GSV's AI News & Updates (07/21/22)
Bard Updates, AppleGPT, Elon Musk Launches xAI, Meta Releases Llama 2 and CM3Leon, FTC Investigates OpenAI, ChatGPT Goes to Harvard, NotebookLM, Teacher ChatGPT Usage
You are receiving this weekly update on the intersection of AI and education, as part of our GSV community – we hope this will be a helpful resource as we navigate this evolving and exciting intersection. Please feel free to send over any relevant resources you may come across (claire@gsv.com) – we’d love to learn and hear from you!
General 🚀
Bard’s new updates: you can now access Bard in 40+ new languages and countries, customize responses (listen to responses or easily adjust Bard’s tone and style of responses), add images to your prompts, boost productivity by pinning chats/sharing chats with friends, export code and more.
Elon Musk announces new AI venture, xAI: its goal is to “understand the true nature of the universe.” The team is headed up by Elon Musk and includes team members that have worked at other big names in AI, including OpenAI, Google Research, Microsoft Research, and DeepMind. Musk has been part of a major AI organization before, co-founding OpenAI in 2015.
Meta releases Llama 2: the next generation of Meta’s open source large language model, free for research and commercial use. Llama 2 was trained on 40% more data than Llama 1, has double the context length, and outperforms other open source language models on many external benchmarks, (reasoning, coding, proficiency, and knowledge tests). This release includes model weights and starting code for pretrained and fine-tuned Llama language models — ranging from 7B to 70B parameters.
Meta announces CM3Leon, their best-in-class new art-generating model: CM3Leon (“chameleon” in clumsy leetspeak) is a transformer model distinguished by being one of the first image generators capable of generating captions for images, laying the groundwork for more capable image-understanding models going forward (vs. modern image generators, including OpenAI’s DALL-E 2, Google’s Imagen and Stable Diffusion, rely on a process called diffusion to create art).
Apple is testing a ChatGPT-like AI chatbot: the tech giant has created a chatbot that some engineers are internally referring to as “Apple GPT.” Apple has yet to determine a strategy for releasing the technology to consumers, but is reportedly aiming to make a significant AI-related announcement next year.
Microsoft Announces Bing Chat Enterprise and Microsoft 365 Copilot pricing: Microsoft 365 Copilot will be priced at $30 per user and Bing Chat Enterprise will be available as a stand-alone offering for $5 per user, per month (included at no additional cost in Microsoft 365).
FTC investigates OpenAI over data leak and ChatGPT’s inaccuracy: The FTC called on OpenAI to provide detailed descriptions of all complaints it had received of its products making “false, misleading, disparaging or harmful” statements about people. The salvo represents the most potent regulatory threat to date to OpenAI’s business in the United States, as the company goes on a global charm offensive to shape the future of artificial intelligence policy.
Shopify Reveals Generative AI ‘Sidekick’ for Shopping: Sidekick can pull info from your store as well as do tasks like redesigning your homepage or putting products on sale.
Superhuman introduces Superhuman AI: turns any idea into an email, fixes errors as you go, summarizes long conversations and threads. Superhuman AI matches the voice and tone in the emails you’ve already sent, applying that to everything it creates.
Wix’s AI Site Generator can create entire websites from prompts: the AI Site Generator leverages Wix’s domain expertise and near-decade of experience with AI to tune the models to generate high-quality content, tailor-made design and layouts. It lets Wix users describe their intent and generate a website complete with a homepage, inner pages and text and images.
The Next Frontier For Large Language Models Is Biology: AI may give us the ability, for the first time in the history of life, to actually compose entirely new proteins (and their associated genetic code) from scratch, purpose-built for our needs.
How to Use AI to Do Stuff: An Opinionated Guide: Ethan Mollick covers the state of play as of Summer, 2023
Actors say Hollywood studios want their AI replicas — for free, forever: “they proposed that our background performers should be able to be scanned, get one day’s pay, and their companies should own that scan, their image, their likeness and should be able to use it for the rest of eternity on any project they want, with no consent and no compensation.”
Education and the Future of Work 📚
National ChatGPT Survey: Teachers Accepting AI Into Classrooms & Workflow — Even More Than Students:
63% of teachers say they’ve used ChatGPT for their job, up 13 points from when asked in February (50%).
Nearly two-thirds of parents think teachers and schools should allow the use of ChatGPT for schoolwork (64%), including 28% who say they should encourage its use.
Teachers use the tool to help with planning lessons and instructional materials (33%), generating creative ideas for classes (30%), and building background knowledge (30%).
ChatGPT goes to Harvard: could ChatGPT-4 pass my freshman year at Harvard?: not only can ChatGPT-4 pass a typical social science and humanities-focused freshman year at Harvard, but it can get pretty good grades. As shown in the report card above, ChatGPT-4 got mostly As and Bs and one C; these grades averaged out to a 3.34 GPA.
Google introduces NotebookLM: an AI-first notebook designed to use the power and promise of language models paired with your existing content to gain critical insights, faster. It’s hard to go from information to insight: Google’s conversations with students, professors and knowledge workers emphasize the difficulty of synthesizing facts and ideas from multiple sources. You often have the sources you want, but it's time consuming to make the connections.
Why AI detectors think the US Constitution was written by AI: Can AI writing detectors be trusted?
The AI Organization: most of the discussion of AI today focuses on 1:1 interactions, but we aren’t talking as much about how AI is going to change organizations. AI will dramatically enhance our collective communication abilities, and this will lead us to completely overhaul the structure of the corporation.
The workers at the frontlines of the AI revolution: the global labor force of outsourced and contract workers are early adopters of generative AI — and the most at risk.
Startups and Tools 🛠️
Voice cloning platform Resemble AI lands $8M: Resemble supercharges your synthetic voice with a text-to-speech generator paired with real-time APIs to build immersive experiences. Resemble has also developed a product, Resemble Detect, that’s designed to validate the authenticity of audio data using an AI model trained to distinguish fakes from real audio.
Hugging Face Is Raising Fresh VC Funds At $4B Valuation: the AI model startup is reviewing competing term sheets for a Series D round that could raise at least $200M at a valuation of $4B.
Air.ai: conversational AI that can perform full 5-40 minute long sales & customer service calls over the phone that sound like a human. And can perform actions autonomously across 5,000 unique applications - kind of like having 100,000 sales & customer service reps at the tap of a button.
Relay: AI-powered portfolio analyzer, driven by your inbox. Extract key details from investment documents and company updates. Remove the manual process from managing portfolio company updates and investment documents.
Formless by Typeform: AI-driven forms that are conversational. The idea was to allow form creators to provide basic instructions, including what information to collect, and let the system handle the rest.
Extend: AI-powered workflows for unstructured data - intelligent data extraction and automation tools built for modern enterprises.
Tech 💻
Replit’s State of AI Development Report: 34000% growth in AI projects, OpenAI's dominance, the rise of open-source, and more
Code Interpreter API: an open-source implementation of the ChatGPT Code Interpreter using LangChain Agents.
Dream-with-Vision-Pro: Text to 3D generation in Apple Vision Pro built with the VisionOS SDK. Type in the text description of the object you envision. This could be anything from an elephant to a sword. Once you've described it, your object will appear before you.
How is ChatGPT's behavior changing over time?: researchers evaluate the March 2023 and June 2023 versions of GPT-3.5 and GPT-4 on four diverse tasks: 1) solving math problems, 2) answering sensitive/dangerous questions, 3) generating code and 4) visual reasoning.
Regulation and Policy ⚖️
China released its official guidelines for generative AI: The move comes following months of consultation between the government and industry players (Alibaba, Baidu, JD.com). The finalized guidelines stipulate that offshore providers of generative AI tools — if they are aimed at Chinese residents — must also comply with the set of rules.
Elon Musk's warning for China: superintelligent AI could take control of the country. Musk shared that following his conversations with leaders in the country, there seemed to be interest in working on a "cooperative international framework" to regulate AI.
AP, Open AI agree to share select news content and technology in new collaboration
AI Safety and Corporate Power – remarks given at the United Nations Security Council
Other Resources:
The GSV BIG 10: weekly coverage corner for the top 10 stories, insights, and major plays in learning and skilling.