Happy Sunday 👋 Some personal news - after nine months on disability / off work recovering from my brain injury, I'm finally going back to my role as a PM @ MasterClass tomorrow! I'm not yet fully recovered and still run into symptoms daily so it's not a slam dunk, but I am so so grateful to be working again 🙏

Welcome to all the new readers and those already tuning in every week. As a reminder, TTS is now a twice weekly email:

1. Friday: Curated list of new features and products from the week

2. Sunday: Breakdown of a new product launch + other good finds

One Big Ship

Open AI shipped ChatGPT, a general-purpose chatbot.

Details: The free chatbot is now live in a "research preview" public beta. Some highlights include:

• Answers general & technical questions

• Writes content such as messages, letters, blog posts, and essays

• Remembers what users have said earlier in conversations

• Allows users to provide follow-up corrections

Of note - ChatGPT is based on new "GPT-3.5" models, which Open AI quietly introduced with the release of the chatbot.

Real-life examples: ChatGPT has to be seen to be believed. I've included links below to a few wild examples, although you really have to check it out for yourself!

• Writing a python script for trading strategies

• Step-by-step redesign of Git

• Writing a letter to a child explaining that Santa isn't real

• Writing an introduction script to a podcast

The big picture: In the past few months, generative AI tools for writing and image-creation have seen significant traction. With the release of ChatGPT, AI increasingly appears to have won the "next big platform" debate.

Google's conundrum: ChatGPT represents a real threat to Google's search business. For several types of questions, ChatGPT already yields strong results that don't require users to sift through multiple articles and ads.

The chatbot doesn't yet include live browsing, but Open AI is working on a WebGPT model that searches the web and give sources for answers. WebGPT is the real Google-killer, and according to Open AI it's only a few months away.

Google's in a tough spot. They likely have very similar AI capabilities, but productizing them would cannibalize Google's most valuable business.

One thought: While the core models have been improving, the underlying capabilities have existed for years. The reason ChatGPT is blowing up is it's conversational UX, which makes it easy (and fun) to play around with the technology.

While model quality will obviously have a huge impact on who wins in the space, UX and customer acquisition will play an equally large role.

What to watch: If ChatGPT is already churning out passable python code and written content, what are we in for when GPT-4 comes out next year?

Good finds

Adam Fishman - No, you shouldn't do a Spotify Wrapped Campaign

Why Wrapped works for Spotify:

1. Fun & personalized reminder of how much value users got from Spotify

2. Very strong data mapping and ranking capabilities across Spotify network

3. Experience is deeply integrated with the core product experience

Why your company shouldn't do Wrapped:

1. Takes months of work for marketing + product, data, engineering

2. Insufficient data leads to bad output and can highlight lack of product usage

3. It will bomb if not given sufficient exposure in the core product experience

4. It will bomb if creative and marketing resources aren't sufficiently allocated

Nikita Bier - Growth v core product efforts for the Gas app

Gas is a viral teen social media app that's been topping the App Store charts.

How has Gas spent their engineering time?

• 10% focused on the core product experience

• 90% focused on the friendfinder system

This shows the importance of initial activation / day zero features in consumer tech. Teams too often focus on building lower-down-the-funnel core functionality, despite the majority of their drop-off occurring higher up the funnel.

Gibson Biddle - Defining product strategy in a new job

Defining a SWAG (Stupid Wild Ass Guess) product strategy in 2 weeks at a new role:

• 1) Define the product you're seeking to build

  • What the product is | what benefits it will deliver | personality of the product

• 2) Iron out brand promise of the product

  • Mission -> product emotional benefit -> product value -> product features

• 3) Outline a product vision

  • GLEE: Get big on ___ | Lead ___ | Expand ___ | Expand ___

• 4) Outline product strategy hypothesis

  • DHM: Delight customers in | Hard to copy | Margin enhancing ways

• 5) Define north star metrics

  • GEM: Growth metric | Engagement metric | Monetization metric

• 6) Articulate tactics

  • SMTs: 3-6 Strategies with an associated Metric and Tactics

Twitter shenanigans

And that’s a wrap for today. Stay tuned for more!

- Amol