Introducing Gauntlet AI

An extremely intensive 12-week AI training for engineers.

4 weeks remote + 8 weeks all-expenses-paid in Austin, TX. 80-100 hrs/week.

Participation is 100% free.

Anyone who completes The Gauntlet receives an automatic $200k/yr job in Austin, TX. > Wait, what’s happening?

I know. Crazy.

About a year ago we began working with a handful of companies to train their engineering teams to use the cutting edge of AI.

We were all completely blown away by what we saw. By what the smartest people wielding AI are capable of.

From the now-declassified 1954 CIA manual on assassination, I can’t stop thinking about this line:

“Obviously, successful secret assassinations are not recorded as assassination at all.” Link: nsarchive2.gwu.edu/NSAEBB/NSAEBB4…

In one of the most fascinating podcasts I’ve ever listened to, the members of NPR’s This American Life got their testosterone levels tested.

All of the males had very low testosterone; some low enough doctors would recommend medication.

And a trans guest talks about how taking testosterone made them suddenly interested in science.
I think it’s highly likely that our hormones and microbiome play a much bigger role in our thoughts and decision making than we give them credit for.

The Supreme Court's Chevron ruling may be most impactful things to happen to startups in a long time, in ways that people don't realize.

A thread: From AP: "The court’s 6-3 ruling on Friday overturned a 1984 decision colloquially known as Chevron that has instructed lower courts to defer to federal agencies when laws passed by Congress are not crystal clear."

It’s difficult to describe how insane a 25% tax on unrealized capital gains is.

Not a one-time 25% hit. It’s compounding, annually taking 25% of every dollar of potential increase before it can grow.

Not an exaggeration to say it could single-handedly crush the economy.

https://twitter.com/mikebeckhamsm/status/1783204586041364772

As my father-in-law (who is a farmer) would say, “You’re eating your seed corn!”

Holy smokes.

OpenAI board member Helen Toner published an article Altman took issue with.

She described it as “an academic paper that analyzed the challenges that the public faces when trying to understand the intentions of the countries and companies developing A.I.”

lol.

👀 Suffice it to say that’s not _all_ the article talks about.

The article is literally an analysis of different ways you can force AI companies (and governments using AI) to slow development, and recommendations on how they can be used and which are best.

PSA: If you’ve wanted to travel to Maui, it’s probably the best time ever to visit (other than Lahaina proper).

Beaches are empty. Airbnbs, hotels, rental cars are at 1/4 the usual price. Flights into Maui are empty.

Lahaina is closed. Rest of the island is safe and desperate. Lahaina is closed. Maui is open.

There people who travel to places only after they’ve stabilized from recent natural disaster.

Pretty clever.

You get the place to yourself and rock bottom prices. Everyone desperate to get the economy going.

That’s where Maui is right now.

There are two options at this point:

1. FDIC moves very quickly
2. Mass layoffs The very brief breakdown of what happens from here:

1. If you had FDIC insured accounts in SVB (<$250k) you should get that cash shortly, next few days.

2. If you had accounts that weren't FDIC insured, you become a creditor. There's an asset firesale, you're first in line.

~90% of the mistakes startups make is realizing x needs to happen but not doing x for some reason that ultimately ends up being unimportant This is why I'm convinced down markets are very, very healthy for startups.

Less time and money to waste, less need to spend all of your time trying to keep everyone happy, less optimizing for what investors want.

Building real businesses.

Calling it now:

SpaceX is going to be one of the biggest companies of all-time, and almost nobody realizes how big it will be.

Short version:

Best team of all-time
Biggest market of all-time
Deepest moat of all-time
Fastest-growing revenue of all-time. Yes, they won the launch business.

SpaceX is reusing rockets for the 15th time, no one else is reusing.

SpaceX launched 61 orbital missions last year and carried 70% of all payload to space in 2022. Insane.

And estimates say starship will drive their prices down another 20x.

I don’t know how many of these painfully low budget, not-quite-Hallmark-channel Christmas movies Netflix is going to make, but what I do know is my wife is going to watch every single one of them I don’t know which one she’s watching now but a recently divorced writer is living in Scotland trying to write another book (her career depends on it) and trying really hard to not fall in love with some single-for-no-particular-reason silver fox who lives in a castle?

When AI gets 99% of the way to replacing a skill we say that skill is being automated away.

I think that’s entirely backwards.

You still need that human, but those humans can now harness AI for insane leverage.

They’re not being replaced; they’re becoming 100x as valuable. Look at programming.

AI is pretty far away from replacing programmers.

But now if you’re a programmer and you can build 10x as much stuff because AI is doing the heavy lifting…

Can you even wrap your mind around how valuable that is?

Business Insider openly admitting that they don’t care (or check) whether what they’re publishing is true or not because they said “reportedly” in the headline News organization in which SBF is a major shareholder claims that SBF owns $100 million of Twitter.

Owner of Twitter says this isn’t true.

Headline: “SBF is reportedly a shareholder in Twitter.”

Zero attempt to actually determine what is accurate.

How headcount explodes in a startup:

You do a critical task. It takes 1/8 of your time, and is easy and cheap to hire for. So you hire someone to do the task.

That person does that task well. Great hire! In fact that person is so smart and resourceful that they start doing two other tasks that, while not critical, are great. What an awesome hire!

Add those two great things to the list of things we do!

Someone referenced the notion that you can't work productively for more than 5 hours/day, and I realized I had always heard that but had no idea what the notion came from.

Googling around you find and endless number of sites linking to each other, but eventually found the study. Stanford researchers measured hours worked vs. productivity in a bunch of easily measurable (generally manual labor) jobs.

Easy to measure hours worked vs. amount of output.

I was surprised to see the results were... pretty much linear?

bullish “Twitter doesn’t have a comms department? 😱”

The CEO already sent a <280 character message directly to 115 million people multiple times today.

There’s your comms department.

IMO what Twitter should actually do:

Blue check for $8/month that means you're a real/legit person (ID verified). Get a bit of priority in search/replies and some premium features.

VIP "Gold check" that is free if you're a celebrity or expensive ($1k/month) if you're not. I think most people underestimate how much value Twitter can bring to power users.

I don't even want to think about how much I would actually pay for it if I were paying for 100% of the value it provides.

Some of the most frustrating times to me as a founder were budgeting, goal setting, and headcount planning, because incentives shift.

By default people ask for maximum team size, set “goals” they’ll hit automatically, and you end up planning for failure.

What I’ve seen work: 1. First, actually optimize for *net income,* (profit), not just revenue.

Yes, some VCs/analysts don’t consider the bulk of your team “COGS,” say you have 99% gross margin, and focus 100% on revenue.

But operationally that’s a fantasy.

You have to care about revenue - cost.

Did a bunch of reading about this over this past week.

Turns out most gamblers aren’t mismanaging risk. They don’t really care at all whether they’re winning or losing.

Playing slots is addictive because it lets you escape distraction and enter an almost flow-like state.

https://twitter.com/Austen/status/1543671237024440321

People described slots as “entertainment” but that doesn’t really do it justice.

And the psychological phenomenon isn’t waiting for the high when you win.

In fact some prefer to lose - it gives them a bigger rush.

“My family and I were homeless at the time I attended BloomTech. I worked hard to complete my classes while staying in hotels and homeless shelters.

In two months I’ll be celebrating a year with the company that hired me.

Bloom helped me go from homeless to 6 figures!”

❤️ The software engineers and data scientists who will power the future of our economy are everywhere—yes, including homeless shelters.

We are SWIMMING in untapped human potential.

To unlock it:

1. Identify it
2. Take person all the way to the outcome
3. Minimize risk of failure

My only takeaway from all this is that some people need to spend way more time outside Full interview: wired.com/story/blake-le…