#16 All of This Has Happened Before, All of This Will Happen Again
Burhan Sebin
Head of Research at Venture Miami | leading the largest tech incubator in Florida | shaping Miami as an AI hub
"All of this has happened before, and all of this will happen again."
Sam Altman tweeted this quote last week. I was unaware that it's a famous line from Peter Pan, nor do I know why Sam shared it. However, a couple of days ago, I watched a video of Steve Jobs introducing the first iPhone in 2007. The audience's reaction—screaming, applause, etc.—as he unlocked the phone by sliding the button on the touchscreen was electric.
Now, in 2024, look at where we are with technologies like ChatGPT, LLMs, large vision models, Sora, Midjourney, etc. I remember the day ChatGPT was released; I was so engrossed that I couldn't sleep for three straight days, bombarding it with every question I had and trying out thousands of scenarios. However, after a few weeks and months, I began to notice and complain about its inaccuracies.
I'm sure you've found reasons to complain about ChatGPT, Gemini, or Claude as well.
We humans adapt so easily to new advancements and invariably crave more. This adaptability is instinctive; we are survivors by nature.
Just as I'm certain people were both terrified and fascinated when cavemen invented the first primitive tools. This duality of fear and amazement at new discoveries has always been a part of our evolution.
I'm convinced that people were amazed when they experienced the first radio broadcast, watched the first TV show, witnessed the first step on the moon, accessed the internet for the first time, or used the first mobile app. Each innovation seemed like it would change the world—and indeed, many of them did. Yet here we are, still working, still finding purpose, and still thriving.
Yes, AI poses risks, and AGI will undoubtedly change the world. We will be amazed at first sight, but after some time, we will yearn for more and strive to achieve more. We will find solutions and continue to accelerate our progress.
All of this has happened before, and all of this will happen again.
What I Read This Week in Artificial Intelligence
Anthropic introduces Claude 3, a family of AI models that reportedly surpasses OpenAI's GPT-4 in performance, including a multimodal capability to analyze text and images. Claude 3 offers various models, with improvements in conversation understanding and output structuring, aiming to address common GenAI issues like bias and hallucinations.
Inflection announces Inflection-2.5, an AI model rivaling GPT-4, designed to enhance the capabilities of their personal AI, Pi. With improved performance in STEM fields and web search, it aims to transform user interactions and has already seen significant user engagement.
Stable Diffusion 3 outshines DALL·E 3, Midjourney v6, and Ideogram v1 in typography and prompt adherence. With its new Multimodal Diffusion Transformer (MMDiT) architecture, SD3 offers improved text understanding. Available in variations from 800m to 8B parameters, SD3 introduces flexible text encoders and advanced Rectified Flows for efficient image generation.
Midjourney claims Stability AI employees caused a service outage, alleging data theft, leading to all Stability AI employees being banned from Midjourney.
Elon Musk and Sam Altman's once collaborative relationship has devolved into a legal and public dispute over OpenAI's shift from a non-profit aiming to benefit humanity to a for-profit entity. Musk's lawsuit accuses OpenAI of breaching their founding agreement, while OpenAI refutes Musk's claims, suggesting professional jealousy might be a factor. This conflict highlights broader concerns about AI's development and its potential impact on society.
Microsoft's Orca-Math, a specialized small language model, outperforms larger models like Llama 2 and GPT-3.5 in grade school math problems with a 86.8% success rate, showcasing the efficiency of models trained on specific tasks.
DeepMind alums launch Haiper, an AI video generation tool aiming to compete with OpenAI's Sora. Haiper has raised $13.8 million in seed funding, focusing on consumer engagement and exploring commercial applications.
领英推荐
OpenAI introduces a Read Aloud feature for ChatGPT, allowing responses to be read in one of five voices across 37 languages on web and mobile platforms.
Answer.AI releases an open-source system enabling the training of a 70 billion parameter AI model on consumer-grade GPUs. This advancement, developed in collaboration with Tim Dettmers, U Washington, and Hugging Face, utilizes FSDP and QLoRA, making high-capacity model training accessible and affordable.
Cohere partners with Accenture, leveraging a $3B AI investment to scale generative AI in enterprises, focusing on privacy and security. Cohere's AI models will enhance productivity and efficiency across various sectors.
AI Tools Worth Your Attention
Huudle.io : Connect all your meetings and messages around the same topic in a single space, track key points via AI generated short highlights, video message your short updates on any progress, let AI follow up the actions.
Heyday: Heyday is an AI copilot that transforms your documents, notes, and conversations into quotes, shareable content, and a queryable database.
Sider: Sider is your AI sidekick, seamlessly integrating into your daily workflow. It starts as a Chrome/Edge extension, making browsing, reading, and writing easier.
Miami AI Hub Presents Generative Gatherings in April!
Miami AI Hub 's Generative Gatherings: Featuring Noelle Russell (Microsoft MVP, AI) ??
Join us at Mindspace Wynwood on April 2nd for an electrifying evening at the heart of Miami's AI innovation. Dive deep into the world of artificial intelligence with our distinguished guest speaker, Noelle Russell. As a Microsoft MVP and a trailblazer in data, cloud technology, conversational AI, Generative AI, and LLMs, Noelle's insights are not to be missed.
Event Highlights:
This is your chance to connect with like-minded professionals and visionaries in the AI space, learn from one of the best, and discuss future AI trends and collaborations.
Spaces are limited. Secure your spot now for a night of AI innovation and networking. Let's shape the future of AI in Miami together!
Register now:
And... cut! ?? That's a wrap on this edition's AI extravaganza. See you on the next one! Don't forget to like, share, and subscribe!
Senior Manager | Legal Counsel
8 个月Great read!
Chief Economic Development Officer - City of Miami/Venture Miami | Chairman – Advanced Air Mobility Committee | Mayor's Representative – Miami-Dade Beacon Council | Mbr - Chicago Taskforce on Innovation & Technology
8 个月Thanks for sharing the insights!
Competitiveness Manager at Alstom | Fashion-Tech Enthusiast | M.Sc. Industrial Engineering at TU Berlin
8 个月It's intriguing to observe similar patterns of awe and skepticism with each technological advance, from early tools to AI. Human capacity for adaptation is astounding! ??
????