I made an AI research tool
It's called Bearly AI.
Thanks for subscribing to SatPost.
Today, I’m excited to announce the launch of an AI research tool for reading and writing: Bearly AI.
Meet Bearly AI
As many of you know, I consume way too much internet content.
And in the past few months, I’ve seen dozens of awesome AI writing and art tools that looked like they could help my daily workflow (all were built on top of open-source models from OpenAI, the $20B AI research lab).
I didn’t try any of the tools, though.
There are just too many options and I’m really bad at juggling lots of open tabs (this is the same reason I don’t use productivity apps and default to iPhone Notes).
But I wanted to use the power of these large AI language models for:
Automating writing ideas based on existing text
Generating first drafts on any topic (email, blog posts, tweets etc.)
Instant summaries (including key takeaways and counter-arguments) for any article or YouTube video
Why? Because I know AI can improve my:
Reading: Retain 10x the knowledge from what I read
Writing: Speed up my writing process 10x by suggesting ideas and autocompleting text
More efficient reading and writing would easily save me 10s of hours a month.
These tools exist but, again, I’m not trying to deal with 30 browser tabs.
What’s the solution? I wanted to create a single app that sits on your desktop and appears whenever you need an AI boost. So, I hit up my former colleague Parham Negahdar, who was head of infrastructure at Kensho (an AI fintech firm acquired by S&P Global for $550m). Turns out he’d been working on the exact AI solution I needed.
So, we partnered up and are excited to announce:
Bearly AI, an app that sits on your desktop and gives you the power of OpenAI from a single click or keyboard shortcut.
After installing the app, you can summon Bearly AI anytime with CMD+SHIFT+P and a dropdown menu appears with OpenAI-powered text tools. Pick an action and then paste in any body of text (that you need work on) or enter a short prompt (to create brand new text).
Need a first draft of an email written? Hit CMD+SHIFT+P.
Need grammar correction on a shared Google Doc? Hit CMD+SHIFT+P.
Need 3 ideas for the next paragraph of your newsletter? Hit CMD+SHIFT+P.
Need 5 punchy headlines for your article (or tweet)? Hit CMD+SHIFT+P.
Based on a handful of people that have demo’d the product, the most popular prompt is for suggesting what to write next (as a way to break through writer’s block).
In this next GIF, I input a few lines into the Bearly app and it gave me 3 paragraph recommendations. Each is written in a different tone and each takes the article in a different direction. Whether or not you use the full paragraphs, it’s an incredible ideation tool.
That’s the writing stuff.
We have a wild tool for reading: instant article summaries. Over the past week, I’ve been using Bearly to summarize every article I read. It helps prime me before I read and is a great memory jogger after I read.
Even better: the Bearly AI reader is just an icon that pops up on every webpage. You click it and ask for “exec summary” and instantly get:
3 key takeaways
If you have to summarize stuff for work, it’s a huge time saver. Or how about all these articles that people send you throughout the day? No one can finish that content canon, but Bearly gives you a top-level overview (at a minimum).
And you can do it for YouTube videos…seriously (one click in the bottom right):
We’ll be adding new OpenAI-powered text prompts every single day and the tools we can build on top are only limited by our imaginations.
Generative text is first. Art is next (coming very soon).
The best part: after you install Bearly AI, every single tool we add will be available by just hitting CMD+SHIFT+P.
SatPost subscribers get priority access.
So, download the app and signup below. The first 1000 SatPost readers will get their invite and accounts activated on Monday morning.
PS. Check out the demo video.
Links and Memes
Elon is now “Chief Twit” of Twitter and entered Twitter HQ holding a porcelain sink as a nod to a famous meme. (Link) He also sent an open-letter to advertisers saying that Twitter “obviously cannot become a free-for-all hellscape” and his view on ads: “Advertising, when done right, can delight, entertain, and inform you…low relevancy ads are spam but highly relevant ads are content” (Link) — ads make up ~90% of Twitter revenue, so that makes sense and one thing that could help: Elon sent Tesla engineers to Twitter HQ to review the code.
What’s next for Twitter? Elon says “comedy is now legal on Twitter” and that major Twitter content moderation decisions will be a decided by a council (Facebook/Meta has a similar council that: 1) takes pressure off of Zuck; 2) but still pisses people off no matter what because so many hot button issues are 50/50).
One way around the moderation issue is to let people choose their own adventure when it comes to algorithms — Elon replied “something like this makes sense” to the following tweet:
Meta is down 71% YTD but Zuck won’t stop investing in the Metaverse…spending $10B+ on the unit this year and expects that sum to “grow significantly” in 2023. (Link)
The rationale behind Zuck’s big bet: Here is a tweet length take on Zuck’s bet by Antonio Garcia Martniez, who helped to build Facebook’s ad business: “Meta's business has been kneecapped by Apple, growth has stopped in favor of rivals like TikTok, and it's too late to build a phone. Zuck is making an existential bet-the-company move at enormous scale. Most companies would just delude themselves and slowly die.” (Link)
Succession: The trailer for Season 4 of HBO’s Succession just came out (the season will drop Spring 2023). Here’s my favorite YouTube vid on the show: it’s by NerdWriter and explores how all the characters speak in a coded language.
Some gold tweets
The UK is on its third Prime Minister in less than two months. And that person — 42-year old Rishi Sunak — is the first British PM of Indian descent…hence this tweet:
On semi-related news, Indian CEOs lead a number of America’s top tech firms including Alphabet, Microsoft, Adobe, IBM etc…hence this tweet:
And a former Indian tech CEO — Parag Agrawal — just rode off into the sunset:
Finally, the Twitter main characters this week…are two people that pretended to be “fired Twitter employees”. These
method actors pranksters stood outside Twitter HQ on Friday with cardboard boxes and conducted very fake interviews about their “firings”.
Their “names” were Daniel Johnson and Rahul Ligma. For those that aren’t familiar, the word “Ligma” is usually a set-up for a joke that ends up with the punchline line “Ligma nuts” (as in lick my nuts). And “Johnson” can sometimes refer to the male member. And if you add them up then you get “Ligma Johnson”.
A few media outlets initially wrote about Ligma and Johnson as if their story was true…but enterprising internet folk soon pointed out the “Ligma” factor and it did turn out to be a prank.
The story peaked when Elon tweeted out a DailyMail article that was fooled by Ligma and Johnson.