When AI goes physical

Everything you need to know in AI this week

October 06, 2023

Hey there! The past month has been crazy in terms of new AI announcements, with each week bringing about more huge news than the last. This week, we’ll be taking a small breather from the incumbent tech giants’ new AI products—but don’t worry, there’s still lots of exciting AI news to cover. Let’s dive right in! :)

—Matt

Let’s Get Physical

Humane on the runway. Photo credit: Luca Tombolini for Coperni.

The AI fever is spilling over to the physical realm. We’ve seen wearable devices that use AI to perform a task since 2013 (shoutout NFC ring), but devices designed specifically for AI applications are new to the scene. AI-specific wearables fall into a new product category—one that’s been seeing major developments.

This week’s hottest launches:

  1. Rewind Pendant. Rewind.AI introduced a $59 neck-worn pendant that records and transcribes everything you say and hear. Its AI software basically creates a searchable audio database from the soundtrack of your everyday life. (Which sounds like it’s straight out of a Black Mirror episode…)

  2. Tab. Founded by 20-year-old tech visionary Avi Schiffman, Tab is a wearable AI device you can ask questions and chat with. Tab is different from your average chatbot—it listens in on users’ daily conversations to integrate context specific to the individual user in the answers it provides.

Other AI wearables on the horizon:

  • Humane (a tech company led by former Apple employees) recently debuted its new AI pin—the first-ever wearable designed specifically for AI—on Paris Fashion Week’s catwalks. Despite the hype, there are still lots of open questions about what the screenless little square pin can actually do.

  • Meta announced a partnership with Ray Ban last week that integrates the new Meta AI into a pair of smart glasses.

  • Jony Ive (aka the iconic designer of the iPhone) is allegedly in talks with OpenAI and SoftBank to develop a mysterious AI hardware product.

  • AI startup Prophetic is building a prototype for a headset that can induce lucid dreams.

Big picture: Tech wearables are notoriously difficult products to get right. Sure, the Apple Watch has seen some success in recent years, but other attempts (*cough* Google Glasses *cough*) have largely failed to make it to the masses. Which begs the question: Will this new category of AI-specific wearables be any different?

Today’s Future Tools brought to you by:
FastBots

AI Powered Chatbots Trained On Your Data

  • ✅ Create a fully automated AI chatbot like ChatGPT to help you or your customers get fast answers about your business with zero coding.

  • 🎓 Easily train the Chatbot using your website link, PDFs, Word Docs, CSVs, XLS files, or even YouTube videos.

  • 💬 See all the message conversation history and export into a PDF.

  • 🎤 Tune the Chatbot to communicate the way you want to match your brand voice.

  • 📋 Have your Chatbot automatically collect leads and email them to you or your team.

For a limited time, we're offering you the chance to claim a FREE AI-powered chatbot.

🗞️ Other AI News & Articles

Stability AI’s New LLM

After the successful releases of Stable Diffusion XL and Stable Audio, Stability AI is keeping up its remarkable pace with yet another promising AI release.

Enter: Stable LM 3B.

The details: Stable LM 3B is a new LLM that’s designed to operate on mobile devices and laptops (instead of huge data centers). Because it only uses three billion parameters, Stable LM 3B is more compact and efficient than larger LLMs—which typically use 7 to 70 billion parameters.

Why’s that exciting? Stable LM 3B hints at a future where AI models are small (but still powerful enough!) to run on personal devices.

The Problem With AI Watermarking

Generative AI creates a dangerous breeding ground for misinformation and misleading deepfakes—which is why the ability to watermark AI-generated content is so important. But as it turns out, current methods of AI watermarking are…pretty terrible.

Why’s that? Researchers at the University of Maryland found that it’s pretty easy for bad actors to evade and remove AI watermarks. They found it even easier to add fake watermarks to human-made pictures and videos, which blurs the lines between reality and deepfakes further.

The findings are humbling; right now, watermarking is one of the most promising strategies to identify AI content. While researchers are seeking more sophisticated authentication tools, they’re reframing expectations around watermarks—viewing them not as the end-all-be-all, but as a form of harm reduction for lower-level AI counterfeits.

Everything else:

  • Anthropic is looking to raise an additional $2 billion from Google and other investors…just days after securing a $1.25 billion check from Amazon.

  • Microsoft entered the competitive field of LLM application frameworks with its open source Python library, AutoGen.

  • Visa pledged $100 million for its new generative AI venture fund.

  • Meta’s new AI-generated stickers are already causing trouble

  • The new Arc Max has some nifty AI features—in what many call the most impressive implementation of an AI browser to date.

📺️ This Week’s Videos to Watch:

Interested in those Meta Ray Bans? I actually covered them (and all the other major AI announcements from Meta) in a recent video:

Food for thought: When it comes to jobs, the rapid advances of AI are often presented in a negative light. Here’s the more optimistic argument for a 3.5-day work week.

And that’s a wrap! Will you be picking up one of those fancy new AI wearables any time soon? Or is this Google Glasses all over again? Shoot me a reply with your thoughts!

—Matt (FutureTools.io)

Interested in working together on a newsletter sponsorship, YouTube integration, or featured listing on the Future Tools website? I look out for the audience above all else, so I'm pretty picky—but get in touch here and let's see if we can make something happen.

If you haven’t already, be sure to follow me over on X. I like to share news and AI updates as they happen over there.

And don't forget to check out all of the newest tools we've just added on Future Tools!

You rock! See ya next week. :)

P.S. This newsletter is 100% written by a human. Okay, maybe 96%.