| Here's a look at your daily AI briefing: - OpenAI unveils AI video generator Sora.
- Google debuts latest version of large language model
- Apple incorporating AI in its programming software.
- Nvidia discloses stakes in AI companies.
- Amazon AGI team develops world's largest text-to-speech model.
Beth
p/beth-duckett | |
| 1 | | OpenAI announced Sora, its new AI system that generates high-def videos from text prompts. Sora, the Japanese word for sky, can create realistic-looking videos of up to a minute long. More: - Sora can create 1080p movie-like scenes from descriptions or images and "extend" video clips by filling in missing details.
- According to OpenAI's blog post, Sora can create videos with "complex scenes with multiple characters, specific types of motion, and accurate details of the subject and background."
- The company shared some Sora-generated videos, including one of a woman walking down a Tokyo street.
Release: - Sora will first launch to "a limited number of creators," according to CEO Sam Altman.
- OpenAI said that safety measures, including "adversarial" testing by experts in misinformation, hateful content, and bias, must be completed before Sora's public release.
- Sora still has weaknesses when it comes to complex scene physics, cause and effect, and spatial details, including confusion over directions and precise camera movements, the company noted.
Zoom out: - Meta, Google, and Runway AI have also developed text-to-video generators, though Sora's videos are the longest at 60 seconds.
- The worldwide market for AI video generators is expected to reach nearly $2.9B in 2032, up from around $566M last year.
|     | |
| 2 | Google DeepMind has revealed the latest version of its major large language model. Gemini 1.5 Pro is capable of processing video, images, audio, and text to answer queries. More: - 1.5 Pro boasts an expansive context window, capable of processing up to 1 million tokens, equivalent to 700,000 words, 1 hour of video, 11 hours of audio, or 30,000 lines of code.
- For example, it could read Leo Tolstoy's epic "War and Peace," spanning half a million words and over 1,000 pages, in a minute.
- Gemini 1.5 Pro operates across "many languages," including Spanish. After learning from a grammar book and a Kalamang language dictionary — spoken by fewer than 200 people in western New Guinea — 1.5 Pro successfully translated a sentence into English.
- DeepMind will first release Gemini 1.5 Pro to developers and enterprise customers through a "research release." The model will launch gradually to the public through a waitlist.
Zoom out: - Gemini 1.5 Pro was announced less than a week after Google launched a premium paid version of its AI model, Gemini Ultra.
- According to Google executives, Gemini 1.5 Pro achieves results similar to Gemini 1.0 Ultra but does so more efficiently, requiring less computational power.
| | | |
| 3 | | Apple is updating its Xcode programming software with AI that can generate code, according to Bloomberg. The Xcode AI feature will predict and finsh segments of code, simplifying the app and software development process. More: - Apple's AI tool, similar to Microsoft's GitHub Copilot, can generate code from natural language prompts and translate code across programming languages.
- The company has been working on the AI tool for a year to include in the upcoming major release of Xcode.
- According to Bloomberg, Apple is currently conducting internal tests of the feature, which could launch to third-party software developers as early as this year.
Zoom out: - Apple also plans to debut AI features like the ability to auto-generate Apple Music playlists and business presentation slide decks in Keynote, Bloomberg reports.
- Craig Federighi, Apple's software chief, has told employees to develop many new AI features for iOS 18, iPadOS 18, and macOS 15.
- Apple is expected to market iOS 18, releasing to the public in September, as one of the biggest updates to the iPhone since it launched.
| | | |
| 4 | | Nvidia disclosed its investments in AI-related companies Arm Holdings, SoundHound AI, Nano-X Imaging, Recursion Pharmaceuticals, and TuSimple Holdings. The AI chip giant disclosed its stakes in its first-ever 13F filing with the SEC. More: - Nvidia held investments in the stocks as of Dec. 31.
- As of that time, Nvidia owned $147M worth of stock in Arm, the chip design company it sought to buy for $40B but abandoned in 2022 due to regulatory hurdles.
- Nvidia also revealed its $3.67M investment in SoundHound, a vocal recognition AI software provider, after having previously invested in the company in 2017.
- Nvidia acquired a stake in Nano-X Imaging, an Israeli medical device company using AI for report analysis, while investing $3M in autonomous driving technology company TuSimple Holdings.
| | | |
| 5 | | Amazon's artificial general intelligence (AGI) team has developed what they say is the largest text-to-speech model to date. The researchers said the BASE TTS high-fidelity model can mimic human speaker traits based on only "a few seconds of reference audio." More: - The model, short for Big Adaptive Streamable TTS with Emergent abilities, showcases "state-of-the-art naturalness" in conversational text, according to the team's not-yet-peer-reviewed paper.
- The model was first trained on 100,000 hours of public domain speech data, mostly in English, to mimic American speech patterns.
- To explore the right size needed for "emergent abilities," the AGI team trained two smaller models on 1,000 and 10,000 hours of speech data, comparing their language naturalness with a larger model.
- The 10,000-hour model ranked highest in the abilities, which included things like understanding punctuation, non-English words, and emotions.
- Audio samples from the model are available here.
| | | |
| |
| 7 | | QUICK HITS - If your business accepted Visa or MC from Jan. 1, 2004, to Jan. 25, 2019, you may be entitled to a share of the historic $5.6B settlement.*
- Nvidia's market cap overtook Amazon's on Tuesday and Alphabet's on Wednesday, making it the world's fourth most valuable company after Microsoft, Apple, and Saudi Aramco.
- U.S. tech giant Cisco said it will cut 5% of its workforce, or about 4,250 jobs, as it plans to shift its focus to AI, software development, and other key areas.
- Microsoft plans to invest nearly €3.B ($3.5B) to expand its AI and cloud computing data centers in Germany over the next two years.
- Conquer any Kubernetes challenge with the Kubernetes Cookbook. Get your copy from DoiT, a global team of cloud experts.*
*This is a sponsored post. | | | |
| Upcoming Events | | MAR
7 | | Portfolio Diversification - Learn from knowledgable experts on the latest trends in alternative investment assets like private equity, crypto, real estate, and the art market. | | | | | | | MAR
27 | | Discover how generative AI can speed up your enterprise, expand your business model, and improve your value proposition. Network with hundreds of AI leaders.* | | | | | | * This is a sponsored event | | | |
 | AI and technology writer | | Beth is a contributing editor and writer of Inside's AI and Tech newsletters. She has written for notable publications like USA Today and the Arizona Business Gazette. During her time as a public policy reporter at The Arizona Republic, she received recognition with a Pulitzer Prize nomination and a First Amendment Award for her collaborative reporting on pension cost increases. You can reach her at Beth.Duckett@yahoo.com. | | This newsletter was edited by Aaron Crutchfield | |
| |
| |
Comments
Post a Comment