ChatGPT-4 is finally unveiled, Microsoft & Google go head to head

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

Before we jump in, I want to pour one out for Reddit. Today is my last day working there. It’s been a fun and informative 5 years and I know they’ll miss me…well at least, I’ll miss them. Monday will be Day 1 of my entrepreneurial adventure. Wish me luck or hit me with some venmo.

In this week’s edition:

• ChatGPT-4 is finally unveiled

• Rival language model, Claude, opens their waitlist

• Duolingo & Khan Academy preview their GPT-4 tutors

• Google and Microsoft continue their PR upstaging attempts

• Image generation takes another big step forward with Midjourney v5

• And, for all the parents, an AI-copilot (take my money, Milo!)

NEW PRODUCTS

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

ChatGPT-4: Another step function in what’s possible. On Tuesday, OpenAI released ChatGPT-4 (the previous version that took the world by storm in November was v3.5). One of the biggest improvements is that you can now input images into the chat and it can comprehend them. OpenAI demo’d this by giving ChatGPT a chicken scratch sketch of a website landing page and GPT-4 produced a basic working webpage that looked like the sketch…it was a basic webpage, but impressive nonetheless (demo). A developer has already hacked together an app that enabled him to snap a picture of the inside of his fridge and GPT-4 gave him a list of potential recipes he could make based on it’s contents (twitter). Beyond understanding images, GPT-4 appears to be better at complex logic, coding, and creative writing. GPT-3.5 often fell short when responding in a particular style, such as a celebrity. GPT-4 is more impressive. Take a look at the differences below between GPT-4 and GPT-3.5. I asked it to explain how to make scrambled eggs in the style of Donald Trump, but to get distracted halfway through and go off topic. [Public Launch | $20 per month]

Claude: Anthropic AI’s rival language model is now available for early access. Claude was viewed by many to outperform ChatGPT-3.5. It was better at creative writing and other creative tasks. It could maintain context in a conversation for much longer and it could tell jokes that were actually funny. I’m very curious to see if this comparison still holds up with the launch of GPT-4. Claude has been in a closed beta until this week, when they announced they are going to be opening it up. Sign up for the waitlist here. [Waitlist]

Duolingo Max: Two GPT-4 powered features for language learners. A good example of the types of products that GPT-4 unlocks due to it’s increased ability to reason and comprehend more user input (ie. longer user sessions). Duolingo’s two AI features are: “Explain My Answer” and “Roleplay.” “Explain My Answer” is a chatbot that will explain why you got an answer incorrect. Super useful since there are so many nuances when it comes to language learning. “Roleplay” allows you to practice real-world conversation skills via chat. I expect we’ll see an increased number of these AI tutors that are trained for specific areas of study. Maybe Duolingo Max will be what finally helps me to learn French…je ne sais pas 🇫🇷 [Public Launch - 14 day free trial]

Khanmigo: Khan Academy’s GPT-4 powered tutor for students and assistant for teachers. In their announcement, they stressed the importance of ensuring that AI doesn’t hallucinate and make up facts when it comes to learning. They also highlighted that their AI tutor won’t tell students the answer to their homework assignment, instead it will help them understand the concepts behind the questions. Kudos. On the teacher side it’s supposed to help teachers come up with lesson plans. [Private Beta - but you can pay to jump the waitlist 💸💸 ]

Google vs Microsoft: Both are bringing AI to their suite of work productivity apps. On Tuesday, Google rushed to announce how generative AI is coming to Google Docs, Sheets, Slides and more. I think the coolest bit is that it will be able to take a document or email and turn it into a Slides presentation, and it will even add in some AI generated images. At Microsoft’s pre-planned event on Thursday, Microsoft previewed their new AI “Copilot” that they are adding to all of their productivity apps — Word, Excel, PowerPoint, Outlook, etc. This is all super exciting, but I wish they’d stop with the sneak peek PR announcements and instead announce their products when they are actually ready to be used by the public. Regardless, there’s going to be a big shift in how people work when the two dominant companies in the work productivity space have generative AI fully baked in. [Announcement]

Midjourney v5: Image generation also took another big step forward this week. Midjourney has been at the cutting edge of AI-generated images that are strikingly lifelike since their launch in 2022. This week, they released v5, which they’ve been working on for the past five months. v5 has a significantly higher image quality (2x resolution increase) and broader aesthetic range. Check out this thread from Nick St. Pierre for some great examples that compare v4 to v5. [Beta]

Milo: A copilot for parents. As a parent with two kids under three, this new product caught my eye. An early partner of OpenAI, the team at Milo have been building a copilot for parents that helps you manage the daily details of running a family. The first version of the product is a GPT-4 powered SMS copilot that “can take in all manners of chaos - grocery list, reminders by voice, bday party screenshots, school newsletter pdfs… and get them back to you (and everyone else in the family) as work cal invites or SMS reminders or easy lists.” I’m hoping this reduces the frequency of the thousand mile stares I give to my son’s teachers when they remind me at pickup (with a bubbly smile) that daycare is again closed early tomorrow. [Beta - Waitlist]

WHAT I'M READING

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

Technology Makes Us More Human by Reid Hoffman

"ChatGPT, a new AI system that sounds so human in conversations that it could host its own podcast, is a test of temperament. Reading between its instantly generated, flawlessly grammatical lines, people see wildly different visions of the future. For some, ChatGPT promises to revolutionize the way we search for information, draft articles, write software code, and create business plans. When they use ChatGPT, they see Star Trek: a future in which opportunities for personal fulfillment are as large as the universe itself. Others see only massive job displacement and a profound loss of agency, as we hand off creative processes that were once the domain of humans to machines. When they use ChatGPT, they see Black Mirror: a future in which technological innovation primarily exists to annoy, humiliate, terrify, and, most of all, dehumanize humanity..." THE ATLANTIC

WHO I'M FOLLOWING

People passionate about AI that are doing cool shit.

Danny Postma is an indie developer from the Netherlands that has been building a ton of unique AI products. From a fake model agency (linked below) to a tool that can add your face into your favorite meme. I also love that Danny shares a ton of behind the scenes from his projects.

HELP ME OUT

If you enjoyed today's issue (and are feeling generous), can you share it with your friends or colleagues? This newsletter primarily grows from referrals and a referral from you would mean a lot to me.

Refer just ONE person and I'll send you my write-up: "Trends and Opportunities of this AI Inflection Point." In this write-up, I have outlined seven AI trends that I have observed over the past six months, along with the opportunities they present.

Share using your personal referral link:

https://www.productreport.ai/subscribe?ref=PLACEHOLDER

PS: You have referred 0 people so far

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 or DM me on Twitter.