• AI Breakfast
  • Posts
  • This MIT Chatbot Simulates Your 'Future Self'

This MIT Chatbot Simulates Your 'Future Self'

Good morning. It’s Friday, June 7th.

Today in tech history: On this day in 2000, a federal judge ordered Microsoft to break up into 2 separate companies. This never happened - and a year later, the government announced that it would seek a lesser antitrust penalty.

In today’s email:

  • Apple’s ChatGPT Integration

  • CRISPR 2.0

  • Chinese AI Video Generator

  • AI Sound Design

  • MIT’s “Future Self” AI

  • 10 New AI Tools

  • Latest AI Research Papers

You read. We listen. Let us know what you think by replying to this email.

Today’s trending AI news stories

Report: Apple To Integrate ChatGPT into iOS 18 As Opt-in Feature

Apple Inc. is expected to integrate OpenAI’s ChatGPT into iOS 18 as an opt-in feature, per Bloomberg. The next major iOS release is set to debut at the WWDC developer event on Monday. Alongside ChatGPT, the update promises a host of AI enhancements in the search bar, emoji library, and built-in apps.

Before choosing ChatGPT, Apple reportedly flirted with Google LLC’s Gemini chatbot but ultimately chose OpenAI for its more favorable terms and superior language models. The exact OpenAI model for iOS 18 remains a mystery, though it might feature the next-gen LLM if ready by September.

In addition to ChatGPT, iOS 18 is expected to offer new machine learning features, including the ability to create emojis and slideshows via natural language, access AI-generated Apple Music playlists, and use the search bar for app tasks. These features are anticipated to be available on an opt-in basis.

Parallel to the iOS 18 rollout, Apple is reportedly working on two LLM-powered home robots. One is a table-top robotic arm with a smart display, and the other is an autonomous vehicle designed to perform chores. Both projects are in the early stages, reflecting Apple's expanding AI ambitions. Read more.

CRISPR 2.0: Kinder, Gentler, More Powerful Gene Editing

The CRISPR-Cas gene editing saga unfolds with CRISPR 2.0, a kinder, gentler iteration poised to revolutionize genetic tinkering. Unlike its clunkier predecessor, CRISPR 2.0 wields surgical precision, sparing DNA the scars of double-strand breaks and opening doors to a broader array of gene therapy applications. Recent developments include base editors and prime editors, offering more targeted approaches to gene editing by nicking one DNA strand instead of causing breaks.

The success of CRISPR-based therapies like Casgevy for sickle cell disease highlights the immense potential of these systems. CRISPR Therapeutics is pioneering the development of allogeneic CAR T-cell therapies and in vivo CRISPR-Cas9 treatments for various diseases, with promising results in clinical trials. Additionally, CRISPR QC and Metagenomi are revolutionizing CRISPR quality control and expanding the CRISPR toolbox, respectively, to optimize gene editing efficiency and safety. Read more.

KLING: The Chinese AI Video Generator That Could Rival OpenAI's Sora

Kuaishou, a Chinese tech giant, introduced KLING, a new AI model for video generation. KLING sets its sights on becoming a rival to OpenAI's Sora, boasting the ability to create videos up to two minutes long in 1080p resolution at 30 frames per second. KLING can reportedly model complex motion sequences with physical accuracy thanks to a 3D space-time attention system, potentially surpassing OpenAI's "world simulator" approach.

KLING goes beyond just physically accurate motions. It employs a "diffusion transformer" to combine concepts and generate entirely fictional scenes, like a cat driving a car, that weren't part of its training data. Currently available as a public demo in China, KLING positions Kuaishou as a strong contender in the race for large-scale generative AI models. Read more.

Watch KLING's AI-generated video here:

Stable Audio: Open-Source AI Sound Design

Stability AI expands its generative audio repertoire with Stable Audio Open, a freely available, open-source model catering to short-form audio needs. Unlike its commercially-driven sibling, Stable Audio 2 which specializes in full-length songs, Stable Audio Open adopts a more granular approach, specializing in audio snippets capped at 47 seconds.

Designed for sound designers, musicians, and creative professionals, the model is trained to produce drum beats, instrument riffs, ambient sounds, foley recordings, and more for music production and sound design purposes. Utilizing audio data from FreeSound and the Free Music Archive during training, Stability AI emphasizes responsible development and protection of creators' rights.

Available for download on Hugging Face, Stable Audio Open encourages user customization and feedback. Read more.

This MIT Chatbot Simulates Your 'Future Self.' It's Here to Help You Make Better Decisions

MIT Media Lab takes a futuristic spin on self-help with the "Future You" project. This initiative leverages AI-powered simulations, including personalized chatbots and AI-rendered avatars, to introduce users to their potential selves at age 60, aiming to foster long-term thinking and behavior change. Participants engage with AI-powered simulations, which include personalized chatbots and AI-generated images, to explore their future experiences, memories, and insights.

Unlike traditional therapy-focused chatbots, this project targets future-self continuity, a psychological construct linked to reduced anxiety and improved decision-making. Through web-based accessibility and personalized interactions, participants report heightened emotional connections and decreased anxiety post-conversation. While promising, the study's scalability and long-term impact remain areas for further exploration. Read more.

Etcetera: Stories you may have missed

10 new AI-powered tools from around the web

Fliki utilizes advanced AI to generate videos, audios, and designs, featuring AI avatars, natural voices, and customizable templates, optimizing content creation, e-learning, and marketing workflows.

Formshare enables the creation of interactive, conversational AI forms without coding, allowing users to generate forms using natural language descriptions. Perfect for productivity and marketing.

OrgaNice is an AI-powered HR platform designed for Slack-first teams, facilitating organization, recognition, synchronization, and engagement directly within Slack.

Taloflow 2.0 is an AI-driven platform that accelerates tech vendor evaluations by replacing lengthy processes with collaborative notebooks and expert insights.

Maple CMS is a headless CMS with AI capabilities, enabling effortless content creation, management, and seamless code integration for any tech stack.

Databutton is an AI-powered tool that enables users to create SaaS applications by converting natural language instructions into React frontends and Python backends.

BeeDone is an AI-powered productivity tool that uses gamification to help users maximize focus, prioritize tasks, build habits, and maintain routines.

Falcon AI is an AI-powered Scrum Master integrated with Slack, enhancing agile project management by automating scrum ceremonies, JIRA boards, and stakeholder updates.

Riffo is an AI-powered tool for renaming and organizing files, offering smart renaming, batch processing, and efficient workspace management.

Doly by AniML is an iPhone app for creating 3D product videos, featuring intuitive capture grids, multi-angle views, and easy export for marketing use.

arXiv is a free online library where researchers share pre-publication papers.

Thank you for reading today’s edition.

Your feedback is valuable. Respond to this email and tell us how you think we could add more value to this newsletter.

Interested in reaching smart readers like you? To become an AI Breakfast sponsor, apply here.