Simon Smith’s Post

Here's a summary for the recent OpenAI spring update - ✨ 1) GPT-4o, a model that equals GPT-4 except is also much faster and will be free for everyone using ChatGPT. 2) GPT-4o on the API will be 50% cheaper than GPT-4. 3) Audio conversational capabilities of GPT-4o are exponentially better. 4) You can talk to GPT-4o through your live camera feed, (e.g., you can have a conversation while solving a math equation on paper and displaying it to ChatGPT as you do.) 5) ChatGPT will also have a desktop app, which can see your screen real-time allowing you to have a conversation as you make changes. Introducing GPT-4o, OpenAI 's new model which can reason across text, audio, and video in real time. It's extremely versatile, fun to play with, and is a step towards a much more natural form of human-computer interaction (and even human-computer-computer interaction) Stay in touch with Geek Room for more updates ✨

Here's a summary of OpenAI's spring update in case you missed it, I highly recommend watching the video if you have the time though. ⛳ A new flagship model, GPT-4o, combines GPT-4-level intelligence with improved speed and capabilities across text, voice, and vision. Notably, it excels in understanding and discussing images, enabling tasks like menu translation and providing recommendations. ⛳ ChatGPT now supports over 50 languages, enhancing accessibility for users worldwide. ⛳ GPT-4o is being gradually rolled out to ChatGPT Plus and Team users, with availability for Enterprise users forthcoming. Free users will also have access to GPT-4o, but with usage limits. ⛳ ChatGPT Free users now gain access to advanced features, including GPT-4 level intelligence, web responses, data analysis with chart creation, photo discussion, file uploads for assistance, GPT Store access, and Memory integration. However, there will be messaging limits, and exceeding them will switch the conversation to GPT-3.5. ⛳ A new ChatGPT desktop app for macOS is introduced, aiming to streamline user workflow. This app allows users to interact with ChatGPT via voice conversations, screenshots, and text queries directly from their computer. ⛳ Voice Mode, initially available in the desktop app, will be enhanced with GPT-4o’s audio and video capabilities in the future. ⛳ The ChatGPT interface receives a friendlier and more conversational makeover, featuring a new home screen, message layout, and more. Report: https://lnkd.in/eWKJNBrK

OpenAI just rolled out GPT-4o, a solid upgrade from GPT-4. Here’s what’s cool about it: 🔸 It’s got pretty realistic: GPT-4o makes ChatGPT way more interactive. You can now talk to it like you’re having a real conversation. Plus, it can handle visuals, so you can show it screenshots or photos and chat about them. 🔸 Personalization: one of the biggest changes is the memory feature. GPT-4o can remember what you talked about before, which makes it feel more personal. It’s like having a conversation that picks up right where you left off. It also supports over 50 languages, which is a huge plus. 🔸 Better user experience: compared to GPT-4, this version feels a lot more natural. The voice mode is impressive—it’s almost like talking to a human. It can even detect emotions and respond in a calming way. 🔸 Accessibility: the new desktop app and developer options are great. Whether you’re just using it for fun or building something, GPT-4o is more accessible. Free users get a taste of the new features, though there are some limits. 🔸More to come: with Google and Meta hot on their heels, it’s clear OpenAI is stepping up their game. The competition is heating up, but GPT-4o definitely sets a new standard. Overall, it’s a big leap forward. If you’re already using #chagpt, you’ll notice the difference. And if you haven’t tried it yet, now’s a great time to start ⏳ #ai #openai #ainews #aicommunity 

Do you know the difference between OpenAI, ChatGPT, and GPT-4? (Let’s not even get into “o1” at this stage, shall we?) Let’s untangle some parts of the GPT web… • OpenAI: The company that developed the GPT family of large language models (LLMs). • ChatGPT: Initially a web app, now also available as stand-alone apps for iOS, macOS, and more. It’s a user-friendly interface built by OpenAI to interact with their GPT models. • GPT-3.5, GPT-4, GPT-4o, GPT-4o-mini, o1, o1-mini, etc.: These are different GPT models trained by OpenAI, forming the core technology behind ChatGPT and other AI tools. And of course there are other players in the LLM game: • Anthropic: The company behind Claude, their family of LLMs. • Meta: The company behind LLaMA (cleverly named after their “Large Language Model Meta AI” family of tools). • Google: Creator of the Gemini family of LLMs. And here’s the question - Copilot to Microsoft is like what to OpenAI? #LLM, #AI, #Digital

New OpenAI Model Time. They've just made a variety of announcements. GPT-4o A new model (unfortunately still no word on GPT-5). It has the same 'intelligence' as GPT-4, with improved performance. But will come with text,audio and vision capabilities built into a single model. Will remove a lot of the latency that you will have experienced when these individual models have been put together. It's also going to be 50% cheaper than GPT-4-Turbo (5x faster too). Desktop App: A new desktop app will be released, available for everyone. Demo implied that you'll be able to feed your screen to the app and it can read what's on it, analyse it as well as being able to talk about it. GPT Store + GPTs: These were hyped originally, kind of fell flat, not even revenue sharing in America could revive them. But they're now being made available to free users (instead of being restricted to ChatGPT+ subscribers) Other interesting notes: Was able to detect changes in voice, can change it's own voice to suit the circumstance. This alone opens up some new opportunities for it's use. A softer voice for a childrens bedtime story, or a more authoritative voice for an instructor. If you ever want to talk about all things #AI or #GenerativeAI. Interesting use cases, questions or ideas. I'm always open to talk about it. Be glad you don't work with me and have to hear about it almost everyday....

OpenAI's o1 - Thinketh before thou speakest !! OpenAI has launched a new model named o1 today, the first in a planned series of “reasoning” models designed to tackle more complex questions more quickly than humans. Now literally a model can "Think (aloud) before it speaks". The o1 model series employs large-scale reinforcement learning and a "chain of thought" approach to enhance reasoning abilities, safety, and robustness. This model, therefore, hallucinates less. Usage: 👉 Complex Code Generation: Facilitates algorithm creation and advanced coding tasks to support developers. 👉 Advanced Problem Solving: Ideal for in-depth brainstorming and addressing intricate issues. 👉 Complex Document Comparison: Useful for analyzing contracts, case files, or legal documents to identify subtle differences. 👉 Instruction Following and Workflow Management: Excels in managing workflows that involve shorter context. ⚠️ Challenges: 🔥 High cost: More expensive than GPT-4o. In the API, o1-preview is priced at $15 per million input tokens (chunks of text processed by the model) and $60 per million output tokens. In contrast, GPT-4o costs $5 per million input tokens and $15 per million output tokens. 🔥 Performance trade-offs: Slower, doesnt do well with factual knowledge and doesn’t have the ability to browse the web or process files and images. 🕒 Availability: ✔ Available to ChatGPT Plus and Teams users now. ✔ Enterprise & Edu users get access next week, with future rollout for free users. (OpenAI says it plans to bring o1-mini access to all the free users of ChatGPT but hasn’t set a release date yet.) For OpenAI, o1 marks a significant step toward achieving human-like artificial intelligence. #GenAI #GenerativeAI #OpenAI #o1 #GPT #AI (Image Source: AIM)

Leveling Up My Information Retrieval Game with AI! Learning project on OpenAI's LLM, 𝐅𝐢𝐧𝐒𝐢𝐠𝐡𝐭 - a user friendly news research tool designed to streamline financial research! 📋Features: ✔️Load URLs or upload text files containing URLs to fetch article content. ✔️Process article content through LangChain's UnstructuredURL Loader ✔️Construct an embedding vector using OpenAI's embeddings and leverage FAISS, a powerful similarity search library, to enable swift and effective retrieval of relevant information ✔️Interact with the LLM's (ChatGPT) by inputting queries and receiving answers along with source URLs. About LangChain: 🔎Langchain is an open-source framework designed to simplify the process of building applications that utilize Large Language Models (LLMs) like ChatGPT. It tackles the messy world of unstructured text data – think articles, emails, social media posts 🔎Streamlit: To create a user-friendly chatbot interface ✏️Uncovering hidden connections: OpenAI embeddings and FAISS unlock powerful similarity search, helping you discover relevant insights buried within financial data. ✏️Answering your questions: Ask 𝐅𝐢𝐧𝐒𝐢𝐠𝐡𝐭 anything finance-related, and receive insightful answers powered by ChatGPT - New!, complete with source citations for credibility. 📜Key Takeaways: Customization: LangChain allows for customization of the chatbot experience, including integrating with APIs for additional functionality. Development Efficiency: LangChain simplifies the development process by providing a streamlined workflow for building LLM-powered chatbots. 𝐆𝐢𝐭𝐇𝐮𝐛 𝐋𝐢𝐧𝐤 : https://lnkd.in/dPdsmvwV 𝐇𝐚𝐩𝐩𝐲 𝐋𝐞𝐚𝐫𝐧𝐢𝐧𝐠😊 #openAi #Langchain #streamLit #LLMmodels #FAISS #LearningTechnology #exploringgenerativeai

Breaking 🚨 : OpenAI just announced GPT-4o Their best AI model ever Here's everything you need to know: Save this post for future reference --- They have a new desktop app now and a cleaner UI --- GPT-4o 1. It is smarter, it is faster than GPT-4 , and it is natively multimodal. 2. Best part: It is available to all ChatGPT users, including on the free plan! 3. It was the gpt-2 chatbot trending recently and it ranks at top compared to other LLMs 4. It performs extremely well at coding too as compared to other LLMs out there 5. For people using the API, GPT-4o is half the price and 2x faster than GPT-4-turbo And with 5x rate limits 6. You can also interact with it with your voice when you ask it to help with tasks like coding, translation, analysing your expressions, and more. --- Note: ChatGPT Plus users can access it now Also, OpenAI will have more updates to share soon ---  Post credit Shushant Lakhyani Video Credit: OpenAI

OpenAI just annouced some really cool stuff and it feels like we're living in the new era! Quick summary - What’s new - Desktop version of ChatGPT - Refreshed UI (seems like minor change though) - New model GPT-4o — “o” for Omni - GPT-4o available to everyone (yes, including free users!) - GPT-4o reasoning across voice, text and vision (real-time multiple inputs in at the same time - explained below) - Custom GPTs can be used by everyone (not just paid users) - Improvement to 50 languages What this means for free users: - GPT-4 series - the latest and most advanced model by OpenAI is available to you - Custom GPTs and GPT Store is available to you - Make use of Memory capabilities With the custom GPT becoming freely available to everyone, this also means like content creators, "GPT creators" will become a thing soon. So why have a free account you ask? Paid users gets up to 5 times the capacity limit of free users. For Free users, once it hits the limit, it switches from GPT-4 to GPT-3 Turbo automatically. What’s more exciting is what’s releasing with GPT-4o though.. Think Siri but 10x better. - The voice is real-time in responsiveness — it can respond to audio inputs in 232 milliseconds, with an average of 320 milliseconds similar to human response time! - Turn on the camera and provide the live feed of the visuals as input - Added “emotions” - It can sing to you - With the Desktop app, imagine it watching your screen and interact with you while your write content or code - You can do live translation - Detect emotion on your face from a live video and much more.. Game changing indeed. The new model is going to give us more natural way to interact with computers. Before 4o, text, audio and visuals are processed separately thus losing the full context but now it combines all 3 and is processed at the same time by the neural network. Imagine doing a video call with ChatGPT and it can analyze your surrounding, the emotions on your face while you’re talking to it in real time. And with that, process and respond fast like a human. I’m keen to see the response from other tech giants on this. Exciting weeks ahead! #EmbraceAI

🔍 Are you aware of the latest tech showdown? Microsoft is pushing Bing's AI-powered chat bot, GPT-4, in a bid to challenge Google's dominance. 💻 But, how does this affect us? Well, AI is revolutionizing industries, including healthcare! For chiropractors, AI can automate complex tasks like lien filing, saving you time and reducing errors. 📚 Want to explore more AI applications? Visit www.gptsplayground.com, a community-driven platform listing the best GPT apps and assistants. There, you can discover, vote, and even list your own GPTs! 🌐 As the tech giants clash, remember, the real winner is us, the users, who get to harness the power of AI. Stay updated, stay ahead with GPTs Playground. 💡 #AI #GPT4 #Microsoft #Google #Chiropractic #Automation #GPTsPlayground