π Grok-1.5 Launch, AI Identity Theft Alert, and AI21's Leap: Long-Context AI Breakthrough
Todayβs pick
Announcing Grok-1.5 Introducing Grok-1.5, our latest model capable of long context understanding and advanced reasoning. Grok-1.5 will be available to our early testers and existing Grok users on the π platform in the coming days. By releasing the model weights and network architecture of Grok-1 two weeks ago, we presented a glimpse into the progress xAI had made up until last November. Since then, we have improved reasoning and problem-solving capabilities in our latest model, Grok-1.5. x.ai
π: Should be available on π next week. Grok 2 should exceed current AI on all metrics. In training now. - Elon Musk (@elonmusk)
AI hustlers stole women's faces to put in ads. The law can't help them. Artificial intelligence is spurring a new type of identity theft β with ordinary people finding their faces and words twisted to push often offensive products and ideas. Michel Janse was on her honeymoon when she found out she had been cloned. The 27-year-old content creator was with her husband in a rented cabin in snowy Maine when messages from her followers began trickling in, warning that a YouTube commercial was using her likeness to promote erectile dysfunction supplements. By Nitasha Tiku via Washington Post
π: βMichel Janse was on her honeymoon when she found out she had been clonedβ¦β Ordinary women online are now being victimized by AI deepfakes for everything from Putin propaganda to ED ads - Will Oremus (@WillOremus)
The best way to reach new readers is through word of mouth. If you click THIS LINK in your inbox, itβll create an easy-to-send pre-written email you can just fire off to some friends.
AI21 Labs' new AI model can handle more context than most. Increasingly, the AI industry is moving toward generative AI models with longer contexts. But models with large context windows tend to be compute-intensive. Or Dagan, product lead at AI startup AI21 Labs, asserts that this doesnβt have to be the case β and his company is releasing a generative model to prove it. Contexts, or context windows, refer to input data (e.g. text) that a model considers before generating output (more text). By Kyle Wiggers via TechCrunch
π: Wow this is a big deal, first large-scale Mamba-based model! Mamba layers brings much longer context and higher inference throughput. Having 4 attention layers seem to be the sweet spot to get the best of Transformer & Mamba architectures. - Tri Dao (@tri_dao)
NYC will test AI gun detectors on the subway. New York City will soon start testing out technology that uses AI to detect guns at subway turnstiles, Mayor Eric Adams said on Thursday. Adams' announcement comes one week after an altercation at a subway station in Brooklyn in which a man was shot with his own gun after pulling it on another passenger. By Gaby Del Valle via The Verge
π: These metal detectors are a non starter and do nothing to fix the issues. - Hayden Clarkin (@the_transit_guy)
Social Media, Authoritarianism, and the World As It Is. Earlier this month, the United States House of Representatives passed the Protecting Americans from Foreign Adversary Controlled Applications Act, a bill that would force TikTokβs parent company to sell the platform to U.S. citizens or be banished from the U.S. market. The same fate could await any other platform the President designates as a βforeign adversary-controlled application. By Meredith Whittaker via lpeproject.org
π: On the TikTok ban, the danger of abandoning defense of speech to extremists, & how the liberal tendency to assume a just state "outside the scope of this paper" leads to confused law/policy that can exacerbate the problems platforms pose. - Meredith Whittaker (@mer__edith)
Thanks for reading to the bottom and soaking in our Newslit Daily fueled with highlights for your morning.
I hope you found it interesting and, needless to say, if you have any questions or feedback let me know by hitting reply.
Take care and see you on Monday!
P.S. Want to advertise with us? Weβd love to hear from you.