Snap's soulless AI companion and a mind-blowing Kendrick Lamar voice filter

A weekly newsletter that highlights new and innovative AI products that are worth exploring.

Happy Friday all! Both of my kiddos are now attending daycare, so I’m writing this update from the bliss of a calm and quiet house…although I still hear phantom cries periodically. In this week’s edition:

• I get my hands on Snap’s new AI chatbot

• A mind-blowing voice filter demo

• A nifty work document organizer

• A chrome plugin that actually does something unique

• Lastly, a thought provoking article about Tesla, their AI “autopilot,” and the philosophical and ethical choices Elon Musk is faced with

NEW PRODUCTS

Here are four products from this week that are worth your time.

Snap’s “My AI”: A soulless AI companion right next to your chats with friends. On Monday, Snap announced they’ve entered the AI space with a chatbot. It’s smart that they reduced the friction to use it by putting the chatbot right next to the chats with your friends. It’s almost as if “My AI” is just another one of your friends. Snap is all about entertainment and conversation, so it’s a natural fit for an AI chatbot, much more than say Bing Search. Snap says they are using OpenAI’s GPT and have added some additional guardrails, but I wouldn’t be surprised to see teens quickly figure out hacks to break it out of it’s guardrails. At the moment, it seems very limited on what it can talk about. To me, it felt soulless. It kept telling me it couldn’t do something and then ending it’s response with, “is there anything specific that you need help with today?” [Public Launch - $3.99 per month]

A real-time Kendrick Lamar voice filter: A mind-blowing demonstration of AI being used to copy someone else’s voice (and identity) 🤯 In the video below, a Tiktok musician demonstrates how they were able to leverage a voice filter to make their voice sound like Kendrick Lamar’s…in real-time…without missing a beat. It’s crazy. As ML models get faster and more powerful we are going to see more and more unique ways that they are applied. This is a cool demonstration, but it raises concerns about the impact to creatives, similar to how generative art is being trained to mimic specific artist’s unique styles. [Preview]

Sense: All of your team's documents and information automatically organized and interconnected. On Tuesday, the team at Sense launched a 2.0 version of their AI organization product. You connect it to all of your various sources at work (Google Docs, Figma, Slack, Github, Jira, etc) and it begins to build out connections between them centered around specific topics or features. IMO the most useful feature is that you can create a new slide deck about a feature (say a new landing page) and Sense quickly tracks down all of the relevant docs, such as the PRD over in Google docs and the design file in Figma, putting all the resources you need to make your slides at your fingertips. Allegedly it can also create tasks for you based on your doc or Slack mentions, but I’m skeptical of that being useful. Most likely it would become a big un-prioritized laundry list of tasks and cause me to throw up. Anyways, I think Sense has some useful bits and it’s worth following. [Public Launch - $9 per user per month]

Magical AI: A chrome plugin AI assistant that automates some of the basics. Seems like everyone is launching an email writing assistant that leverages ChatGPT. Yes, Magical AI can do that, but that’s not what caught my eye. What I found to be unique about Magical, is that their AI can help you quickly fill in forms (think Salesforce CRM) or a Google sheet. It does this by analyzing the content of the page you are on AND what’s open in your other tabs. For instance, let’s say you are creating a list of competitors in a Google sheet and you’ve found a new competitor that you want to add. You hop over to your sheet, type “//” in the new row to pull up Magical AI and it will pre-populate the fields in that row with info pulled from the competitor’s website that you have open in the other tab. Pretty f*cking nifty. [Public Launch - Free]

WHAT I'M READING

If you only read one thing this week let it be this.

Elon Musk’s Appetite for Destruction by Christopher Cox (New York Times)

"Early on, the software had the regrettable habit of hitting police cruisers. No one knew why, though Tesla’s engineers had some good guesses: Stationary objects and flashing lights seemed to trick the A.I. The car would be driving along normally, the computer well in control, and suddenly it would veer to the right or left and — smash — at least 10 times in just over three years. For a company that depended on an unbounded sense of optimism among investors to maintain its high stock price — Tesla was at one point worth more than Toyota, Honda, Volkswagen, Mercedes, BMW, Ford and General Motors combined — these crashes might seem like a problem. But to Elon Musk, Tesla’s chief executive, they presented an opportunity. Each collision generated data, and with enough data, the company could speed the development of the world’s first truly self-driving car..." CONTINUE READING

WHO I'M FOLLOWING

People passionate about AI that are doing cool shit.

Rob Lennon is an experienced digital marketer that has become an expert at Prompt Engineering. He regularly shares nuggets on how to get the best out of a chat based AI. Checkout his thread below for some of his best techniques.

CLOSING THOUGHT

Because we shouldn't take ourselves too seriously.

Until next week! ✌🏻

-Tyler

P.S. Have tips or suggestions for next week's issue? Reply to this email and send them my way.