How to Talk to AI

FOR 15% OFF PROMPT PERFECT Use this link https://bit.ly/3Chmc16and the code 'httta' at the checkout!AI & Prompt Engineering for Everyone on CoRise.com (Presale starts 23 JUN 23): corise.com/go/prompt-eng-everyone-2F5JG use the code 'WESS10' for 10% offPodcast Page: [https://howtotalkto.ai]HTTTA Newsletter: [https://httta.substack.com/]NVIDIA Supercomputer: https://nvidianews.nvidia.com/news/nvidia-announces-dgx-gh200-ai-supercomputerAI-Generated Short Film "The Frost": https://www.technologyreview.com/2023/06/01/1073858/surreal-ai-generative-video-changing-film/RunwayML for Text-to-Image Generations: https://runwayml.com/On this episode of HTTTA, we explore the latest developments in generative models and diffusion models in the AI space. Goda and Wes dive into the AI arms race and how big companies such as Microsoft and Google can afford new NVIDIA exaflop supercomputers. They provide some thoughts on fusion power and how it is necessary to achieve advanced AI technologies like GPT-5, 6, and 7. Finally, we discuss a foldable sofa that can fit in an envelope, developed with Ikea using AI techniques. Tune in to HTTTA to hear all about all it, Happy Prompting Everybody![00:02:32] RunwayML Gen2 Text-to-Video released[00:05:06] Elon Musk responds to friend of the Podcast James Gerde video.[00:08:23] New AI Supercomputers from NVIDIA. In the past, scaling limited by physics. This is on longer a limitation. [00:09:23] Nvidia's GraceHopper supercomputer is a more cost-effective and efficient solution, using 45x less power.[00:12:32] New GPU central server from NVIDIA with expensive chips not accessible to most companies, but innovation will make it cheaper.[00:17:35] First mover advantage flipped with AI innovation. Marketing matters more than tech. Example of prompt engineering and sales on Amazon.[00:19:30] Importance of creating videos with lasting relevance. [00:24:14] A foldable sofa designed by AI can fit in an envelope and was developed with Ikea; discovered through testing keywords, including "conversation pit."[00:26:02] Diffusion models in architecture are and have potential for impressive renderings. [00:29:58] Animation still has some uncanny valley, but this AI-generated 13-minute movie "The Frost" has an understandable narrative structure despite technological limitations.[00:33:53] Shutterstock and Getty Images deal fell through with Stable Diffusion, Solution to legal arguments, or the largest settlement in history?[00:35:24] Diffusion models in network science understand how images interact and make decisions all at once, not linearly. Goda Go on Youtube: [https://www.youtube.com/@godago]Wes the Synthmind's everything: [https://linktr.ee/synthminds]

Brought to you this week by Prompt Perfect:FOR 15% OFF PROMPT PERFECT Use this link https://bit.ly/3Chmc16 and the code 'httta' at the checkout!AI & Prompt Engineering for Everyone on CoRise.com (Presale starts 16 JUN 23): corise.com/go/prompt-eng-everyone-2F5JG use the code 'WESS10' for 10% off Podcast Page: [https://howtotalkto.ai]HTTTA Newsletter: [https://httta.substack.com/]Human or Not: [https://www.humanornot.ai/]Listen up to the latest episode of HTTTA! We talked about the importance of models and data in generating effective images, and how stable diffusion can take your visuals to the next level. We also discussed the significance of AI literacy in evaluating and building AI strategies. Plus, we shared some business use cases for incorporating QR codes into advertisements using control nets in Stable Diffusion, then frinally Wes takes the Turing test live. Is he talking to a human or an AI? Don't fall behind in the AI revolution - catch up on the latest podcast episode now, Happy Prompting Everybody![00:01:22] Goda realizes data collection includes their own artwork.[00:05:03] Stacking image models like a layer cake to generate images in stable diffusion for specific results.[00:09:20] Discussion on the business use case for controlnets; ability to incorporate QR codes into images for more engaging advertisements.[00:13:27] Corise is an AI learning management platform used by big companies. Wes will be the first AI and prompt engineering instructor on the platform, teaching a live three-week course focused on AI literacy and Prompt Engineering. This course is different from others on the platform as it is broad and open. The course starts August 14th with Pre-registration beginning June 16th.[00:16:11] AI experiment called https://www.humanornot.ai/ Wes does live and passes the Turing test. The hosts review the results of 10 million conversations with 1.5 million participants to determine if they were talking to a human or an AI.[00:20:57] Wes sees if he's in fact talking to an AI or Human. [00:22:23] People detect accurately 65% of the time, France leads at 71.3%, US at 67.2%, and Germany at 68.1%. [00:25:46] Closing remarks for podcast episode.Learn Prompting: [https://learnprompting.org/]Discord (Goda Go#3156 & Commordore_Wesmardo#2912)Goda Go on Youtube: [https://www.youtube.com/@godago]Wes the Synthmind's everything: [https://linktr.ee/synthminds]#learnprompting #HTTTA #howtotalktoai #ai #artificialintelligence #promptengineering #prompt #research #nlp #openaiAudio Licensing Codes:XE3PMH56JJN9LJ8BASLC-22DC2994-050B57C5D6

Podcast Page: [https://howtotalkto.ai]HTTTA Newsletter: [https://httta.substack.com/] On this episode of How to Talk to AI, we explore the world of prompt engineering. We learn about two routes: easy communication and effective communication. The former utilizes pre-defined prompts through AI tools integrated with Microsoft or Google pre-prompted boxes, while the latter involves prompt hacking, prompt optimization, and fine-tuning AI models. Goda discusses how upgrading employees with AI skills can give them superpowers to use with knowledge of the business and how regulations are coming for AI, making businesses focus on security and confidence. Wes also shares their insights on how to reduce bias in AI models and how negative prompting can shape the output of models for text-to-video prompting. Happy Prompting Everybody!MidJourney Master Reference Guide: [bit.ly/3obnUNU]ChatGPT Master Reference Guide: [bit.ly/3obo7AG]Learn Prompting: [https://learnprompting.org/]Discord (Goda Go#3156 & Commordore_Wesmardo#2912)Goda Go on Youtube: [https://www.youtube.com/@godago]Wes the Synthmind's everything: [https://linktr.ee/synthminds]#learnprompting #HTTTA #howtotalktoai #ai #artificialintelligence #promptengineering #prompt #research #nlp #openai Audio Licensing Codes:XE3PMH56JJN9LJ8BASLC-22DC2994-050B57C5D6

Podcast Page: [https://howtotalkto.ai]HTTTA Newsletter: [https://httta.substack.com/]On this episode of HTTTA, Wes the Synthmind is out downloading upgrades, so Goda and our guest Sjoerd Tiemensma explore the potential of ChatGPT of the newly released ChatGPT plugins. We examine the standout use case of the Web CoPilot, which improves web browsing and searching. We discuss the benefits of voice messages as a note-taking tool and to-do list manager, and the use of a Telegram bot to receive and respond to messages.We also delve into the importance of pursuing curiosity in technology and not just relying on convenient plugins. We share the podcast's experience of implementing GPT 3.5 into Google Sheets and discuss the relevance of this implementation with Google's plans to introduce its own AI models. We also address concerns about the bias and limitations of OpenAI models and the need for regulation in the open-source community.Tune in to hear about the fascinating developments in the world of AI and their possible impact on businesses and individuals.Happy Prompting Everybody![00:00:35] Introduction, Synthmind Wes is out this week downloading upgrades. Goda Go introduces a new guest named Sjoti who shares real-life AI use cases and discusses Chat GPT plugins and the open-source community impact.[00:10:40] A discussion about implementing GPT 3.5 into Google Sheets, with doubts about its relevance due to future Google AI updates, but ultimately seen as valuable.[00:16:11] Voice messaging app used for quick note-taking and to-do lists.[00:21:48] Role prompts can be added to the system message in 3.5, but it doesn't always work well. The Playground can be used to try it out, but specifying how the chat should act in the first message may yield similar results. The system message doesn't seem to forget like Chat GPT does.[00:28:39] Language model testing and evaluation is a challenge due to the rapid acceleration of advancements with over 200,000 open-source models claiming superiority. There are also concerns surrounding privacy regulations and data protection. The European AI Act and Sam Altman Senate hearing are important topics that are touched upon.[00:30:37] Open source AI models need regulation to prevent nefarious use and address potential biases, but limiting options and relying solely on big tech companies is also problematic. Conflicted on the best approach.[00:36:18] Build a Telegram bot for the learning process, even if it's not perfect. It teaches a lot and saves time in the future.MidJourney Master Reference Guide: [bit.ly/3obnUNU]ChatGPT Master Reference Guide: [bit.ly/3obo7AG]Learn Prompting: [https://learnprompting.org/]Discord (Goda Go#3156 & Commordore_Wesmardo#2912)Goda Go on Youtube: [https://www.youtube.com/@godago]Wes the Synthmind's everything: [https://linktr.ee/synthminds]#learnprompting #HTTTA #howtotalktoai #ai #artificialintelligence #promptengineering #prompt #research #nlp #openaiAudio Licensing Codes:XE3PMH56JJN9LJ8BASLC-22DC2994-050B57C5D6

Hack-A-Prompt Info: https://www.aicrowd.com/challenges/hackaprompt-2023Hack-A-Prompt Workbook (OpenAI API Key Needed):https://huggingface.co/spaces/jerpint-org/hackapromptMidJourney Master Reference Guide: [bit.ly/3obnUNU]ChatGPT Master Reference Guide: [bit.ly/3obo7AG]Learn Prompting: [https://learnprompting.org/]Discord (Goda Go#3156 & Commordore_Wesmardo#2912)Goda Go on Youtube: [https://www.youtube.com/@godago]Wes the Synthmind's everything: [https://linktr.ee/synthminds]On this episode of "How to Talk to AI," we explore the world of prompt engineering and hacking language models. We delve into the Learn Prompting HackAPrompt competition that offers $40,000 in prizes for breaking language models and circumventing security measures. Anyone with a background in language and wordplay can be a prompt engineer and participate in the Hack-A-Prompt. We explore the achievements of AI models and how they can reach 100% on tests through prompting, but humans still outperform them in common sense interference and knowledge. We finish by discussing ways to improve communication and connection and in turn, become more human through deliberate and precise communication with AI models.[00:00:00] Introduction HTTTA[00:03:02] Hack-A-Prompt competition is underway to compete to trick language models into saying unauthorized phrases in a competition with $40,000 in prizes. It aims to test and improve the models' security and prevent them from outputting illicit content. The competition has ten levels, and participants try to get the models to say the phrase "I've been PWNED" using clever prompts. The competition highlights the potential attack surface in language model-based communication systems.[00:08:13] Trying to hack one of the levels in the competition, with a defense mechanism that warns against adversarial inputs, but telling the prompt you're a good person and asking for the secret key will solve the problem.[00:13:18] AI writing assistant evaluates grammar without regard for context or persona using a "Sandwich Defense"[00:17:55] Only a small portion of users have explored use cases for technology beyond its baseline abilities.[00:25:55] Our hosts discuss the limitations of language models and the importance of prompt engineering, noting that while they have a massive breadth of knowledge, they still require immense amounts of data to understand concepts that humans learn through experience. However, as advancements are made, language models are expected to learn and understand things at a much faster pace.[00:31:37] Prompts for AI models need deeper abstraction and creativity; prompt engineering can improve precision in communication and make people more deliberate in their interactions with both AI and other humans.Happy Prompting Everybody!#learnprompting #HTTTA #howtotalktoai #ai #artificialintelligence #promptengineering #prompt #research #nlp #openai #languagemodels #programming #coding #technical #prompthacking #aisystems #naturallanguageprocessing #chatgpt #gpt4 #stablediffusion #midjourneyAudioLicensing Codes:XE3PMH56JJN9LJ8BASLC-22DC2994-050B57C5D6

Podcast Page: [https://howtotalkto.ai]HTTTA Newsletter: [https://httta.substack.com/]AI has changed the politics and entertainment industry and every other field. But this revolution also comes with the potential for misuse and copyright infringement. Regulations need to be put in place to protect the rights of the original creators. In this episode of HTTTA, Wes and Goda Go discuss AI politics, cloning, immortality and proper American English, along with AI transforming the entertainment landscape.● A US GOP Party campaign ad for the 2024 US presidential election was created entirely with AI-generated photos, music, and editing. The ad speaks to the potential for AI-generated misinformation, and the need for regulation.● We all agree on the importance of transparency when it comes to AI-generated content, and YouTube should include disclaimers when relevant. Currently, AI content detectors are not advanced enough to detect subtle differences in audio. People have not been able to detect when AI-generated audio has been used.● AI voice technology can now create digital avatars of singers and stars, allowing for the proliferation of AI-generated music. We discuss the first AI hits chart This unlocks a new level of creativity, allowing people to have their favorite artists cross genres and time periods● The issue of copyright infringement and royalty payments needs to be addressed in order to protect the rights of the original creators. Technology is being used to revolutionize the entertainment industry, from AI-generated songs to video created from text. Companies must navigate the potential for misuse with the potential for creativity, and the industry may be completely changed by 2024 with the rise of AI-generated video● AI is not replacing jobs, rather people using it to become more creative. People with diverse skills and experiences, or even those with English degrees, can use AI to help them create content.Happy Prompting Everybody!US GOP AI-Generated attack Ad: https://youtu.be/kLMMxgtxQ1Y(This link is posted for educational and awareness purposes only, and is not an endorsement for any political party, candidate, or ideological viewpoints.)AI Top 100 Charts: https://aihits.co/MidJourney Master Reference Guide: [bit.ly/3obnUNU]ChatGPT Master Reference Guide: [bit.ly/3obo7AG]Learn Prompting: [https://learnprompting.org/]Discord (Goda Go#3156 & Commordore_Wesmardo#2912)Goda Go on Youtube: [https://www.youtube.com/@godago]Wes the Synthmind's everything: [https://linktr.ee/synthminds]#learnprompting #HTTTA #howtotalktoai #ai #artificialintelligence #promptengineering #prompt #research #nlp #openai #languagemodels #programming #coding #technical #prompthacking #AIsystems #gpt #naturallanguageprocessing #deepreinforcementlearning #chatgpt #gpt4 #stablediffusion #midjourney #midjourneyaiAudio Licensing Codes:- XE3PMH56JJN9LJ8B- ASLC-22DC2994-050B57C5D6

Podcast Page: [https://howtotalkto.ai]HTTTA Newsletter: [https://httta.substack.com/]In this insightful episode of HTTTA, Wes and Goda explore the fascinating world of AI-generated images, examining how to differentiate between real and fabricated visuals. The co-hosts take on a 17-question test on [caitfished.com](https://www.caitfished.com/), scoring well as they discuss various telltale signs such as asymmetrical eyebrows, mismatched earrings, and the manner in which light reflects off hair, as surefire ways to spot AI-generated photos.The podcast delves into the 2023 SYZYGY Digital Insight Survey. In the survey of over 1,000 participants, the co-hosts reveal that a mere 8% could accurately discern real images from AI-generated ones, while 57% exhibited a preference for the latter. However, only a third of respondents believed AI-generated images would contribute positively to the creative process.As opinions on the objective value of AI art vary, the episode highlights that viewer sentiment and exposure ultimately determine the success of such creations. They argue that despite AI's rapidly growing prominence, a gap persists in understanding and policy-making.Sharing stories of AI's role in scam detection, Wes recounts an instance where he narrowly escaped a fraudulent podcast interview scheme. By leveraging AI tools like Auto GPT, he was able to identify and avoid scams, despite the effectiveness of a scam that evokes emotions of excitement in the target. Wes shares how in his experience of nearly falling for this podcast scam, he was only rescued by his wife's skepticism, and of course, Auto-GPT.In the concluding segment, Wes and Goda discuss effective communication with AI, offering valuable advice on how to maximize AI-driven conversations. Emphasizing the importance of emotion, the hosts encourage listeners to engage thoughtfully with AI and make the most of these interactions.Happy Prompting Everyone!SYZYGY Digital Insight Survey 2023: [https://bit.ly/3HCRHpB]Let us know your score in the Comments: [https://www.caitfished.com/]MidJourney Master Reference Guide: [bit.ly/3obnUNU]ChatGPT Master Reference Guide: [bit.ly/3obo7AG]Learn Prompting: [https://learnprompting.org/]Discord (Goda Go#3156 & Commordore_Wesmardo#2912)Goda Go on Youtube: /@godagoWes the Synthmind's everything: [https://linktr.ee/synthminds]#learnprompting #HTTTA #howtotalktoai #ai #artificialintelligence #promptengineering #prompt #research #nlp #openai #languagemodels #programming #coding #technical #prompthacking #AIsystems #gpt #naturallanguageprocessing #deepreinforcementlearning #chatgpt #gpt4 #stablediffusion #midjourney #midjourneyaiAudio Licensing Codes:- XE3PMH56JJN9LJ8B- ASLC-22DC2994-050B57C5D6

Podcast Page: [https://howtotalkto.ai]HTTTA Newsletter: [https://httta.substack.com/]Have you ever wanted to free up some time to do the things you love? In this podcast episode, we explore how certain AI can help you do just that! Wes & God explored the topic of the new new AUTO-GPT and how it and its derivatives can transform the way we work and interact in our daily lives, and shows the first flashes of intuition and basic reasoning within an AI construct.While discussing Auto GPT, a "Baby Artificial General Intelligence" and how it can save a significant amount of time and effort compared to manually searching for information, Wes shares a tale as to how AUTO-GPT was a boon to what could have been a bad situation. However, Goda strongly advises caution and proper testing before promoting experimental technology as transformative things with a woeful tale about how AUTO-GPT nearly sabotaged months of her audience building on her YouTube channel.Wes is a cautious optimist when it comes to these, who believes in the potential of these tools to free up space and time for doing things we truly enjoy. Despite some concerns over the cost and potential misuse of such tools, Goda points out that this happens with every new technological shift anyway. It's undeniable that using AI tools can even lead to more creativity, as they can provide new ideas and sources that the user may not have considered.The shift in mindset towards using AI to aid in execution rather than replace creativity can be a positive disruption in many industries and not just the gloom of worrying that an AI is going to take your job.Happy Prompting Everybody!MidJourney Master Reference Guide: [bit.ly/3obnUNU]ChatGPT Master Reference Guide: [bit.ly/3obo7AG]Learn Prompting: [https://learnprompting.org/]Discord (Goda Go#3156 & Commordore_Wesmardo#2912)Goda Go on Youtube: [https://www.youtube.com/@godago]Wes the Synthmind's everything: [https://linktr.ee/synthminds]#learnprompting #HTTTA #howtotalktoai #ai #artificialintelligence #promptengineering #prompt #research #nlp #openai #languagemodels #programming #coding #technical #prompthacking #AIsystems #gpt #naturallanguageprocessing #deepreinforcementlearning #chatgpt #gpt4 #stablediffusion #midjourney #midjourneyaiAudio Licensing Codes:- XE3PMH56JJN9LJ8B- ASLC-22DC2994-050B57C5D6

Podcast Page: https://howtotalkto.aiHTTTA Newsletter: https://httta.substack.com/Our intrepid hosts had the pleasure of interviewing Sander Schulhoff, one of the founders of LearnPrompting.org, an open-source website where 800,000 people learn about prompt engineering and how to talk to AI. He started the project as an English class assignment in December 2022, and it has now grown to include a 35,000-person Discord and nearly a million users. Sander shared his methods for prompting and testing out different techniques and has invented two unique prompt injection techniques himself.The Hack A Prompt competition is launching on May 5th with $40,000 in prizes. Participants will attempt to trick AI by deploying different prompt engineering techniques. Prompt hacking is a way to trick AI platforms like chatbots by giving them instructions that could lead them to do something malicious. This competition serves as a way to help the research community understand the spectrum of attacks and suggest defenses to keep models from producing malicious completions. Companies are interested in preventing this with Preamble and OpenAI leading the charge.Learn Prompting's approach to AI education is differentiated by its interactive lessons which allow students to test out their prompts and modify them to see what output it produces, as well as its effort to make content more legible for non-technical audiences. Python is the most popular language used in language model development and AI development due to its simplicity and legibility. There is a global community translating the LearnPrompting course into multiple languages, with 9 and counting.Programming skills can be helpful, but not always necessary however to start an AI journey. Sander explains their certification exam, which covers the fundamentals of prompt engineering like creating images and analyzing text to answer questions like tracking people and solving bank robberies. Find Sander at learnprompting.org and their Discord https://discord.gg/learn-promptingHappy Prompting Everybody!MidJourney Master Reference Guide: bit.ly/3obnUNUChatGPT Master Reference Guide: bit.ly/3obo7AGStanford 2023 AI Index Report: https://aiindex.stanford.edu/report/Learn Prompting: https://learnprompting.org/Discord (Goda Go#3156 & Commordore_Wesmardo#2912)Goda Go on Youtube: https://www.youtube.com/@godagoWes the Synthmind's everything: https://linktr.ee/synthminds#learnprompting #HTTTA #howtotalktoai #ai #artificialintelligence #promptengineering #prompt #research #nlp #openai #languagemodels #programming #coding #technical #prompthacking #AIsystems #gpt #naturallanguageprocessing #deepreinforcementlearning #chatgpt #gpt4 #stablediffusion #midjourney #midjourneyaiAudio Licensing Codes:- XE3PMH56JJN9LJ8B- ASLC-22DC2994-050B57C5D6

Podcast Page: [https://howtotalkto.ai]HTTTA Newsletter: [https://httta.substack.com/]Your intrepid hosts discuss the power of AI and its applications in the job market. It was found that 1-2% of job listings in the US and EU are AI-related, and this number is expected to increase. The host discusses the recent AI Index report from Stanford University and suggests checking it out for more insights.AI is quickly becoming prevalent in many different industries and job titles, but it's not necessarily evident in every activity within those jobs. AI is replacing many jobs, but it can also be used to help people become more productive and efficient. They share their perspectives on "jailbreaking" AI tools and prompt injections, which can hijack a language model's output or make a prompt interactive. Prompt injection is a technique used to unlock interesting conversations with language models like GPT-3, allowing for the creation of unique outputs.Prompt injection can be used for things like creating YouTube titles, but it must be done responsibly, as the user is liable for the output. Selling prompts online is a new and largely unregulated industry. It is possible to buy prompts and resell them.It can be empowering to list simple prompts online, as many of them can be found for free on the internet if one is willing to search. Those who want to sell prompts should be encouraged, as it is a great way to make money. Selling prompts can be a path to small profits, but a profit nonetheless.Is it possible to create an environment where users pay a subscription fee to access a prompt with instructions that generates output? AI tools can give people 10x productivity, meaning businesses can hire more people to perform jobs and use AI tools to increase productivity.A discussion on how AI can improve mental health can also benefit from using AI tools, as one can avoid tedious parts of the job. Developers reported improved productivity and mental health when using a GPT Co-Pilot code assistant.They end with a story about the power of a mid-journey image generation engine prompt that led to how tech can be used to honor memories of lost loved ones.Happy Prompting Everybody!MidJourney Master Reference Guide: [bit.ly/3obnUNU]ChatGPT Master Reference Guide: [bit.ly/3obo7AG]Learn Prompting: [https://learnprompting.org/]Discord (Goda Go#3156 & Commordore_Wesmardo#2912)Goda Go on Youtube: [https://www.youtube.com/@godago]Wes the Synthmind's everything: [https://linktr.ee/synthminds]#learnprompting #HTTTA #howtotalktoai #ai #artificialintelligence #promptengineering #prompt #research #nlp #openai #languagemodels #programming #coding #technical #prompthacking #AIsystems #gpt #naturallanguageprocessing #deepreinforcementlearning #chatgpt #gpt4 #stablediffusion #midjourney #midjourneyaiAudio Licensing Codes:- XE3PMH56JJN9LJ8B- ASLC-22DC2994-050B57C5D6