We are excited to announce Radical’s newest portfolio company, Twelve Labs, whose seed raise was covered last week in TechCrunch. Twelve Labs is building next-generation video search technology by fusing cutting-edge natural language processing and computer vision. Think “Ctrl+F for video.” We believe multimodal approaches to AI like Twelve Labs’ use of both NLP and computer vision represent the future of artificial intelligence.
Video has become the dominant medium for our digital lives. 80% of the data on the internet today is video. Yet, remarkably, there is no effective way to search through all that video data to find what you’re looking for: a particular moment, individual, or concept. Twelve Labs is out to change that search experience.
There are countless use cases for high-accuracy video search: from social media to streaming content, from digital asset management to workplace productivity, from content moderation to cloud storage. Twelve Labs makes its video search technology available via API, so that it can be applied across industries and applications.
Radical invested alongside Index Ventures in Twelve Labs’ $5M seed round. We introduced Fei-Fei Li, AI luminary and close friend of Radical Ventures, to the company and she also participated in the round.
5 Noteworthy AI and Deep Tech Articles: week of March 21, 2022
1) The AI Index Report – Artificial Intelligence Index (Stanford University)
AI made the jump from an emerging technology to a mature one in 2021. AI is being integrated into the economy, and the effects are showing across AI funding, research, and deployment. Some findings in the report include private investment in AI more than doubling from 2020 to 2021, a 5x increase in publications on AI fairness and transparency in the past four years, and AI systems becoming less expensive and time-intensive to train. Since 2018, the cost to train an image classification system has decreased by 64%, while training times have improved by 94%. Robotics arms also became 4x cheaper in the same time frame. Stanford’s HAI also released its interactive Global AI Vibrancy Tool to visualize which nations are leading the global AI race based on different indicators. Users can compare 29 countries across 23 indicators.
2) ‘No-Code’ brings the power of AI to the masses (New York Times – subscription may be required)
The “no-code” AI revolution will empower a world of citizen science. While not professional scientists, citizen scientists are curious or concerned people who collaborate with professional scientists in ways that advance scientific research on topics they care about. The citizen developer is able to “create AI-enabled software in much the same way that teenagers today can create sophisticated video effects that would have required a professional studio a decade or two ago.” Radical Ventures portfolio company Cohere is part of this technological revolution. Through a simple API, Cohere’s platform enables users to leverage the world’s most powerful NLP toolkit to build products that can read and write.
3) Using AI, people who are blind are able to find familiar faces in a room (Microsoft Research)
Microsoft Research is testing a new research technology to help people who are blind (called learners in the study) and their peers interact more easily. Learners leverage AI to build a ‘map’ and track the people around them in a space and assist in effectively signalling communicative intent. The technology, in turn, indicates to the learner’s peers when the peers have been ‘seen’ and can interact — a replacement for the eye contact that usually initiates an interaction between people. Most recently, the system has been applied to children who are blind in school settings, supporting their skills in proactively interacting with classmates. The innovation explores how we can design human-AI interaction that moves beyond discrete task support to provide a continuous stream of dynamic information.
4) Will AI turbocharge the hunt for new drugs? (Financial Times – subscription may be required)
In the last five years, AI’s application to drug discovery and repurposing has accelerated. AI has made meaningful progress despite its many challenges and “all data suggests AI is here to stay.” During the pandemic, AI was primarily used to save scientists’ time, accelerating the notoriously slow drug discovery process. AI optimized Covid vaccines, vastly sped up pharmaceutical companies’ ability to search their databases and identify new potential medicines, and designed drugs (still in various trial phases). The pandemic highlighted two key hurdles for AI and drug discovery: making sense of incomplete and disaggregated data, and changing the strategies of large pharmaceutical companies that have been cautious toward AI. However, larger pharmaceutical companies are becoming more interested in AI investment marking a break from the past with fewer barriers for the future of AI and drug discovery.
5) Presidential candidates’ computer-generated avatars heat up debate (KoreaTimes)
Yoon Suk-yeol, who won South Korea’s March 9 election, campaigned using videos that featured an AI-generated likeness of himself answering voters’ questions. No deception was involved. Viewers were informed that they were watching a computer animation. Yoon Suk-yeol is known for his direct way of speaking. The avatar was designed to be more charismatic to appeal to young voters. Yoon’s avatar was created using only 20 hours of audio and video of the candidate captured in front of a green screen, totalling around 3,000 spoken sentences, according to France24.
Share this edition of Radical Reads