My Perspective on the Chief of Staff Role

 Check it out on a much prettier version on this Notion page

⚡️ A great Chief of Staff (Cos) is a high-EQ, operationally excellent, and strategic leader who helps move the entire organization forward every single day. 

Legal AI Is a Crowded Market

These are companies that just happened to flow through my inbox in the last 24 hours, no proactive searching 

• robinai.com 
• finchlegal.com 
• spellbook.legal 
• draftwise.com

And then there are the bigger players, like 

• ironcladapp.com 
• harvey.ai 
• evenuplaw.com

I haven't used these products, but I have a hard time thinking the value they produce is differentiated. 

Seems like a race to the bottom and war of CAC, which favors established players who can distribute Gen AI features through existing implementations.

On Perplexity - Timeline and Hypotheses

Perplexity baffles me! 

It’s a fast-growing startup that has built a great product and is stirring competitive reactions from the likes of Google and OpenAI.

At the same time, it's sustainable competitive advantage is unclear, and it's saddled with growth expectations attached to a $3B valuation (maybe $9B). 

To help wrap my had around Perplexity, I created a 2 year timeline to understand where they've been and help me think through where they might be going.

Read on for a fully sourced timeline of Perplexity's rise, covering:

• Fundraising
• Product Launches
• MAUs
• Revenue 
• Legal action
• and more

GenAI Litigation to Watch

Millette vs. Nvidia - https://www.404media.co/nvidia-sued-for-scraping-youtube-after-404-media-investigation/

UMG vs. Suno/Udio - https://www.theverge.com/2024/6/24/24184710/riaa-ai-lawsuit-suno-udio-copyright-umg-sony-warner

NyTimes vs. OpenAI - https://harvardlawreview.org/blog/2024/04/nyt-v-openai-the-timess-about-face/

Bartz vs. Anthropic - https://www.reuters.com/technology/artificial-intelligence/authors-sue-anthropic-copyright-infringement-over-ai-training-2024-08-20/

Dow Jones & New York Post vs. Perplexity - https://www.theverge.com/2024/10/21/24275924/news-corp-wall-street-journal-perplexity-lawsuit-copyright-infringement

You Get What You Click For

Algorithmic feeds are a mirror. 

For a remarkable number of folks, this reflection remains surprising, uncomfortable and frustrating.

Take Y Combinator partner David Lieb for just one case-in-point example. He recently shared his frustration with X/Twitter's algorithmic "ForYou" feed:

It’s largely clickbait videos and maga-Elon garbage.

The uncomfortable truth for David - and the rest of us - remains that algorithms across X/Twitter, Facebook, Threads, Reddit, TikTok, etc. don't pull content out of thin air. They respond to our behaviors, showing us more of what we engage with. 

The fact that David "For You" feed is populated with "fistfights and car crashes" (as mentioned in another tweet) suggests he has a habit of engaging with such content. If he truly found that content repulsive, he'd only use his "Following" feed.

Rather than get frustrated with the algorithms, you can bypass them with a click. However, that requires effort and awareness—two things often in short supply on these platforms.

In short, you get what you click for.

3 Good Recent Reads (Apple Vision, Choice Screens, AI)

9 Takeaways from the Vision Pro After 6 Months 

Matthew Ball

If one assigns 25% of this budget (5,000 of 20,000 patents) to the Vision Pro, the product has a price tag of $33 billion

‘Choice Screen’ Fever Dream: Enforcers' New Favorite Remedy Won’t Blunt Google’s Search Monopoly 

Megan Gray

...the EU Android choice screen has not changed Google’s market share.

Investing in the Age of Generative AI 

Kevin Zhang

SearchGPT's Example Results are Worse than Google

Only July 25th, 2024 OpenAI announced the SearchGPT prototype, and the twitter-sphere / threads-sphere reacted with "Search/Google is dead!"

I'm not so sure about that for a number of reasons, but OpenAI's own post demonstrates two

• "Designed to give you an answer" is big promise that may be hard to uphold
• Google is actually very good at producing useful results today

Read on 👇

Doug Shapiro on GenAI

You've probably never heard of Doug Shapiro, but he is writing some of the most cogent, grounded pieces on GenAI today. 

From his website (bolding is his, not mine)

He is not a futurist. His work is grounded in the very practical challenges of investors seeking returns and executives who must manage change even while balancing the needs of multiple constituencies (employees, investors and customers), combating institutional inertia and running a business.

Read his work 

• GenAI is Foremost a Creative Tool
• Fear and Loathing (and Hype and Reality) in Los Angeles
• Is GenAI a Sustaining or Disruptive Innovation in Hollywood?

The two best answers to "what is a startup"?

Steve Blank - What’s A Startup? First Principles. [2010]

Your startup is essentially an organization built to search for a repeatable and scalable business model. As a founder you start out with:

1) a vision of a product with a set of features,

2) a series of hypotheses about all the pieces of the business model: Who are the customers/users? What’s the distribution channel. How do we price and position the product? How do we create end user demand? Who are our partners? Where/how do we build the product? How do we finance the company, etc.

Your job as a founder is to quickly validate whether the model is correct by seeing if customers behave as your model predicts. Most of the time the darn customers don’t behave as you predicted. 

Paul Graham - Startup = Growth [2012]

A startup is a company designed to grow fast. Being newly founded does not in itself make a company a startup. Nor is it necessary for a startup to work on technology, or take venture funding, or have some sort of "exit." The only essential thing is growth. 

To grow rapidly, you need to make something you can sell to a big market. That's the difference between Google and a barbershop. A barbershop doesn't scale. 

I don't buy the "crypto 🤝 AI" narrative

For the betterment of the consumer, may I be wrong and eat my words 🫡