521: How to Save Intel and Give U.S. its TSMC, Nvidia Antitrust Probe, 3D-Printed Blood Vessels, Internet Archive, and Gravity Falls

"decades of tacit knowledge"

When will we realize that the fact that we can become accustomed to anything, however disgusting at first, makes it necessary to examine carefully everything we have become accustomed to?

—George Bernard Shaw, A Treatise on Parents and Children (1910)

🧭 🗺️ ⏱️ Someone recently mentioned isochronic maps — I think it may have been Ethan Mollick — and I wasn’t familiar with the term, so I looked it up.

It turns out I had seen them a long time ago, but I didn’t know their name.

They’re travel time maps showing what is geographically accessible from a starting point within a specified time period, often using a particular mode of transportation.

The one above shows the rail network of the Austro-Hungarian Empire in 1912, starting in Vienna.

Here’s one showing how far you can get from London in X number of days in 1881:

🍀 Doing is the best way to learn. Trying things is the best way to increase your surface area of luck!

It's no surprise that people who rarely try anything are most likely to complain about how unlucky they are ¯\_(ツ)_/¯

📖👀🏄‍♂️🧠 As a follow-up to the intro of Edition #520, friend-of-the-show Mark Heyer left a very interesting comment regarding balance boards:

Enjoy your balance board. You might be interested to know that during the 1960s, NASA discovered that after exercising on the balance board, reading comprehension improved. 

I used one at my standing desk for a while, which is cool, but it also inspired an app, yet to be realized. Use the phone in your pocket to sense your movements on the board and create an app that synchronizes visual displays on the computer screen in front of you. The possibilities are endless. 

Among other things, I think we could train ourselves to read better and faster. Reading harnesses our proprioception which synchronizes our physical movements with our vision. This is a whole world still waiting to be explored.

I had no idea. Fascinating how everything is connected!

🏎️ 🚧👷‍♂️🛠️🏗️ The history of buildings things fast needs to be studied and learned from:

In the beginning (i.e., World War II), America built things quickly. The Pentagon was built in 16 months. The Manhattan Project ran for less than four years. The Apollo Program put a man on the moon in under a decade. Kelly Johnson, of Lockheed’s Skunkworks, designed the SR-71 Blackbird, the fastest manned aircraft ever made, with pencils and a slide rule. In the 1950s alone, America built five generations of fighter jets, three generations of manned bombers, two classes of aircraft carriers, submarine-launched ballistic missiles, and nuclear-powered attack submarines.

Source.

---

🏦 💰 Business & Investing 💳 💴

I made this with FLUX 💚 🥃

✂️🐜💰☁️ How to Save Intel and Give the U.S. its TSMC 🏗️👷🚧🛠️

I’ve been closely following the commentary and analysis on Intel lately. Ben Thompson (💚 🥃 🎩) has great stuff, including this interview/conversation with Ben Bajarin and his piece Intel Honesty.

When I zoom out and think about what truly matters, the bottom line is:

• It’s all about Intel’s foundry business. 🐜

  • The chip design division isn’t a national security matter and it’s not a scarce asset. There’s plenty of competition for what they do and more coming on the market all the time.

• The Foundry business is an almost unique asset. 🌟

  • Startups are not entering the space. Few entities on the planet have the financial firepower to even consider being active at the bleeding edge.

  • Intel Foundry has decades of tacit knowledge embedded in its workforce that would take a long time to replicate even with infinite money.

• For the Foundry business to have a chance of succeeding at transitioning to a TSMC-like entity, it needs

  • Lots and lots of capital 💰💰💰💰💰

  • Enough customers to reach scale 🙋🏻‍♂️🙋🏻‍♀️🙋

• Intel itself probably *can’t* provide that capital or be that customer if it truly wants to change the Foundry’s incentives and culture to be third-party customer-oriented. 🛑

So far, the muddle-along approach seems to have been slowly sprinkling government subsidies and Intel keeping the band together because the foundry unit would run out of cash long before it could transition. Can we improve the odds?

Intel needs to be split up ✂️ ("We had to destroy the village to save it.")

Also made with FLUX 💚 🥃

The best choice is to let the part of Intel that makes money continue doing so. Let the chip design operation go its separate way and become an unshackled fabless entity like AMD and Nvidia. Then set up the Foundry — the entity that is important to national security — in a standalone way that can actually succeed.

Intel’s foundry could offer a highly sought-after product. That demand comes from Big Tech. We just need to do a bit of restructuring and matchmaking to get from here to there…

The consortium idea mentioned in Edition #520 would allow Big Tech to deploy a little bit of the cash on their balance sheets into an asset that has the potential to someday save their bacon if something ever happens to TSMC. The absolute numbers are big, but considering the cumulative market cap that rests on semiconductor supply, it would be wise and inexpensive insurance (remember the inverted pyramid with trillions and trillions of market cap resting on TSMC at the bottom?).

It doesn’t even have to be that dramatic. Even if TSMC just slowly loses its way and becomes worse (because when you’re alone at the top for too long, you usually become slow and fat…), it would be great to have an alternative to keep them on their toes and keep wafer prices lower by avoiding a monopoly situation at the leading nodes.

There’s a coordination problem to make this happen, but this is where the US Government can come in. Instead of handing out subsidies that have little chance of success, it could instead herd those cats and coordinate this project and very carefully target things to create huge leverage with whatever it spends. It could also assure Big Tech that this is a national security matter and they won’t face legal troubles for joining the consortium.

In other words, rather than giving Intel $ X billion, what if it offered Big Tech to pay for the difference between TSMC chips and Intel Foundry chips for X years? That same $ X billion in guarantees could be leveraged into multiples of that amount in spending from Big Tech by reducing their risk. This would help Intel Foundry ramp up to enough scale from 3P customers to make it viable.

✋ Intel announces cancellation of 20A node for Arrow Lake CPU, will use TSMC instead of its own foundry 🤔

Speaking of Intel’s trouble, this is one more example of why they are headed in the direction of being a fabless chip designer, only slowly and painfully rather than ripping off the band-aid:

In a surprise move, Intel announced today that it no longer plans to use its own 'Intel 20A' process node with its upcoming Arrow Lake processors for the consumer market. Instead, it will use external nodes, likely from partner TSMC, for all of Arrow Lake's chip components. Intel's only manufacturing responsibilities for the Arrow Lake processors will be packaging the externally manufactured chiplets into the final processor.

Here’s how Intel explained things:

One of the benefits of our early success on Intel 18A is that it enables us to shift engineering resources from Intel 20A earlier than expected as we near completion of our five-nodes-in-four-years plan. With this decision, the Arrow Lake processor family will be built primarily using external partners and packaged by Intel Foundry.

Of course, Intel is spinning this as a good thing because 18A is doing so well, but who really thinks that they wouldn’t be fabbing their own chips if they could rather than give that margin and wafer volume to TSMC instead?

🤖😎🔍⚖️👩🏻‍⚖️ ‘Nvidia Faces DOJ Antitrust Probe Over Complaints From Rivals’ 🤔