437: Paradox of Entrepreneurship, Disney to Sell ESPN?, Elon Musk xAI, FTC vs OpenAI, Sony & Microsoft, Ozempic, and American Psycho

"Not every experiment has to last forever."

I see my age in my children.

—Dave Chappelle

💍😴🛏️🌙📊 I’m finding myself checking the sleep metrics from my Oura ring less often lately.

After my ring dies, I probably won’t replace it.

I’m *not* saying it’s useless or that you shouldn’t get one if you think your sleep could use help.

I’ve been wearing an Oura since 2019, and it has made my life a lot better by helping me optimize sleep. It created a highly visible feedback loop between my actions and their impact on my sleep quality. What gets measured gets managed. It’s easier for me to be convinced by hard numbers and graphs and long-term trends than by the vibes of how I’m feeling.

It helped me change a bunch of behaviors. f.ex. I drink a lot less alcohol — even though I never drank much in the first place — because it really kills sleep quality. When you feel like dogcrap the next day, part of it is the alcohol directly, part of it may be dehydration, but a big part is also bad sleep.

But I’ve been tracking things for so long that I now have a pretty well-calibrated mental tracker that gives me most of what the Oura does, so I’ve reached diminishing returns. It’s similar to how I’ve been tracking what I eat in an app for years; now I can just look at a plate of food and have a much better idea of macros and calories than I did before I started tracking things.

I was getting huge benefits in the first few years, medium benefits after that, and much smaller benefits today.

Which is fine!

Not every experiment has to last forever. On to the next one…

🪡🧵🐦 If Threads really wants to kill it on the power-user experience, they should offer a setting panel with a dozen pre-made types of timelines that you can enable and then switch between in the app, plus the ability to make a customized type by mix & matching elements.

For example, maybe you’d keep the default algorithmic timeline but you could also have a chronological timeline for only people you follow. Or an algorithmic timeline but *only for people you follow* (that’d be great, actually!).

Sky’s the limit, I can imagine all kinds of ways of slicing and dicing:

How about an algo timeline but only for people you follow *and* for algorithmically selected posts from the universe of the people who you follow also follow (ie. go out in the network one degree of Kevin Bacon, so that you don’t see too many random influencers, but draw a wider net in the sub-communities you follow).

A timeline just for shared links from people you follow if you’re looking for good content to read/watch, etc.

None of this has to get in the way of casual users who will stick to the default algo view, but it would raise the ceiling for power users (who are the people who create most of the content — and thus most of the value — on these apps).

Of course, what’s good for Threads would also be great for Twitter, if they could only stop punching themselves in the face….

🎶🤖 Unless I’m mistaken, assistants like Siri and Alexa can’t be told to play music by more than one band at a time. That seems like a basic thing and I can’t believe they haven’t implemented that yet.

Sometimes you know what you feel like listening to, but don’t want to sit down and create a playlist for it.

It would be so convenient to ask to “play music by Modest Mouse and Dr. Dog and Houndmouth and Nathaniel Rateliff” or whatever and just have it shuffle a bunch of songs from all these artists mixed together.

Or the same kind of thing, but where you name albums or genres instead of artists.

Hopefully, when all these ‘assistants’ get upgraded with LLMs they’ll get smart enough to do this basic multi-artist stuff…

---

🏦 💰 Liberty Capital 💳 💴

🔨 The Paradox of Entrepreneurship 🔄

Friend-of-the-show Cedric Chin (🇸🇬🥋) writes:

The paradox of entrepreneurship: 

Have almost too much self belief, because you want to persist long enough to get lucky.

Also: if you don’t listen to people who call you out on your bullshit, don’t be surprised when the market punches you in the face.

Here’s another one I thought of:

• You can make better decisions by being well-calibrated.

• But if you are very well-calibrated, you probably won't even attempt to work on some of the most difficult problems out there where the odds favor failure and defeat.

It’s a good reminder that life is usually about finding the right balance to be effective, rather than going out as far as possible to one extreme (which rarely turns out well — especially when looking at the whole picture and not just at one metric).

See also:

• 🎧 Cedric Chin: What Operators can Teach Investors, the Study of Expertise, Knowledge Shields, and More (Podcast #10)

🪡🧵 ‘Instagram to shut down Threads as Meta consolidates its messaging platforms’ (2021 headline) 🪦🧶

I think few people remember the original Threads, or know that the current version that has taken over the Zeitgeist is actually a reworked V2.

Here’s what the original offered:

Threads was a similar idea to Facebook Messenger — it allowed users to send and receive Instagram DMs in a dedicated messaging interface outside of the main app. It had a collection of quirky features, including automatic statuses that the app could set for you based on what your phone was doing. All of those tools will be available in Instagram itself by the time Threads shuts down, the company said in a statement to TechCrunch.

This does show that having a huge user base doesn’t mean that anything you do will automatically be successful:

not a heck of a lot of people were using Threads to begin with. The app had just over 200,000 users as of last year, which is a tiny percentage of Instagram’s billion-plus user base. (Source)

You need at least some level of product-market fit and the right conditions for people to be willing to try your thing. ie. The growing dissatisfaction with Twitter.

But the above wasn’t even the first time that FacebookMeta had experimented with a Twitter-like app:

The company’s Instagram unit even began building a text-based prototype similar to Twitter, which they referred to by a similar name, Public Threads, in 2015, according to a person with direct knowledge of the effort.

That project didn’t go anywhere. In 2016, Instagram shelved the Twitter clone in favor of new visual features it was building—copying Snap rather than Twitter—the person said.

There’s no doubt an alternate timeline out there in the metamultiverse where Twitter has been firing on all cylinders for a few years and Threads has had a very different launch (in fact, it probably wouldn’t have been launched yet by this point if Twitter wasn’t struggling).

📺🐭 Bob Iger is open to selling ESPN and ABC 💰🏈🏀