169: Twilio CEO Jeff Lawson, Nvidia Reveals CGI Jensen Spoke During Keynote, Do You Have 2 Remote Jobs?, Reddit Going Tiktok, Intel's Time to Build, and McCartney & Rubin
Vacation Edition #1
Investing & Business
‘Ask Your Developer’ Book Highlights (Jeff Lawson, Twilio co-Founder & CEO)
Jeff Lawson on his first internship in 1997:
On my first day, my manager welcomed me to the company and outlined my task: I needed to convert the files from the old format to the new format. He gave me my company-issued computer, descriptions of the old and new file formats, and showed me to my desk.
I wrote a program in C that read in the old files and spit out the data in the new format. At around lunchtime, I returned to my manager’s desk to let him know I’d finished my first task. He was stunned. That, apparently, had been my job for the whole summer. He figured that I would spend the days copy and pasting data from one file to another—for thousands of files—and that this would make for a great internship. I asked what I should do with the rest of the day, and he said he’d get back to me. Turns out, he didn’t have anything for me to do—not just for the remainder of that day, but for the entire summer.
Lawson on pitching new products:
When you pitch a new product idea to potential customers, one of two things happens, especially when they’re an acquaintance. If they genuinely like the idea and it seems to solve a pain point in their life, they ask you questions. Does it do this? Does it do that? They’re trying to match your solution to their problem. That’s a good sign. If they don’t see your idea as solving some pain point they have, the conversation goes differently. They try to be polite, and say: “Oh, that sounds nice . . .” as their voice trails off. After a few moments of awkward silence, they’ll change the subject.
On innovation:
This is how innovation works: experimentation is the prerequisite to innovation. The more quickly and cheaply you can run experiments, the faster you’ll eventually find something that works. So I kept looking for ideas. [...]
The superpower of software to me was how quickly you could take an idea and get it in front of customers. Once customers could play with your idea, they’d give you feedback, and tell you what was good and what was bad. That would inform the next thing you built, and so forth. You could release a new version every day if you wanted to. [...]
having seen big companies like Amazon as well as tiny startups like StubHub, I’ve learned that the key ingredient to unlocking innovation is cultural more than anything else, and that culture starts at the top.
On real devs vs cliché pop-culture devs:
there’s an often-overlooked reason software has so many parallels to the music and film industries:
Code is creative.
Pop culture depicts developers as science and math nerds. [...]
Writing software is more similar to making music or writing a book than it is to doing math or science. [...]
Yet most companies don’t understand this, and don’t create an environment where developers can exercise this creative muscle—and everybody loses. Developers don’t do their best work, and they dream of quitting to start their own company.
companies lose because some of their best talent goes underutilized. Customers lose because the products are the sad results of a dispassionate software factory. [...]
too often I witness executives leave Silicon Valley, taking away the wrong lessons. It’s easy to latch on to the superficial stuff that is evident when you walk around tech offices, like free food, or letting people wear T-shirts and hoodies, and bring dogs to the office. [...]
Casual observers of innovation often overlook the thing that matters most, which is that developers are given responsibility and freedom. Not just freedom in working hours or what to wear, but freedom in terms of creativity.
He talks about the need to give devs *problems* and then let them come up with solutions, rather than giving them *solutions* and asking them to just implement them without much creative input. It seems so obvious when you say it like that, yet it’s so rare out in the wider world…
As an aside, I now realize I should’ve bought a Kindle years ago. If you’re on the fence, I recommend giving it a try — it has increased how much book-reading I’ve been able to do lately.
E-Ink and a single-purpose device are really nice things to have in many situations…
Do you have two remote jobs at the same time yet?
A small, dedicated group of white-collar workers, in industries from tech to banking to insurance, say they have found a way to double their pay: Work two full-time remote jobs, don’t tell anyone and, for the most part, don’t do too much work, either.
Alone in their home offices, they toggle between two laptops. They play “Tetris” with their calendars, trying to dodge endless meetings. Sometimes they log on to two meetings at once. They use paid time off—in some cases, unlimited—to juggle the occasional big project or ramp up at a new gig. (Source)
Considering how many in-person jobs are mostly about busywork, pointless meetings, and spending hours commuting back and forth, with only a few hours a week of deep work, this isn’t too surprising.
But I’m impressed at how organized this sub-culture has gotten, with the website Overemployed basically providing guidance on how to successfully pull this off.
Even Reddit is Getting more TikTok-like + Raising $700m @ $10bn Valuation
most iOS users should have a button on their app directly to the right of the search bar — when tapped, it will show a stream of videos in a TikTok-like configuration. When presented with a video, (which shows the poster who uploaded it and the subreddit it’s from), users can upvote or downvote, comment, gift an award or share it. Like TikTok, users can swipe up to see another video, feeding content from subreddits the user is subscribed to, as well as related ones. (Source)
🤔
I don’t know how I feel about this. I suppose the mechanism in itself isn’t worrying, but Reddit raising $700m at a $10bn valuation does make me a bit more nervous… Why do they need all this money? How do they intend to give a multi-X return to their investors and justify that valuation? Could it push them to do stupid things?
Reminds me a bit of when Twitter thought it was expected to perform like Facebook, and did all kinds of hare-brained schemes to grow into that rather than recognize what it actually is and focus on making that better.
Intel’s Gelsinger is Subsidy-Shopping
Intel’s new CEO has met with US officials to sell them the company’s new strategy:
Mr. Gelsinger has struck a common note in those meetings, according to people familiar with the talks and documents, namely that Intel has big plans to build more chip factories that also can help fix an overconcentration of chip-making in Asia driven by lucrative incentives there. All it will take to level the playing field is a few billion dollars in subsidies. [...]
To get back on top, Mr. Gelsinger is elbowing to be the first in line for U.S. government grants of as much as $3 billion per factory project—potentially more with a presidential signoff—from a program making its way through Congress. [...]
The U.S. accounts for about 12% of global chip-production capacity, Europe 9%. Mr. Gelsinger has said he would like to see the U.S. at 30% of global capacity within the next 10 years and Europe at 20%. [...]
Soon after, Mr. Gelsinger headed to Germany and to Brussels, where he presented European Commissioner Thierry Breton with a plan suggesting a modern manufacturing site with two factory modules would cost €17 billion, equivalent to roughly $20 billion (Source)
Science & Technology
COVID19 Rapid-Testing MIA in N-A
Why hasn’t the US and Canada gotten, as far as I know, any significant amount of COVID rapid-testing yet?
Seems a no-brainer to have scaled this up and be using these paper tests everywhere, mailing them to everyone for free, etc.
Even if they are not nearly as sensitive and accurate as PCR tests, the idea is that if you do lots of them, you still catch way more cases than you would without them, and often earlier. And my understanding is most of the errors are false negatives (because viral load is too low to trigger a positive in the early days of infection), not false positives, so if you do them at close enough intervals, you eventually get a positive and catch more cases than it would seem from just the % specs (ie. You shouldn’t think of it as comparing one rapid-test to 1 PCR test, but rather to compare 10 or 50 or 500 rapid-tests to every PCR test).
This talks a bit about Germany’s use of rapid testing (as just one part of the testing system — I’m not saying it should be the only thing, just that it would no doubt make things better).
How much blood & treasure has been lost because some regulatory body has let the perfect be the enemy of the good (which is probably what happened)? How many people would still be alive today if we had blanketed North-America with these tests months ago? ¯\_(ツ)_/¯
Nvidia Expands Omniverse Software Platform
NVIDIA Omniverse makes it possible for designers, artists and reviewers to work together in real time across leading software applications in a shared virtual world from anywhere. [...]
Blender, the world’s leading open-source 3D animation tool, will now have Universal Scene Description (USD) support, enabling artists to access Omniverse production pipelines. Adobe is collaborating with NVIDIA on a Substance 3D plugin that will bring Substance Material support to Omniverse, unlocking new material editing capabilities for Omniverse and Substance 3D users. (Source)
This is very cool. Improving tools has a lot of leverage, I’m looking forward to what creators can make with this.
Also, this short video showing 3D character sculpting in a VR environment is pretty neat:
Big Reveal: Computer-Generated Jensen Did 14 Secs of Keynote Kitchen Speech
Nvidia says part of its April keynote was led by a virtual replica of CEO Jensen Huang, created by scanning Huang and then training an AI to mimic his gestures
You can see how they did it 23 minutes in this video.
h/t Friend-of-the-show Adam Brass
The Arts & History
McCartney 3,2,1 — Did you do your homework?
Finished watching the 6 episodes of this docu-series of, basically, Paul McCartney and Rick Rubin sitting in front of a studio mixing board and listening to master tapes of various Beatles and McCartney songs and talking about them, their back-stories, anecdotes, isolating various tracks/instruments that we may not hear as well in the final mix, etc.
I gotta say, I really enjoyed it.
One of my main takeaways is the quality of Paul’s bass playing, and how he was an accidental bass player (nobody in the band wanted to do it when they lost their original bass player) and yet his varied playing style was obviously very influential to many generations of musicians. He helped make the instrument cooler and more exciting than it was at the time, at least for that genre of music.
Even more impressive, he often came in the studio not even knowing what he was going to play, not having anything written down, and just figuring it out on that day. So many classic bass lines, with all kinds of cool melodies and counter-points and genre-blending choices, were just figured out on the spot.
Legendary.
I also really enjoyed looking at how Rubin works, he he gets familiar with a song and tries to uncover the original nugget of the idea by stripping away all the extra arrangements, and then build things back up to understand how things were made to gel together.
If you did follow along, I hope you enjoyed it too! Let me know your thoughts in the comments on the site or by replying to this email.
Interview: Nick Kokonas (Alinea Restaurant, Tock app)
I’ll put this one here, even though there’s also some business and tech talk in there.
For more context on Kokonas, check out editions #57 and 58 where I talk a bit more about the Chef’s Table episode that features him and chef Grant Achatz from Alinea.
Here’s a recent interview that Jason Calacanis did with Nick that I enjoyed:
Hi,
What is your view on this?
https://semianalysis.com/tesla-dojo-ai-super-computer-unique-packaging-and-chip-design-allow-an-order-magnitude-advantage-over-competing-ai-hardware/
Rolling it all up, cost equivalent versus Nvidia GPU, Tesla claims they can achieve 4x the performance, 1.3x higher performance per watt, and 5x smaller footprint. Tesla has a TCO advantage that is nearly an order magnitude better than an Nvidia AI solution. If their claims are true, Tesla has 1 upped everyone in the AI hardware and software field. I’m skeptical, but this is also a hardware geek’s wet dream