So our model is a manager Agent and we should train it to be a manager... ie lots of planning an grading: LOL: