Bhupesh Daheria’s Post

OpenAI continues to push the boundaries of what's possible with AI, and the release of ChatGPT-4o marks another significant leap forward. This new model is a testament to OpenAI's relentless innovation and commitment to enhancing our interactions with technology. The advancements in ChatGPT-4o promise even more nuanced understanding, faster responses, and greater versatility. Whether it's for business, education, or creative projects, the potential applications are truly limitless. As someone passionate about AI, it's incredibly exciting to witness and be part of this rapid evolution. Kudos to the OpenAI team for their dedication and ingenuity! Looking forward to seeing how ChatGPT-4o will transform industries and empower users around the globe. #AI #Innovation #TechNews #OpenAI #ChatGPT4o #FutureOfWork #ArtificialIntelligence #MachineLearning #TechTransformation https://lnkd.in/d8_MgyAQ

Today OpenAI announced #ChatGPT 4o (o standing for Omni), a more powerful but cheaper model than GPT4. Not only that, it is available for free (although paid subscribers will have a higher usage limit). On the topic of LLMs, I'm not huge on buzzwords, and following fads, but you're a fool if you're not using all the tools at your disposal to achieve your goals. I have had much success utilizing ChatGPT, to assist with writing code and other infosec related tasks, and it will continue to be a part of my workflow going forward. Sharing some key info from the OpenAI website below: "It can respond to audio inputs in as little as 232 milliseconds, with an average of 320 milliseconds, which is similar to human response time(opens in a new window) in a conversation. It matches GPT-4 Turbo performance on text in English and code, with significant improvement on text in non-English languages, while also being much faster and 50% cheaper in the API. GPT-4o is especially better at vision and audio understanding compared to existing models." Also here is the YouTube link to the press release. https://lnkd.in/e4Wa77H2 #cybersecurity #informationsecurity #ai #llm #genai #infosec

Scalable Extraction of Training Data from (Production) Language Models Researchers have found a way to extract large amounts of training data from machine learning models, even when the training data is not publicly available. This technique, called extractable memorization, can be used to attack a variety of models, including open-source models like Pythia and GPT-Neo, semi-open models like LLaMA and Falcon, and closed models like ChatGPT. #extractablememorization #llm #adversary

Extremely thrilled to have successfully completed the "Artificial Intelligence A-Z™ 2023: Build an AI with ChatGPT4" course put together brilliantly, which clearly explained the concepts of Q-Learning, Deep Q-Learning, Convolutional Neural Networks, Deep Convolutional Q-Learning and the A3C algorithm. I extend my heartfelt thanks to Hadelin de Ponteves, Hon. PhD, Kirill Eremenko, SuperDataScience, Luka Anicin, Ligency and Jordan Sauchuk for their valuable guidance and expertise, and also for providing an extremely comprehensive and engaging course, that enriched my knowledge with a lot of different techniques. Would highly recommend this to anyone who wants a deeper understanding of reinforcement learning. #MachineLearning #ReinforcementLearning #Python #AI #QLearning #ArtificialNeuralNetworks #DeepQLearning #ConvolutionalNeuralNetworks #DeepConvolutionalQLearning #A3C #Udemy

OpenAI has just released the new GPT-4o model. Check it out here: OpenAI GPT-4o I've had a chance to use it briefly, and I must say, it's impressively fast! The language translation demos are particularly awesome. I can already see how useful this will be in various applications and day to day life. One of the highlights for me is the inclusion of multiple Indian languages in the new language tokenizer, with reduced token usage. I'm especially thrilled to see Tamil, my mother tongue, represented. Tamil is not only one of the oldest languages but also has a rich cultural heritage. However, what is not clear is how they are managing data security and privacy. It’s been cleverly mentioned that GPT-4o does not score above medium risk in cybersecurity, CBRN (Chemical, Biological, Radiological, and Nuclear), persuasion, and model autonomy categories. While it's good to see no high risks, medium risks are still a concern. Should we be worried about these medium risks? I would expect more clarity on this front #OpenAI #GPT4o

ChatGPT Can Teach Using Khan Academy Videos Hey everyone! Have you heard the latest? OpenAI just dropped the bomb! GPT-4.0 is breaking the internet with its new voice capabilities. Imagine chatting with AI, and suddenly it's speaking, cracking jokes, and maybe even dropping some beats! This is a wild ride into the future! Even more exciting, we can now use ChatGPT to teach using Khan Academy videos. Imagine an AI explaining complex math, science, or history in an interactive way, tailored to each student’s needs. But let's think about the impact on social interactions. Could these AI tools lead to people prioritizing machines over real-life connections? How can we ensure technology enhances social development? Let’s discuss! Embrace this future and use these tools to improve education and our lives. How do you think AI like ChatGPT will change education and social interactions? Share your thoughts! #AI #OpenAI #GPT4 #KhanAcademy #Future #Education #SocialInteraction #Technology #Innovation https://lnkd.in/dv_QgvA6

Just finished watching the Open AI live discussion for key Chat GPT/GPT4 announcements. Although there was no announcement of Open AI entering the search marketplace yet, there was some exciting updates focused on user experience. The new model update "GPT4-o reasons across voice, vision, and text. My favorite announcements were a desktop is coming for mac users only, sorry windows computer users. I can now share links to my chat threads on social platforms (LinkedIn,Reddit,X,FB) with a public link, and has the option to make public link discoverable for web searches. Finally, last but not least a quick button update (option+space bar) will view your screen and you can prompt it to help with doc outlines, presentations, etc. The new model is available to paid accounts and free users still only have access to the chat gpt 3.5 model which is good for basic requests but no advanced reasoning capabilities. I recommend anyone wanting to get more familiar with Open Ai GPT models give it a quick watch, they did some very cool demos for capabilities #openai #chatgpt #gpt4o #modelupdate #aivoice #aitext #aivision

Elevate your career with our 𝐂𝐞𝐫𝐭𝐢𝐟𝐢𝐞𝐝 𝐏𝐫𝐨𝐦𝐩𝐭 𝐄𝐧𝐠𝐢𝐧𝐞𝐞𝐫𝐢𝐧𝐠 𝐄𝐱𝐩𝐞𝐫𝐭 (𝐂𝐏𝐄𝐄) 𝐂𝐞𝐫𝐭𝐢𝐟𝐢𝐜𝐚𝐭𝐢𝐨𝐧! 🎓  Dive into the world of Large Language Models (LLMs) and learn how to effectively utilize prompts with practical examples. 🎯 Gain expertise in various prompt engineering techniques and explore the practical applications of GPT-4 and the OpenAI API. 🚀 Become a Certified Prompt Engineering Professional and master the latest industry trends and skills for using LLMs in diverse applications. 💪   𝐆𝐞𝐭 𝐂𝐞𝐫𝐭𝐢𝐟𝐢𝐞𝐝 𝐍𝐨𝐰 👉 https://lnkd.in/dXKGvZJc #PromptEngineering #LLMs #CertificationCourse #AI #LLM #PromptEngineer #LearnPromptEngineering #AIEducation #101Blockchains

Did you see the new CHATGPT Updates? -The big news is the launch of GPT-4o, providing GPT-4 intelligence to everyone, including free users. -GPT-4o's intelligence is now available to all users, allowing them to create work, learn, and access advanced tools previously only available to paid users. -The future of AI includes natural language interfaces that could change the paradigm of programming, allowing ordinary people to get complex tasks done without needing programming expertise VIEW ALL UPDATES HERE - https://lnkd.in/e8t7cFJb

***What is an insight? I played with openai to get an answer.*** "True insights enable us to predict behaviors and outcomes, enhancing decision-making processes and strategies." a sentence from a response written by #chatgpt4 to my prompt to define the term "insight" in strategy work. Basic but not inaccurate! I am impressed. I was particularly concerned with fake data that sometimes I have seen #chatgpt4 creates and relies on. What strategies have you used to define an "authentic" and "reliable" source in your prompts? I used some of my background in logics from my computer science days to explain concepts without falling into signification loops! #semiotics for #ai #logics #syntax #promptengineering