GSV's AI News & Updates (02/16/24)
OpenAI's Sora Text-to-Video, Gemini 1.5 Upgrade, V-JEPA Learns by Watching, Slack's AI Features, Watermarking AI Images, Scribe $25M Series B, Colossyan raises $22M
AIR (AI Revolution) Show is the world’s leading event at the frontier of AI in EDU for educators, administrators, professionals, and AI explorers. There will be demos, exhibits, booths, workshops, hackathons, panels and more! REGISTER here now, tickets are free but capacity is limited.
General 🚀
OpenAI introduces Sora, their Text-to-Video model: this latest model takes text prompts and turns them into ‘complex scenes with multiple characters, specific types of motion’ (up to 1 minute). The model can understand how objects “exist in the physical world,” as well as “accurately interpret props and generate compelling characters that express vibrant emotions.” The model can also generate a video based on a still image, as well as fill in missing frames on an existing video or extend it. It’s not yet available to the general public, but the announcement said it is going through testing with “red teamers” who are assessing the model for potential harms and risks.
Why this matters: I don’t know whether I am impressed or terrified that we can now create synthetic videos from thin air. This does feel like another ChatGPT moment (it’s crazy to think that this all progressed in one year!). It's a potential step function change for learning, creativity, personalization, storytelling. But there’s also *absolutely zero* ways in which this technology might be abused…this year’s election? Who do we turn to for truth? Huge changes coming to the video game and film industry? Do half of the YC batch of startups now have to pivot? Goodbye reality?
Google unveils upgraded Gemini 1.5 model: even though Gemini 1.0 just launched in December, the new model boasts massive upgrades from its predecessor, including a longer context window (groundbreaking 1M tokens), better understanding, and an overall enhanced performance. The model can now understand entire books, full movies, and podcast series' all in one go.
Meta’s new V-JEPA model learns by watching videos: a new model that’s trained in a similar way to today’s large language models, but instead of learning from written words, it learns from video. “Our goal is to build advanced machine intelligence that can learn more like humans do, forming internal models of the world around them to learn, adapt, and forge plans efficiently in the service of completing complex tasks.”
Slack's highly anticipated AI features are finally here: including AI powered search, channel recaps, thread summaries, and more. “These new AI capabilities empower our customers to access the collective knowledge within Slack so they can work smarter, move faster, and spend their time on things that spark real innovation and growth”, said Denise Dresser, Slack CEO.
YouTube’s 4 Big bets for 2024 are all focused on AI: YouTube paid over $70B to creators from 2020-2023, and is now leveraging generative AI to further empower and democratize content creation, while ensuring AI enhances human creativity.
OpenAI rumored to be developing Web Search product in challenge to Google
OpenAI is testing memory with ChatGPT: ChatGPT can now learn and adapt based on your chats - no more repeating yourself and personalized outputs. ChatGPT automatically learns these new instructions, but you also get to control what it remembers and forgets.
These $349 smart glasses have multimodal ‘AI superpowers’: the open-source eyewear comes from a startup called Brilliant Labs, which touts Frame as a way to get AI translations, web search, and visual analysis right in front of your eyes.
Education and the Future of Work 📚
Scribe raises a $25M Series B: Scribe automatically creates visual step-by-step guides, making it easy to capture and share how-to, everyone can access the info they need to do their jobs well.
Colossyan raises $22M to grow its AI-powered corporate training video production platform: the round follows an impressive year for the company, which grew its revenue by more than 600%, off an undisclosed base, after landing deals with large enterprises.
Early Adopters of Microsoft’s AI Copilot Wonder if It’s Worth the Money: AI aide handles email, meetings and other things, but its price and limited use have some skeptical.
AI Is Starting to Threaten White-Collar Jobs. Few Industries Are Immune: as AI adoption grows, it’s likely to reconfigure management hierarchies. Entry-level workers are likely to bear the initial brunt as more of their duties are automated away. In turn, future entry-level work will look more like first-level management roles. The cascading effect could flatten layers of middle management.
How Walmart, Delta, Chevron and Starbucks are using AI to monitor employee messages: Aware, an AI firm specializing in analyzing employee messages, said its data repository contains messages that represent about 20B individual interactions across more than 3M employees.
Google pledges 25M euros to boost AI skills in Europe: the firm will also run a series of “growth academies” to support companies using AI to scale their companies and has expanded its free online AI training courses to 18 languages.
India’s Upskilling Players on Growth Trajectory as AI and Layoffs Increase Demand for Learning: rapid changes in the job landscape demand continuous learning and adaptability, with upskilling and reskilling emerging as essential strategies. Companies like Eruditus, upGrad, and Simplilearn are meeting this demand
How Teachers Are Pondering the Ethics of AI: fiery debates over AI ethics and its potential to outmode teachers, while uncovering deep divides in how educators grapple with tech's double-edged sword.
Why the Rush toward Generative AI Literacy in K-12 Schools May Be Premature: the breakneck pace of development, coupled with underlying complexities and unknowns, makes it exceptionally challenging to provide evidence-based education.
Startups and Tools 🛠️
OpenAI:
OpenAI on track to hit $2B revenue milestone as growth rockets
Sam Altman Seeks Trillions of Dollars to Reshape Business of Chips and AI: the project could require raising as much as $5T-$7T.
Apple acquired 32 AI Startups in 2023: in the overall AI startup acquisition, Google trails Apple with 21, Meta with 18, and Microsoft lags with 17.
Salesforce to Invest in Together AI, Nvidia Chip Reseller, at $1B+ Valuation
Mindy: your email-first Chief of Staff, raised a $6M Seed from Sequoia and Founder’s Fund
Kompas: sophisticated AI agents at work
AI Creative Content Market Map (a16z):
Tech 💻
Evaluating LLM Applications: unlike traditional software and non-generative machine learning (ML) models, evaluation of LLMs is subjective, hard to automate and the risk of the system going embarrassingly wrong is higher.
Regulation and Policy ⚖️
OpenAI and Microsoft warn of state-linked actors’ AI use: OpenAI said it terminated accounts of five state-affiliated threat groups who were using the company’s large language models to lay the groundwork for malicious hacking campaigns
NVIDIA CEO Jensen Huang: Every Country Needs Sovereign AI: “it codifies your culture, your society’s intelligence, your common sense, your history – you own your own data”
Artificial Intelligence in the News: How AI Retools, Rationalizes, and Reshapes Journalism and the Public Arena: the report analyzes AI's impact on news organizations' editorial, commercial, and technological aspects, and its potential to reinforce their dependency on the tech sector, with implications for the public sphere.
Microsoft launches several partnerships in journalism: with news organizations like Semafor and institutions such as the Craig Newmark Graduate School of Journalism to deploy generative AI in journalism, aiming to enhance research, audience engagement, and sustainability of AI in newsrooms.
Inside the Underground Site Where ‘Neural Networks’ Churn Out Fake IDs: the site, called OnlyFake, threatens to streamline everything from bank fraud to money laundering, and has implications for cybersecurity writ large.
Your AI Girlfriend Is a Data-Harvesting Horror Show: the privacy mess is troubling because the chatbots actively encourage you to share details that are far more personal than in a typical app.
USPTO says AI models can’t hold patents: but ‘the use of an AI system by a natural person does not preclude a natural person from qualifying as an inventor.’
Other
Airbnb plans to use AI, including its GamePlanner acquisition, to create the ‘ultimate concierge’: “Imagine an app that you feel like knows you — it’s like the ultimate concierge — an interface that is adaptive and evolving and changing in real-time, unlike no interface you’ve ever seen before.”
Resources
The GSV BIG 10: weekly coverage corner for the top 10 stories, insights, and major plays in learning and skilling.
SORA! WOAH