This week the TechCrunch crew (including yours truly) is at TC's annual Disrupt conference in San Francisco. We have a big selection of speakers from the AI ​​industry, academia and policy, so as an alternative of my usual commentary I believed I'd take a have a look at a number of the great content you may expect.
My colleague Devin Coldewey might be interviewing Perplexity CEO Aravind Srinivas on stage. The AI-powered search engine is on the rise, recently hitting 100 million searches per week – but it is usually being sued by News Corp's Dow Jones over what the publisher calls “content kleptocracy.”
Kirsten Korosec, transportation editor at TC, will chat over a hearth with Zoox co-founder and CTO Jesse Levinson. Levinson, who has been within the thick of autonomous vehicle technology for a decade, is now preparing the Amazon-owned robotaxi company for its next big adventure – and we'll be covering it.
We will even host a panel on how AI is flooding the web with disinformation with Meta Oversight Board Member Pamela San Martin, Center for Countering Digital Hate CEO Imran Ahmed, and UC Berkeley CITRIS Policy Lab Founder Brandie Nonnecke . The trio will discuss how generative AI tools, as they change into more widespread, are being misused by a spread of actors, including state actors, to create deepfakes and spread disinformation.
And we'll be joined by Jingna Zhang, CEO of Cara, Sarah Myers West, co-executive director of the AI ​​Now Institute, and Aleksandra Pedraszewska of ElevenLabs to speak concerning the legal and ethical minefields of AI. The meteoric rise of AI has created latest ethical dilemmas and exacerbated old ones, while complaints fall left and right. This threatens each latest and established AI corporations, in addition to the creators and employees whose work powers the models. The panel will address all of this – and more.
This is only a small excerpt of what's on the agenda this week. Expect appearances from AI experts comparable to Elizabeth Kelly, Director of the US AI Safety Institute, Scott Wiener, Senator of California, Jessica Newman, Co-Director of the Berkeley AI Policy Hub, Amit Jain, CEO of Luma AI, Mikey Shulman, CEO of Suno, and Kakul Srivastava, CEO of Splice.
News
Apple Intelligence launches: Through a free software update, iPhone, iPad and Mac users can access Apple's first AI-powered Apple Intelligence features.
Bret Taylor's startup raises latest money: Sierra, the AI ​​startup co-founded by OpenAI chairman Bret Taylor, has raised $175 million in a funding round that values ​​the startup at $4.5 billion.
Google expands AI overviews: Google Search AI Overviews, which display a snapshot of knowledge at the highest of the outcomes page, are rolling out to greater than 100 countries and territories.
Generative AI and e-waste: The immense and rapidly advancing computational demands of AI models could end in the industry disposing of the e-waste equivalent of greater than 10 billion iPhones per yr by 2030, researchers predict.
Open source, now defined: The Open Source Initiative, a long-standing institution dedicated to defining and “governing” every little thing related to open source, released version 1.0 of its definition of open source AI this week.
Meta releases its own podcast generator: Meta has released an “open” implementation of the viral podcast generation feature in Google’s NotebookLM.
Hallucinated transcriptions: Researchers say OpenAI's Whisper transcription tool has hallucination problems. Whisper reportedly included every little thing from racist comments to imagined treatments within the transcripts.
Research paper of the week
Google says It taught a model to convert photos of handwriting into “digital ink.”
The InkSight model was trained to acknowledge written words on a page and output strokes that roughly resemble handwriting. The goal, say the Google researchers behind the project, was to “capture the stroke-level trajectory details of handwriting” in order that a user can save the resulting strokes within the note-taking app of their alternative.
InkSight isn't perfect. Google realizes it’s making mistakes. However, the corporate also claims that the model performs well in a spread of scenarios, including difficult lighting conditions.
Let's hope this doesn't end in forged signatures.
Model of the week
Cohere for AI, the nonprofit research lab of AI startup Cohere, has released a brand new family of text-generating models called Aya Expanse. The models can write and understand text in 23 different languages, and Cohere claims that they outperform models including Meta's Llama 3.1 70B on certain benchmarks.
According to Cohere, a way called “data arbitrage” was key to training Aya Expanse. Cohere was inspired by the way in which people learn by turning to different teachers for unique skills and chosen particularly powerful multilingual “teacher” models to generate synthetic training data for Aya Expanse.
Synthetic data has its problems. Some studies suggest that over-reliance on it might result in models that change into increasingly poor in quality and variety. But Cohere says data arbitrage effectively mitigates this. We'll soon see whether the claim stands as much as scrutiny.
Lucky bag
OpenAI's Advanced Voice Mode, the corporate's realistic-sounding voice feature for ChatGPT, is now available without cost on the ChatGPT mobile app for users within the EU, Switzerland, Iceland and Norway, and Liechtenstein. Previously, users in these regions needed to subscribe to ChatGPT Plus to make use of enhanced voice mode.
A youngest An article in The New York Times highlighted the professionals and cons of Advanced Speech Mode, comparable to its reliance on tropes and stereotypes when trying to speak the way in which users want. Advanced Voice Mode has blown up on TikTok for its uncanny ability to mimic voices and accents. But some experts warn that this could lead on to emotional dependence on a system that lacks intelligence and empathy.
