122: TSMC is already making Apple's M2!, Microsoft's Regalia, Amazon's Video & Music Big Bucks, ARM's new Data-center CPUs, Dropbox, Roper, and Josh Wolfe

"That’s about twice as fast as a 'top form' human"

Teachers should prepare the student for the student's future, not for the teacher's past.

—Richard Hamming

(I’m sure many of you have read Hamming’s 1986 speech ‘You and Your Research’, but if you haven’t, it’s classic for a reason:

  • You and your research (Richard Hamming, Transcription of the March 7th, 1986, Bell Communications Research Colloquium Seminar)

) 1

⏫ Follow-up on standing desks:

Friend-of-the-show Willis Cap shared his setup, which is basically a way to turn a regular desk into an adjustable standing/sitting desk for about $70:

I don’t think it’s quite what I’m looking for, but it’s very cool and I want to share it (Amazon link) because I bet some of you have been considering trying a standing desk and may be more likely to give it a try with a cheap, non-permanent option like this.

Friend-of-the-show and supporter (💚 🥃) Daniel Abood also shared his setup, and it sounds very close to what I want (though I’m still hesitating on the width, and may go a bit wider than my current desk to see if I can find productive use for the extra surface — but I’m also afraid it’ll just end up more cluttered).

Daniel has an Uplift 24” like this:

He says:

I’ve gone through a ton of these things and this one is high quality, durable and the motor works very fast between standing and sitting. It’s a little bit more expensive than the others but its customizable and keeping with the tone of your apple section, well worth it!

I love how Uplift and Jarvis allow you to customize everything about their products.

It’s the desk version of when I used to build PCs as a kid, and spent hours agonizing over which version of each component (CPU/GPU/case/motherboard/storage/fans) I’d get…

🛀2 The best thing about having multiple interests is that you can always find something new that’s interesting in at least one of the fields you follow. And you can bounce between things — I find that especially powerful with long form (books, papers).

ie. it’s not because I’m too tired to read about X that I’m necessarily too tired to read in general, so switching to something else may allow me to keep going a while longer.

That’s why I usually read 2-4 books in parallel (well, I did before the pandemic… been much harder to read long-form in the past year, but t’is but a bump in the long road).

I read that way because there’s almost always one of the many books that I feel like reading at any time, while if I had just one, I may not feel like that specific one on that day.

🤔 If you had unlimited resources, what would you do all day? I don't mean for a vacation, but for hours each day, over years. If money and diplomas and training weren't an issue, what would you do?

Is there really something stopping you from getting there? Be careful about the sunk cost fallacy.

💚🥃 My current goal for this project is to get to 5% of readers who are $ supporters, and 95% who read for free.

That’s my Bill Gates line, if such a concept can be mangled enough to apply to newsletters (if you’re not familiar, check out Ben Thompson’s excellent explanation).

This elite 5% group is basically paying a round for the whole table once in a while to keep the party going, and keep me talking (except that instead of costing hundreds of dollars like a real round of drinks, it’s $10/month for 12 editions, or 83 cents/ed).

This is the most win-win model that I can think of, because I know there are many at the table who are students or in between jobs or who live in low-wage countries where $10 isn’t just couch-cushion change like it is for many of the highly-paid finance and tech people that I know are reading this (I can see your emails — I see all the .com domains from hedge funds and big banks and tech companies!).

So yeah, if you want to help me move the needle to that 5% (currently at 3% — about 130 paid subs) to help make this project sustainable over time, you can become an elite round-buyer in 20 seconds by clicking here:

💙 Subscribe now 💙

(what happens if we get above 5%? Good question… I may have to start making more 𝕤𝕡𝕖𝕔𝕚𝕒𝕝 𝕖𝕕𝕚𝕥𝕚𝕠𝕟𝕤 — interviews and single-topic essays and company features — and maybe podcasts or group chats with readers on Twitter Spaces or Zoom or something… It’ll be easier to consider this kind of stuff when the pandemic is over and I get back some of my time and mental energy. Right now my ‘to do’ grows faster than it shrinks because of remote-school and constant interruptions to my days)


Investing & Business

‘Amazon Spent $11 Billion on Prime Video and Music Content in 2020, up 41% From Year Prior’

Amazon spent a cool $11 billion on TV series, movies and music for its Prime services last year, an increase of 41% from $7.8 billion in 2019, the ecommerce giant disclosed in its annual report Thursday.

In edition #118, I wrote about the ginormous price tag of Prime’s Lord of the Rings series (“roughly NZ$650 million – $465 million in U.S. dollars – for just the first season of the show”), but clearly, even accounting for that, there’s a lot more going on and their ambitions are still ⬆️.

Muscular men throwing pointy balls probably accounts for another big chunk of the increase:

Last month, Amazon inked a 10-year deal with the NFL nabbing exclusive rights to “Thursday Night Football” starting in 2023; the pact works out to $1.32 billion per year

On the other side of the fence:

On a cash basis, Netflix spent $11.8 billion on streaming video content in 2020, compared with $13.9 billion the year prior, as the coronavirus halted productions for the better part of the year. However, like Amazon, Netflix also has content payments due over several years, and those content obligations totaled $19.2 billion at the end of 2020. (Source)

h/t Friend-of-the-show Jerry Cap

Dropbox & Scuttleblurb

Friend-of-the-show and 𝕖𝕩𝕥𝕣𝕒-𝕕𝕖𝕝𝕦𝕩𝕖 supporter (💚💚💚💚💚 🥃) David Kim is showing all of us the power of the written word. He recently wrote a piece (sub $ required) about Dropbox and Box:

And thanks to the magic of Twitter, on the very same day, here’s the president of Dropbox:

Cool, ain’t it?

Roper Graphs

Roper released Q1 results yesterday. Seemed good to me, but what I want to note is a change that they made to their segment graphs:

I like it. Showing both organic and total growth vs previous year in a compact way inside the bar. It’s a nice addition to this even better thing that they’ve been publishing in every presentation appendix for a few years, disaggregating organic growth, FX, acquisitions/divestitures for every company segment.

More companies should do this.

Microsoft Segments Over Time

Speaking of graphs I like, this one by CNBC is quite good too:

Regalia: “The term can refer to the rights, prerogatives, and privileges that are held exclusively by any sovereign, regardless of title (emperor, grand duke, etc.). An example of that is the right to mint coins.”

Interview: Josh ‘de-Lux’ Wolfe

Another great conversation (because that’s what these things are, conversations, more than interviews) by friend-of-the-show and supporter (💚 🥃) Jim O’Shaughnessy3:

I like that while most of his peers spend their time thinking about how to lower CAC on some ERP SaaS app and which part of Chinese superapps should be copied or whatever, he's thinking about the national security implications of AI and doing high-precision manufacturing in space.

So many of the ideas he mentions in passing could — and should — be a whole podcast episode on their own. Some of my fave podcasts are kind of like that, like Sam Altman going on a tangent about nuclear power rather than about how to raise a series A.

I mean, it’s fine that there’s a big rush of talent and resources to some areas, but let’s not neglect the important hard things that are less popular.

It reminds me of a great Aubrey de Grey quote (just replace “scientists” with “VCs” and “entrepreneurs”):

It has always appalled me that really bright scientists almost all work in the most competitive fields, the ones in which they are making the least difference. In other words, if they were hit by a truck, the same discovery would be made by somebody else about 10 minutes later.

If what you’re working on/funding would’ve gotten made/funded easily without you, what you’re doing is more zero-sum — you’re just trying to get in for the ride. If the thing wouldn’t exist without your support and you’re providing capital where it is scarce, that feels a lot more productive to me.


Science & Technology

Did someone say M2? 🤤 (Apple + TSMC)

The next generation of Mac processors designed by Apple entered mass production this month [at TSMC] [...] using the latest semiconductor production technology, known as 5-nanometer plus, or N5P [...]

Shipments of the new chipset -- tentatively known as the M2, after Apple's current M1 processor -- could begin as early as July for use in MacBooks that are scheduled to go on sale in the second half of this year (Source)

Ok, here’s my dream scenario: iMac Pros with 30” screens and Macbook Pros 13”/16” are unveiled at WWDC this summer, shipping in the fall.

They have the M2 rather than the rumored M1x (which would’ve been a M1 with more cores, RAM capacity, and maybe more thunderbolt lanes and I/O), based on the same core as the A15 that will ship in the next iPhone (will they skip “13” because some people are still superstitious? Can humanity finally get over that numerology BS please?!).

ARM’s Data-Center Cores, Neoverse V1 and N2

2020 has been an extremely successful year for Arm’s infrastructure and enterprise endeavours, as it was the year where we’ve seen fruition of the company’s “Neoverse” line of CPU microarchitectures hit the market in the form of Amazon’s new Graviton2 design as well as Ampere’s Altra server processor. [...]

Today, we’re pivoting towards the future and the new Neoverse V1 and Neoverse N2 generation of products. [...] this generation of Neoverse CPU microarchitectures differ themselves in that we’re talking about two quite different products, aimed at different goals and market segments. The Neoverse V1 represents a new line-up for Arm, with a CPU microarchitecture that is aiming itself for more HPC-like workloads and designs oriented towards such markets, while the Neoverse N2 is more of a straight-up successor to the Neoverse N1 and infrastructure and cloud deployments in the same way that the N1 sees itself today in products such as the Graviton or Altra processors.

Interesting that they’re segmenting and specializing their chips a bit to better serve different use cases:

the V1 focuses on maximised performance at lower efficiency, with features such as wider SIMD units (2x256b SVE), while the N2 continues the scale-out philosophy of having the best power-efficiency while still moving forward performance through generational IPC improvements.

The flexibility of the ARM ecosystem is one of its main strengths, so it makes sense that they would try to lean into that at their level of the pipeline, and then further downstream, others can use the lego blocks that they provide and further customize things for their specific needs.

The original Neoverse N1 as seen in the Graviton2 and Altra Q processors had been a derivative, or better said, a sibling microarchitecture, to the Cortex-A76, which had been employed in the 2019 generation of Cortex-A76 mobile SoCs such as the Snapdragon 855. [...]

the new generation V1 and N2 microarchitectures are related to newer designs in the Cortex-portfolio. The V1 is related to the Cortex-X1 which we’ve seen in this year’s new mobile SoCs such as the Snapdragon 888 or Exynos 2100. The Neoverse N2 on the other hand is related to an upcoming new Cortex-A microarchitecture which we expect to hear more about in the following few months.

From memory, the X1 seemed like a pretty impressive chip.

the V1 was designed at the same time as the Cortex-X1 by the same team at Arm’s Austin design centre

One thing that distinguishes the data-center variant from the mobile core is the extra SIMD capabilities (these “Single instruction, multiple data” units are useful for accelerating highly-parallel vector math).

The V1 core is expected to be about 50% faster than the N1 and The N2 is expected to see about a 40% boost over the N1, which is also a big deal, but these are ARM’s numbers, so we’ll have to wait and see what they can do in the real world.

The N2 also ARM’s first Armv9 core (Apple’s M1 and A14, for example, are using the Armv8 architecture). If you want more details, there’s plenty more geekery on Anandtech.


The Arts & History

‘Animals Of The World Map’

May not be ecologically accurate, but I think it looks cool. Source.

1

Gotta close the parenthesis even if it looks really weird, right?

2

Reminder that I’m using this emoji to signify “shower thoughts”.

3

Speaking of Jim, you should check out his thread on the ‘Thinker and the Solver’ if you haven’t seen it. Great stuff.