From Chatbots to Computer Control

Read time: under 4 minutes

The AI showdown just shifted from chatbots to computer control. Days after Anthropic taught Claude to use computers, rumors are swirling about Alphabet’s own browser-controlling AI: Project Jarvis.

Today’s Insights

• Today in AI: Humanoid Robots, Meta’s Turn, and SLIViT

• Tutorial: How to repurpose and caption videos with AI

• After Anthropic's Claude, another computer-using AI

• Everything else you should know today

• 5 new AI tools to boost your productivity

• AI-Generated Images: Scary Monday

TODAY IN AI

1. When I move, you move, just like that: EngineAI, a Chinese robotics company, introduced the SE01, a life-sized humanoid robot designed to walk in a way that closely mimics human movement, thanks to advanced neural networks and a specially designed joint system. The company plans to ramp up production to 1,000 units a year by 2025, with ambitions to bring the robot into both home and industrial settings.

2. Meta has released NotebookLlama: an open-source take on Alphabet’s NotebookLM, which uses Llama models and text-to-speech to turn text files like PDFs or blog posts into podcast-style audio. Right now, it sounds a bit robotic and less polished compared to the original version. It generates a transcript, adds some dramatization, and converts it to speech, though Meta researchers say quality could improve with better models and a two-speaker setup for a more dynamic effect.

3. UCLA researchers have created a new deep-learning tool: Called SLIViT, it can analyze 3D medical images, like MRIs and CT scans, with expert-level accuracy. What makes SLIViT stand out is its unique design, allowing it to train on smaller datasets by using insights from 2D images—meaning it can learn faster and with less data than traditional 3D models. This innovation shows great potential for clinical use and could even be expanded in the future to help predict diseases.

PRESENTED BY TAPLIO

Struggling to grow on LinkedIn? Here’s how to do it in just 10 minutes a day.

LinkedIn has over 930 million users, but only 1% of them actively create content.

That’s a massive opportunity to stand out and grow your influence.

With Taplio, it takes just 10 minutes a day.

Here’s how we make it simple:

• Our AI generates posts and carousels based on your interests in seconds.

• Schedule your posts with one click.

• Track your growth with advanced analytics—far more detailed than what LinkedIn provides.

Don’t just take our word for it:

"Taplio brought me a million views in 6 months, up from just 15k in the prior 6 months." – Emil Sterndorff, Founder of The Capital.

Join 6,200+ founders, marketers, and professionals and try Taplio today.

THE AI ACADEMY

How to repurpose and caption videos with AI

• Go to the Captions AI website and sign up.

• Upload your video, once uploaded double click on the video to open the editor.

• Go to templates and choose your favorite one from the list.

• You can customize the font and color accordingly.

• It will auto-generate captions for you. Add words if anything is missing.

• Adjust the location of captions on screen and you are good to go.

• Export the video by clicking on the export button at the top right. (you need to have a Pro account to export videos)

FROM THE FRONTIER

Alphabet joins race to create computer-controlling AI with 'Project Jarvis'

According to a report from The Information, Alphabet’s Google is reportedly working on a new AI system called Project Jarvis. It’s designed to handle everyday tasks for users by operating web browsers. Jarvis could take over Chrome to assist with things like research, online shopping, and booking flights. This system works by analyzing screenshots of the user’s screen, enabling it to perform actions like clicking buttons and filling in forms. It could be unveiled in December 2024, alongside the next Gemini 2.0 model.

Interestingly, this news comes shortly after Anthropic introduced a similar feature for its Claude AI, which can interact across different software programs, not just browsers. The timing of Jarvis’s release seems to reflect the ongoing race among major tech companies—like OpenAI, xAI, and Meta—all aiming to launch their next-generation AI models around the same time.

While Alphabet hasn’t confirmed these reports, Project Jarvis represents a major step forward in creating AI systems that can interact directly with computer interfaces to assist users.

PRESENTED BY NOTION

The New Notion AI is finally here

The New Notion AI is more knowledgeable, capable, and secure than the first iteration.

At Superhuman AI, we use Notion AI regularly to summarize our research notes, brainstorm ideas for articles, and get feedback suggestions that help us improve the quality of our writing.

Our favorite productivity features: 

• Chat: Uses knowledge from models like Claude and GPT-4

• No “prompting” required: Tons of one-click functionality

• Generate docs and edit text using any style guide

Unlock the New Notion AI for just $10/month.

AI & TECH NEWS

Everything else you need to know today

🎙️ Prompt Me: ElevenLabs released Voice Design, which allows users to generate a unique voice from a text prompt alone.

👋 Who’s next: Miles Brundage, who was OpenAI’s senior adviser for the AGI readiness team, left the company to focus on independent policy research.

🖱️ Omni Present: Eagle-eyed users noticed Microsoft silently dropped OmniParser, a tool for parsing clickable buttons and icons from screenshots.

📚️ More than Books: Goodreads’ co-founder launched Smashing, a new app that curates web content like news articles, blog posts, etc.. with an added AI and community component.

✏️ School’s In: The co-founders of Anchor (which sold to Spotify) are working on a new educational startup called Oboe, which aims to democratize access to learning.

PRODUCTIVITY

5 AI Tools to Supercharge Your Productivity

✅ ClickUp: Chat and work in one place, with AI superpowers.

✅ Agent Gold: Auto-create new content for social networks or email based on your own voice.

✅ Klue*: AI isn't your competition, but it can help you beat it. Klue AI combines market, competitor, and buyer insights in a unified platform. Take Klue for a spin today.

✅ CodeAnt: Automatically detect and fix code quality issues, bugs, and security vulnerabilities in real-time with every code commit.

✅ Moonbeam: Write essays, stories, articles, blogs, and other long-form content using AI.

^{* indicates a promoted tool, if any}

PROMPT OF THE DAY

Design Online Courses

Prompt: Design an engaging online course with personalized learning experiences. Customize content, pace, and assessments to optimize student outcomes. Leverage interactive multimedia and real-time feedback for maximum engagement. Blend self-paced and instructor-led activities to support diverse learning styles.

_{Source: promptpal}

AI-GENERATED IMAGES

Spooky Monday

Midjourney Prompt: Vintage photo of an old, large demon with horns and black skin standing next to his family in front of their house. The year is 2036, with a sepia filter applied.
--ar 9:16 --v 6.1

Acquire new customers and drive revenue by partnering with us

Superhuman is the world’s biggest AI newsletter for businesses and professionals with 800,000+ readers and 1.5 Million followers on socials working at the world’s leading startups and enterprises. Companies like Amazon, Hubspot, and Salesforce feature their products in Superhuman. You can learn more about partnering with us here.

🧞Your wish is my command

Got more feedback or just want to get in touch? Reply to this email and we’ll get back to you.

Thanks for reading.

Until next time!

Zain & the Superhuman AI team