An AI assisted calendar is this week's top pick

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

Hi all! This week we surpassed 800 subscribers 📈👏🏻

A big thanks to everyone that shared this newsletter with their friends. My goal was to hit 1,000 subscribers by the end of May, but I think I'm going to have to set the goal higher 😀😀

In this week’s issue:

• Mayday, an AI assisted calendar, is my product of the week

• Learn how to write better image prompts

• A copilot for data analysts

• And more AI website builders

• Plus, a recommended read on navigating the conflicting emotions of this moment in tech

PRODUCT OF THE WEEK

After testing dozens of new AI products this week. Here’s my top pick.

Mayday: An AI assisted calendar that aims to move the calendar beyond just being a dumb ledger of all the things in your day. I’ve tested out a variety of calendar tools and also pay for one (I ❤️ you Vimcal)…and I’m impressed with what the team at Mayday has built. They’ve taken all of the essentials of calendar management (eg. external time slots) and layered in AI so that it almost feels like you have a personal assistant. In the onboarding experience you tell Mayday what you are trying to achieve (”find more time for deep work”, “find more time for myself”) and the AI will help you organize your calendar to achieve that. For instance you can set a daily meeting threshold and the calendar will begin to protect your time if meetings start to stack up on a particular day. Their co-founder, Adam Day, told me that for AI they are starting with intelligent suggestions and natural language processing on events, but they have lots of AI features on their upcoming roadmap. Here’s a 3 minute demo from their co-founder and CEO, Jeremy Bell. Available on Mac and iPhone. [Public Beta | Free for now]

RUNNER UPS

Four more AI products that are worth mentioning.

UnPrompt: Learn how to write better image generation prompts. For those of you that are interested in learning how to get the most out of an AI image generator, UnPrompt enables you to input an image, such as the pope in a Balenciaga puffer jacket, and it will tell you the prompt that would have created that image. Here’s their 20 second demo video. [Public Launch | Free]

ProbeAI: A copilot for data analysts. This copilot can help data analysts with those repetitive tasks, like writing (or fixing) their SQL code and identify the relevant tables to query. This would have saved me countless hours of SQL writing at Reddit if I could have just been able to ask it “what’s the median number of days it takes for a new community to reach 50 comments per day”… and bam I have my query ready to go. I’m excited to see how this product evolves. Maybe they’ll start with a new name? [Private Beta]

EnhanceAI: Help your customers fill out forms on your website using AI. If you have any sort of customer inputs on your website, such as lead gen forms or reviews, EnhanceAI looks to be a super useful tool. It takes less than a minute to add the proper code to your website builder (Webflow, Wix, etc) and then when a customer fills out your lead gen form, they’ll get AI autocomplete prompts directly in the field making it easier for them to complete it. What’s really neat is that you can add a bunch of pre-prompt context so that the autocomplete is more relevant to your website. [Public Launch | Free or Paid]

Butternut AI: Build a fully-functional website landing page in 20 seconds. Butternut AI launched this week and boasted that their tool enables you to build a fully functional website in 20 seconds, but I didn’t find that to be the case. Below is a 3 minute video of me trying to create a jobs board (a common website template on many no-code builders). Butternut AI created a landing page that said the website was a jobs board with search and filter functionality, but there wasn’t any actual search or filter functionality. Butternut asked me to input my email to go to the next step, but they never emailed me next steps 😞 It’s got potential, but needs some work. [Beta | Free or $20 per month]

OTHER AI THINGS HAPPENED

Some other notable news from this week

Segment Anything: Meta released a model called Segment Anything that can identify anything in a photo, even something it hasn’t been trained on before. The AI model can quickly learn and do new tasks without needing more training, which is really impressive.

BloomberGPT: Bloomberg created a large language model for financial natural language processing (NLP) tasks. The model has 50 billion parameters and leverages Bloomberg's vast archive of financial data and documents. Another example of how companies with large troves of proprietary data are going to benefit from the AI boom.

Imagica AI: The startup, Brain AI, released a sneak peek of their no-code website builder that let’s you build a high functioning website with text prompts. It appears to be extremely powerful. I signed up for the beta, will report back when I actually get my hands on it.

Google Search's Chatbot: In a Wall Street Journal interview published this morning, Google's CEO Sundar Pichai said that Google Search will soon feature an LLM driven chatbot. TBD on when this thing will actually ship. 

WHAT I'M READING

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

Awe, Anxiety, and AI by Dan Shipper

"This weekend, I sat down to watch AlphaGo, a documentary about the eponymous Go-playing AI that beat the reigning Go champion Lee Sedol in 2016. Sedol is an easy hero: He’s skilled, intelligent, hardworking, and gracious. He arrived at the match with AlphaGo projecting public confidence that he could beat it handily having studied public demonstrations of its powers. Go is a 3,000-year-old game that is considered one of the most complicated board games on the planet—orders of magnitude more complicated than chess. Chess has 10123 board positions, while Go has 10360 board positions. There are more board positions in Go than there are atoms in the universe. Go is so complicated that Go masters often can’t explain why they make certain moves. In 2016, the game was considered far out of reach for even the best machine learning models, and so Sedol’s confidence was warranted. He was wrong. From the first move, he got completely smoked by an unfeeling, relentless machine that had no idea who Sedol was or the stakes in the game it was playing. In one particularly poignant moment..." CHAIN OF THOUGHT

Until next week! ✌🏻

-Tyler

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

P.P.S. Interested in having me give you private feedback about a product that you are building? Learn more here.