413: Constellation Software, Nvidia + Stable Diffusion, Microsoft, Skunk Works, Grimes AI, STEM Shortage, Google Bard, and The Last of Us
"Snatching defeat from the jaws of mediocrity."
When negotiating, dont aim for a bigger piece of the pie; aim to create a bigger pie. –Kevin Kelly
🚮🗑️👨👩👦👦 Last Saturday, my family and I went to the park.
Not to play, but to pick up trash.
My wife had seen a call for volunteers to help clean up the neighborhood. We thought it was a good opportunity to talk with the kids about civics — ie. things don’t just happen by magic, someone has to do them, and through our actions we can make the place where we live better or worse, the tragedy of the commons, etc…
As we were picking up empty water bottles and plastic bags, my oldest son said to me:
“That sucks! There’s trash everywhere on the day that we’re doing this, but on other days there’s almost none.”
The mind of a child!
I shared with him the concept of the Baader-Meinhof phenomenon (without using the name). It highlights how, when we aren't actively seeking something or paying attention to something, we often fail to notice it—even when it's right in front of us.
The amount of trash today is no different from other days, but we simply haven't been looking for it. It's like when you buy a yellow car. Suddenly, you start spotting yellow cars everywhere…
🚨🎙️1️⃣ I was a guest on the podcast of friend-of-the-show and supporter Chris Powers (💚 🥃):
I really enjoyed it!
Chris is a great interviewer and we ended up talking about everything from Constellation Software to the impact of AI on various industries to the internet-era creator economy to the business model of David Senra’s Founders Podcast and a lot more!
🚨🎙️2️⃣ In case you missed it, here’s my podcast conversation with Kevin Kelly.
It’s not something I ever expected to happen, but it was great!
The part about family traditions and rites-of-passage was particularly impactful. I’m trying to implement my own version of these ideas in my family. 👨👩👦👦
✍️🤔 I thought I may have coined a phrase:
Snatching defeat from the jaws of mediocrity.But after looking it up on Google, I found 2 previous hits for it. I guess 2 hits out of billions of sentences on the web isn’t too bad!
💚 🥃 🐇 This is the free edition of Liberty’s Highlights with 16,900+ subscribers.
You can get 1-2 extra editions/week full of juicy stuff + access to the private Discord 🗣🗣🗣 community by becoming a paid supporter (it’s quick & easy).
Paid post since last week 🐇:
If you click the link above, you can see the intro for free and there’s a link to get a risk-free 7-day free trial of the paid version. You can only gain from this!
🏦 💰 Liberty Capital 💳 💴
👨🏻🔧🧰"Kelly’s rule was: never put an engineer more than fifty feet from the assembly area" -Ben Rich, Skunk Works 🦨📕
I’ve recommended this book many times over the years, and I’ll do it again!
Stealth technology development:
If we made this shape into a full-size tactical fighter, what would be its equivalent radar signature… as big as what—a Piper Cub, a T-38 trainer… what?”
Denys shook his head vigorously. “Ben, understand, we are talking about a major, major, big-time revolution here. We are talking infinitesimal.”
“Well,” I persisted, “what does that mean? On a radar screen it would appear as a… what? As big as a condor, an eagle, an owl, a what?”
“Ben,” he replied with a loud guffaw, “try as big as an eagle’s eyeball." [...]
The model was mounted atop the pylon and then rotated in front of the radar beam. Well, two very funny things happened. The first day we placed our model on the pole, the pole registered many times brighter than the model. The technicians had a fit. They had thought their poles were invisible, but the trouble was nobody had ever built a model that was so low in radar signature to show them how wrong they really were. [...]
SR-78 Blackbird:
MORE THAN THIRTY YEARS after its first flight, the Blackbird’s records will not soon be surpassed: New York to London in one hour and fifty-five minutes; London to L.A. in three hours and forty-seven minutes; L.A. to Washington in sixty-four minutes.It’s a lot of fun and full of business insights about high-performance and innovative cultures.
🎤🤖🎶 Grimes offers 50/50 split for AI-generated songs using her voice (turning it into a kind of API 🤔)
Whenever something new comes on the scene, some people will try to stop it while others will embrace it.
In the wake of the viral song with AI-generated vocals by Drake and The Weeknd, Grimes (aka Claire Boucher) wrote:
I'll split 50% royalties on any successful AI generated song that uses my voice. Same deal as I would with any artist i collab with. Feel free to use my voice without penalty. I have no label and no legal bindings.
We're making a program that should simulate my voice well but we could also upload stems and samples for ppl to train their own
Talking with my team: if u register music with us we can collect & pay out royalties direct to anyone who uses A.I. Grimes vocals using smart contracts!? The future rly is now! this is so cool.When I saw this, my next thought was that this could backfire on her. Artists won’t want a flood of crappy, low-quality, or extreme lyrics that they would never say associated with their brand.
It looks like Grimes also had second thoughts and had to add some caveats to her offer:
Ok hate this part but we may do copyright takedowns ONLY for rly rly toxic lyrics w grimes voice: imo you'd rly have to push it for me to wanna take smthn down but I guess plz don't be *the worst*. as in, try not to exit the current Overton window of lyrical content w regards to sex/violence. Like no baby murder songs plz.
I think I'm Streisand effecting this now but I don't wanna have to issue a takedown and be a hypocrite later. ***That's the only rule. Rly don't like to do a rule but don't wanna be responsible for a Nazi anthem unless it's somehow in jest a la producers I guess.
- wud prefer avoiding political stuff but If it's a small meme with ur friends we prob won't penalize that. Probably just if smthn is viral and anti abortion or smthn like that.
Rly rly don't like adding rules so I apologize but this is the only thingIt’ll be interesting to see what comes out of this, and if other artists are encouraged to follow suit, or scared away from the idea of essentially turning their voice into a kind of API that can be purchased by third parties through revenue-sharing. 🤔
It’s also sure to clash with record labels who have been trying to prevent generative models from being trained on the music to which they own the rights…
✨ Constellation Software’s ✨ VMS Venture Fund is planting seeds 🌱
Back in Edition #208 (wow, time flies!), I wrote about Constellation launching a $200m VMS venture fund with multiple goals over the next few years. Here’s how they described it:
The Fund will provide financing for start-up and rapidly growing vertical market software businesses, most of which will have been either incubated or identified by a sponsoring Constellation business unit.
While Constellation regularly invests in dozens of small initiatives and will continue to do so, VMS Ventures will invest in larger initiatives and take a different approach:
The Fund will only invest in businesses that have the potential to become standalone business units within Constellation.
The managers and employees of the Fund's Investees will be significant shareholders in their businesses.
The Fund will not back start-ups that have to rely upon a Constellation business unit for on-going sales, marketing, customer support, development, etc.
Many of the Investees' staff will be hired from outside of Constellation.
The Fund's employees will be compensated based upon the Fund's results.
Investing only $200 million over a three-to-five-year period, Constellation does not expect the Fund to add meaningfully to Constellation’s organic revenue growth rate. The Fund's objective is to develop and refine organic growth processes which can eventually be rolled out more broadly by Constellation's operating groups.
Fast-forward to today, and there are now at least three businesses in the fund:
TIC Systems: A management system for the Testing, Inspection and Certification Industry
Roibos: A marketplace to empower hoteliers and travel operators
Fero: Deliveries Logistics for the Rental Industry
I don’t know if there are others that haven’t been announced or will be kept in stealth mode, but it’s nice to see activity there, and it’ll a fun experiment to track over time.
h/t friend-of-the-show C.J.
🛂 🧳🧑🔬🇺🇸 Arizona needs more semiconductors engineers & STEM immigration is the low-hanging fruit 🐜
If you can get the smartest people from around the world to move to your country and use their skills to improve your companies, research labs, and economy, why wouldn’t you?
While it may seem obvious when put that way, the situation apparently proves to be not-so-simple for the US and it’s creating obstacles in their effort to revive semiconductor manufacturing within the country:
As semiconductor factories go up across Arizona, there's a shortage of engineers qualified to do the job. [...]
According to the US Bureau of Labor Statistics, from 2016-2024 there will be a shortfall of about six million engineers. Kozicki says many of these roles are highly specific.
"We're short of talented engineers in the space," he said. [...]
According to an ASU representative, more than 4,200 engineering students are set to graduate this spring and will then be able to find work in the microelectronics industry, from electrical and chemical to computer science and manufacturing technologies. (Source)
Why not staple green cards and work visas to any STEM degree earned by foreign students in the US, and do the same for immigrants with STEM skills trying to enter the country?
It would probably be bad for Canada if the US implemented this, but it would certainly be very good for the US…
h/t Alec Stapp
🤖🎓 Google Bard gets a Computer Science Degree (not literally!)
Starting now, Bard can help with programming and software development tasks, including code generation, debugging and code explanation. We’re launching these capabilities in more than 20 programming languages including C++, Go, Java, Javascript, Python and Typescript.
In addition to generating code, Bard can help explain code snippets for you. This is particularly helpful if you’re learning about programming for the first time, or if you need some additional support to understand what a block of code might output.
Bard can also help you debug code, even code that Bard wrote. If Bard gives you an error message or code that doesn't do what you intended, just tell Bard “this code didn’t work, please fix it,” and Bard can help you debug.
GitHub Copilot and ChatGPT are getting some more competition!
Once this is in the hands of coders for a while, I’ll be curious to see if Bard is better at certain things and Copilot at others, or if one of the two is better than the other pretty much across the board.
And if it’s the former, will developers devise methods to automatically ask the same questions to multiple coding LLMs and then cherry-pick the best code generated? 🤔
There’s so much leverage to good code, I wouldn’t be surprised to see that.
🕵️♂️ 💾 Deep Dive: Microsoft
From the stuff of legend in its early history to today’s revamped behemoth, my friend MBI (💎🐕) does a great job of looking under the hood of Microsoft:
📄 Microsoft: The Quintessential Technology Company (sub required 🔒)
Some of my favorite parts of the deep dive are the graphs and tables, especially the ones showing Microsoft’s revenue growth and EBIT margins in the 1980-90s and Azure vs AWS growth.
🧪🔬 Liberty Labs 🧬 🔭
✨ 🔭 I don’t even know how to describe this one…
Kind of a tour-de-force romp through a lot of science, physics, and philosophy of science, but in a very entertaining way. I hope you like it. 👩🔬 ⚗️🧬
🤖 🎥 Nvidia turns StableDiffusion into Text-to-Video model
There’s a cool new paper (with lots of video examples!) by Nvidia about a new approach to doing text-to-video efficiently:
Latent Diffusion Models (LDMs) enable high-quality image synthesis while avoiding excessive compute demands by training a diffusion model in a compressed lower-dimensional latent space. […]
We first pre-train an LDM on images only; then, we turn the image generator into a video generator by introducing a temporal dimension to the latent space diffusion model and fine-tuning on encoded image sequences, i.e., videos. Similarly, we temporally align diffusion model upsamplers, turning them into temporally consistent video super resolution models. […]
Furthermore, our approach can easily leverage off-the-shelf pre-trained image LDMs, as we only need to train a temporal alignment model in that case. Doing so, we turn the publicly available, state-of-the-art text-to-image LDM Stable Diffusion into an efficient and expressive text-to-video model with resolution up to 1280 x 2048.
The generated videos aren’t flawless — in many of them, you can see the model hallucinate all kinds of things, especially in the background. But it still shows just how fast things are progressing.
It wasn’t that long ago that the only videos you could make with generative AI looked like hallucinogenic experiences… 😵💫
For more video examples, visit this page, which also includes the full paper if you want to dig all the way to the bottom layer.
CATL *may* have a battery breakthrough coming to mass production 🔋🔋🔋🔋🪫
I’ve been following battery tech for about 20 years, which is why my pulse so rarely quickens when I read about a new "battery breakthrough". They are a dime a dozen, and almost nothing makes it out of the lab and to mass production.
But this one sounds more legit than most:
the world’s largest battery manufacturer CATL has announced a new “condensed” battery with 500 Wh/kg which it says will go into mass production this year. [...]
CATL chief scientist Wu Kai says the condensed battery integrates a range of innovative technologies, including the ultra-high energy density cathode materials, innovative anode materials, separators, and manufacturing processes, offering excellent charge and discharge performance as well as good safety performance.
For comparison, Tesla’s 4680 cells are at 272-296 Wh/kg.
But I'll truly believe it when it's shipping in volume and tested by third parties...
🎨 🎭 Liberty Studio 👩🎨 🎥
🎮 How The Last of Us was Remade for the PlayStation 5 🍄
The original version of the game was released in 2013 for the PlayStation 3.
A remastered edition followed in 2014 for the PS4. It improved game controls and enhanced the graphics a bit (better rendering distance, higher framerate, etc), but it remained largely faithful to the original version.
However, the 2022 remake for the PS5 is something else, as the video above shows!
For more Last of US, check out my podcast discussion and analysis of the HBO TV version with two friends-of-the-show:
We are in the Creator Era 👩🎨 (What is scarce now? 🤔)
Ted Gioia shares some stats on the sheer volume of creative content is being released:
A hundred thousand songs are uploaded daily to streaming platforms.
In the last year 1.7 million books were self-published.
2,500 videos are uploaded to YouTube each minute.
There are now 3 million podcasts—and 30 million podcast episodes were released last year.
86% of youngsters want to grow up to become influencers, and contribute to these impressive numbers
With this loooooooong tail of content, is it any wonder that the scarce thing is now discovery? Curation? Taste?
It’s *impossible* for anyone to find the needles in all those haystacks, but some people are naturally very curious, work hard at it, and over time they develop a skill for it.
They should feel very thankful for living at a time when that talent is becoming increasingly valuable!