Nowism - Edition 37

Mid April musings

The Big Ones.

Allbirds don’t fly

From $4bn to $39m. Once-viral sneaker maker Allbirds sold this month for a mere $39 million, and I’d like to use this moment to be miserable.

We have been in this strange window of time, starting around 2014, where the assumption was that contemporary companies were no longer in the business they were in, but “tech” companies.

I don’t mean the classic “we’re not a pizza company, we’re a tech company that happens to make pizza.” I mean a complete failure of the market to understand that technology didn’t change the fundamentals of most businesses, while assuming that tech changed the very economic physics of the world.

Somehow, because we had the internet, anything could go viral. Because we had consumer data, any small company could scale fast. The TAM of anything was basically the world , because why not.

Whether it’s Blue Apron, Oatly, Beyond Meat — people forgot that selling food isn’t the best industry ever to be in, and it’s not an industry destined to follow hockey-stick graphs (other than ad spend).

So from Shake Shack to Stitch Fix to Casper to SmileDirectClub to FIGS to Away to Outdoor Voices, to Peloton, to Chewy, to Rent the Runway, to Glossier to BarkBox, we’ve got a plethora of wonderfully implemented ideas that are often now worth 1/50th of their peak.

Remember stay-at-home stocks? Remember “P&G will be destroyed”? Remember “you’re either a startup or left behind”? It’s just not true. Pretty much every single sexy company used in trend decks from 2014 to 2021 as “the new way to do things” , the ones that were supposedly going to destroy P&G , turned out to have four massive errors:

1. New tech didn’t change unit economics. You still had to get stuff to people.It was still better to buy bulk. Reverse logistics still sucked.

2. Out of touch. Just because a VC bloke in SF thinks a $2000 bike is cheap, doesn’t mean anyone in the normal world can spend that much. All these brands were aimed at hundreds of millions of wealthy people who, it turned out, don’t exist. Your friends at the Harvard club are not illustrative.

3. Flawed exit expectations. None of these brands really wanted to IPO. They all wanted to grow fast and big enough to irritate a massive company into buying them — which was always risky. Especially as if they got big enough to be solid, they also got tricky to gobble up.

4. They forgot to try to build a mass-market brand. They thought they could performance-market their way into people’s hearts. It doesn’t work like that.

The internet didn’t change that much in some industries.

What people forgot is that while technology may instigate a bit of a shift, the longer you are in an industry, the more the questions become the same ones the industry has always faced. They become about logistics, about brand, about retail, about NPD. You may start out as a totally new car company, but you end up talking about warranties and lobbying. You may set up with a focus on codecs and streaming, but you end up talking about how to make great TV and sell ads. Etc.

And that means knowing how an industry actually works is always rather handy. I bought a pair of Allbirds once and thought they were great. But a couple of years later, bored in Auckland, I went to one of their flagship stores , and it was, without doubt, the saddest, emptiest shop I’d ever been to in my life. And I went shopping in Belgrade just after it had been bombed.

2) The Infrastructure Trap

I found this chart interesting.

What it doesn’t ask, and what really matters, is:

Did the investment actually pay back and more crucially, who actually made the money?

Oddly, every major technological wave follows the exact same script. The pattern is eerily consistent:

- Insane capital expenditure spikes as a percentage of GDP

- Wild hype and frantic overbuilding

- Bubbles, busts, bankruptcies, and a mountain of broken balance sheets

- Almost none of the people who poured the money into the infrastructure actually made the big money

Railroad investors in the 1870s? Many lost everything.

Telecom and fiber players in the late ’90s? Trillions evaporated.

The infrastructure layer always looked like the future… until it became a commodity.

So who actually made all the money?

The real fortunes went to the companies that showed up after the infrastructure was in place and dirt cheap.

The ones who built entirely new business models, platforms, and experiences on top of it.

After the railroads came the national retailers and manufacturers — think Sears and Roebuck.

After electrification came the appliance makers and the modern factory — General Electric sends its thanks.

After the internet buildout came Google, Amazon, Netflix, and the entire digital economy. Not the fiber companies that went bankrupt.

Even today, compare the razor-thin margins of T-Mobile with the eye-watering profitability of Meta.

The infrastructure investors took the risk. They enabled the revolution. The application layer captured the value and now here we are again with AI.

Hyperscalers are spending hundreds of billions on chips, data centers, and power. The capex numbers are staggering. The hype is deafening.

You would have to be nuts to think that this time it’s different.

The winners will be the ones who figure out what to do with all that intelligence once it’s abundant and dirt cheap.

There’s rarely a moat in pure infrastructure. There never has been.

One final thought:
Every technological revolution so far has yielded real societal benefits , new prosperity, productivity, human progress. Let’s hope this one does too.

Little ones.

• Nearly Half of Planned 2026 U.S. Data Centers Face Delays or Cancellations

  This is the great irony of our age: the most advanced civilization in history is still dependent on 20th-century electrical infrastructure made overseas. Progress is limited by supply chains, public opinion physics. Same as it ever was. And somehow people keep forgetting to see this coming.

• The first “one man AI unicorn” story was interesting. Lots of people piled on to point out the flaws, illegal ads, regulatory shortcuts, lies, all good points. Something that hit me, It wasn’t really an AI startup. Everything they were doing could have been done five years ago. The real story was our software is now compartmentalised / containerized, how building even complex sites is merely a matter of orchestrating plugins. I’m not saying what they did is easy; I’m just saying it’s not new, outside of Fake people made easily with AI.

• Unilever has agreed to acquire Grüns, the explosive U.S. greens supplement gummy brand launched in 2023, in a deal estimated at around $1.2 billion.
  
  Proof that sometimes the fastest path to a billion-dollar exit is still a well-timed consumer fad , and maybe my first post on Allbirds is wrong’o’rama. In fact 11 brands have been acquired for more than $1bn in the last 2 years per Iris. I’m always confused that Pepsi needs to buy a Poppi, or Nestle had to acquire Blue Bottle, when these companies are the very best placed entities on the planet to create such units, more on that another day.

• It’s weird that people think half the US thinks the other half is morally bad. I don’t know how we got to a situation where we assumed people were evil, rather than just a different vantage point, different lived experience, different outlook, different priorities, different information sources. Is empathy that hard? People can make terrible decisions without being terrible people. Other countries manage to see things differently , Canadians are somehow so nice.

• For a long time I’ve said the economy is weird. We may have to accept it’s just really shitty and people don’t want to look down.

  

• Social media is populist and polarising; AI could be the hope to correct that.

• Somehow I ended up in a rabbit hole and got fascinated by illegal numbers

• The Redpoint Ventures market update has some fantastic figures, like the chart in my second piece

• Revolut becomes the first NeoBank to make serous bank, whether this is the sign of an industry that’s resisted disruption cracking, or just an exception, we’ll soon find out

• OpenAI tried to make an app store, What didn’t they try to make?, It’s failing, but it would be an amazing project to work for. What should apps on LLMs look like?

• For some reason, neither Kalshi nor Polymark let you bet on OpenAI’s demise, It seems almost guaranteed to me. Not to pile on the misery, but I do think there’s a high chance Sam Altman will end up remembered like a true scammer

• The Unsung Women Inventors Who Transformed Household Chores

• I couldn’t explain why lots of articles drive me nuts, but this piece explained it. Everything is just being repeated rather than questioned. The CEO SAID A THING

Me Me Me

I’m going to be based in Europe for the summer, which means I can do Keynotes / Workshops there more readily ( July- Oct)

Something that comes up in my Workshops often is how varied LLM outputs can be for similar but different types of tasks.

So I’m working on a how-to guide for LLM’s for research, writing and general text generation.
Not something awful like “the 10 best prompts” or “make no mistakes”

But “what are things that it’s a lot better than we expect to be”, and “what are things that it’s a lot worse than we hoped” . It should end up being a very practical guide for when to and how to use it. I think a lot of people don’t realize how good it can be. And an equal number of people don’t realize how bad it can be

My workshops……Bespoke and starting from $25k USD….