π¨π€π Exciting news! Sequoia just dropped a study on the Language Model Stack, and it shows rapid growth in AI applications. π€― 1/9
π The study shows that nearly every company in the Sequoia network is building language models into their products, with 65% already in production. #sequoia#languagemodelstack 2/9
π» Companies are using language model APIs for everything from auto-complete features for code to chatbots for customer support. And OpenAI's GPT is the clear favorite foundation model API, with 94% of companies using it. #AI#chatbots
3/9
π Retrieval mechanisms, like vector databases, are also key parts of the stack, with 88% of companies using them. And 38% are interested in an LLM orchestration and application development framework like @LangChainAI. #development#frameworks 4/9
π But the most exciting part? Companies are customizing language models to their unique context in three ways: training a custom model from scratch, fine-tuning a base model, or using a pre-trained model and retrieving relevant context. #customization#Aiwanger
5/9
π Keep an eye on the market as it evolves, and don't be afraid to train your own language model from open source. With these insights, you can stay ahead of the curve in the world of AI. #DataScience#Aiwanger
6/9
π€ As more companies adopt LLM applications, new opportunities, and hurdles will arise, creating exciting potential for founders. If you're interested in getting involved in the AI industry, now is the time to act and seize these exciting opportunities. #entrepreneurship
7/9
ππ½ So what does this mean for developers? Use pre-trained models and retrieval methods to make unstructured data easily searchable. And language models will become increasingly multi-modal, unlocking a new level of conversational experience. #development#multimodal
8/9
π It's still early, but this is an exciting time for language models and their potential. With AI seeping into every crevice of technology, it's important to stay informed and ahead of the curve in order to succeed in this rapidly evolving industry. #AI#SuccessionHBO 9/9
Thatβs a wrap!π¬ Hope this was helpful!π€
If you enjoyed this thread, please:
1. Follow me @_aigeek for more amazing contentπ
π¨ GPT-4 has outperformed human entrepreneurs in a pitch competition! Investors and business owners found GPT-4 generated pitch decks to be more compelling, convincing, and of higher quality than those created by humans. #GPT4#pitchcompetition
A quick recapπ
π€π€― This game-changing solution not only saves time in creating a pitch deck but also has the potential to be a highly lucrative option. Entrepreneurs should consider using GPT-4 to increase their chances of securing funding and impressing investors. #AI#investing#funding
π‘The implications are clear: if you want to captivate investors, you should reconsider writing your own pitch decks. GPT-4 has unveiled a groundbreaking opportunity to disrupt the traditional approach to securing funding. #Pitching#Investment
Bessemer Venture Partners is investing $1 billion in #AI because they believe it will fundamentally change the way billions of people work, creating trillions of dollars in value. This is a seismic shift with enormous potential! @BessemerVP 1/4
Sameer Dholakia, a partner at Bessemer, predicts that AI adoption will be mind-blowingly fast and easier to adopt than previous platform shifts like the cloud. They predict many winners in the AI landscape, with growth in verticalized LLMs specialized for specific tasks.
2/4
Bessemer's portfolio companies, Jasper and DeepL, are already using generative AI and scaling quickly. Other portfolio companies have adopted AI into their products, like Intercom's Fin using GPT-4, Ada's integrated chatbot, and Canva's five generative AI-based products.
3/4
ππ€ The future of AI is here! Google CEO, Sundar Pichai, recently spoke at Google I/O about how AI can be more helpful for people, businesses, and communities. Let's explore some highlights from his speech! #FutureofAI@sundarpichai@GoogleAI
A quick recapπΏπ
ππ With generative AI, Google is reimagining all its core products, including search. Pichai gave examples of how generative AI is already helping evolve products like Gmail and Maps. #AIadvancements#GoogleMaps#Gmail
π§πΊοΈ In Gmail, Help Me Write uses AI to draft emails for users, and in Maps, Immersive View for Routes allows users to see their whole trip in advance. These advancements in AI will continue to improve and make our lives easier. #HelpMeWrite#ImmersiveView
π In the Age of AI, personalized customer experiences are crucial for competitive advantage. Companies like Home Depot, JPMorgan Chase, and Starbucks are leading the way. #CustomerExperience#AI#Personalization@HBR
A quick recapπΏπ
π Competitive advantage is based on the ability to capture, analyze, and utilize personalized customer data at scale. AI is the key to understanding, shaping, customizing, and optimizing the customer journey. #Data#AI#CustomerJourney
2/7
π‘ Brands like Sweetgreen and Stitch Fix have designed transformative first-party, data-driven experiences as well. Personalization is the design target for every touchpoint. #ChallengerBrands#Personalization#DataJustice2023
3/7
Generative AI is the next productivity frontier! According to a recent @McKinsey study, it has the potential to add trillions of dollars πΈπΈ in value to the global economy. Are you ready to learn how it can impact your business? #GenerativeAI#Productivity
A quick recapπ
πWith the highest impact on customer operations, marketing and sales, software engineering, and research and development, generative AI can significantly increase labor productivity, contributing to economic growth. #AIinBusiness#EconomicPotential
π°The retail and consumer packaged goods industry could see an additional 400 to 660 billion in value annually, while the banking industry could see an additional 200 to 340 billion in value annually. That's a huge opportunity for your business! #ProductivityFrontier
Have you heard about Otter: A Multi-Modal AI Model with In-Context Instruction Tuning? This cutting-edge model leverages a large-scale dataset called MIMIC-IT to achieve state-of-the-art performances in perception and reasoning benchmarks. @Microsoft@NTUsg#AI 1/5
π€ Otter is not your average AI model. With 2.8 million multimodal instruction-response pairs, Otter demonstrates remarkable proficiency in multi-modal perception, reasoning, and in-context learning. It effectively aligns with the user's intentions. #MultiModal#AI
π The MIMIC-IT dataset, instruction-response collection pipeline, benchmarks, and the Otter model are all available for use. This means that Otter can be applied to a variety of industries, from healthcare to finance, to improve perception, reasoning, and planning. #DataScience