- Is autonomous driving years away?
- $100 genome sequencing? It’s closer than many think…
- AI-generated people…
Dear Reader,
They called it “rhinoceros-sized.” A “mega-gourd.”
Weighing in at 2,560 pounds, “Maverick” was a sight to behold at this year’s Safeway World Championship Pumpkin Weigh-Off in Half Moon Bay, California.
Maverick the Mega-Gourd
Source: weighoff.miramaevents.com
Maverick’s largess is impressive. And it got me thinking…
How, and why, is it possible to grow a gourd that large? What is it that makes these pumpkins so different than something like a cucumber? And why is it that these gargantuan gourds seem to be getting larger year after year?
After all, the winner of the championship in 1974 weighed in at just 132 pounds. And it took almost 30 years to breach the 1,000-pound mark in 2001. But now, every winner since 2017 has been well above the 2,000-pound mark.
What sets these pumpkins apart?
The answer lies within their cellular makeup. The period of cell division in a cucumber lasts about four or five days. In pumpkins, it’s about 20 days. After the cell division stops, then cell expansion kicks in.
Cucumbers will expand for around 20 days, whereas a pumpkin’s cells will continue to expand for about 50 to 60 days. And it’s not just the cellular structure that makes these pumpkins so beastly.
Farmers will water the best pumpkins with up to 150 gallons to 200 gallons of water a day in order to maximize the growth potential of the giant gourds. After all, water makes up about 90% of a pumpkin’s weight.
But it raises a question… What’s the limit? These distorted and ghoulish gourds certainly can’t grow as big as a house, can they?
No, they can’t. When a farmer “pumps” too much water into the pumpkins and they become too heavy, the burden results in an ugly split, or fracture, and ultimately it collapses under its own weight.
This process reminds me of the world’s central banks, which have been printing money and pumping their economies with a seemingly endless and monstrous level of debt for decades.
Largely, they’ve gotten away with it. There have been a few bumps and bruises along the way, but this monetary trick has lasted longer than most thought it would. And the treats have been far too tempting to stop.
And here we are on All Hallows’ Eve, where we can sense the horrors of what’s to come.
There’s a limit to how long the U.S. government can trick us before its current monetary and fiscal policy collapses under its own weight. And there’s not much time left. In fact, something massive will break – most likely the bond markets – within the next six months.
As scary as that seems, and it is, what follows will be extraordinary. Easing will find a new lease on life, liquidity will flood the markets, and there will be a massive swing back to investment and growth.
We’ll be smart to steer clear of the ghouls and goblins for now and prepare to stay out late once the liquidity returns.
Until then, I hope you all have a Happy Halloween.
The autonomous driving industry is sorting itself out…
In a surprise announcement, one of the most prominent companies in the autonomous driving space – Argo AI – announced that it’s shutting down operations.
This came as a shock to the industry.
Argo AI was one of the five most well-funded private companies working on self-driving technology. Regular readers may remember we showed footage of Argo’s tech in action just a few months ago.
For the sake of newer readers, Ford acquired Argo AI for $1 billion back in 2017. That was back when self-driving tech was both nascent and largely experimental. I commended Ford at the time for being so forward-looking. It was an unusual move for such a legacy company.
Then, in 2020, Volkswagen (VW) came in and invested $2.6 billion into Argo AI. This signaled that Ford and VW would partner on developing the technology.
This made sense. Both companies could share in the research and development costs and then potentially adopt the technology into their own respective vehicles.
But neither company is a tech company, and their competencies don’t include managing large, complex software projects that utilize artificial intelligence (AI). This is where things fell apart.
The mistake Ford made was in holding Argo AI within its own corporate structure. I know the intention was good. But obviously the environment didn’t foster the same innovation we expect to see at bleeding-edge start-ups.
Technologists working at the bleeding edge need operational autonomy to move fast, iterate, and improve the tech as quickly as possible. But most legacy corporations are very uncomfortable with that operating model.
The opposite is true as well. Innovators are often stifled by the kinds of operational processes that bureaucratic companies tend to put in place. So now Argo AI is shutting it all down. And Ford and VW will write off billions of dollars in losses.
What they should have done is followed General Motors’ (GM) model.
If we remember, GM invested heavily in Cruise for its self-driving technology. But GM only partially acquired the company. And it allowed Cruise to remain largely independent. Not only that, but GM continued to allow a wide range of investors into Cruise.
We can now see the stark difference. Cruise has autonomous ride-hailing services up and running. And Argo AI is shuttering its doors.
Some have already said that this is the beginning of the end of the industry. That’s completely foolish.
Ironically, this is a sign that the industry is maturing. In any emerging technology, winners will emerge, and the weakest players will fall out of the game. Different teams experiment with different approaches. And consumers eventually benefit from this creative destruction. This is normal.
Whenever a new industry shows promise, billions of dollars’ worth of investment capital floods into it. That money is spread between a number of different companies at first. But only a few of those companies survive and thrive in the end. Argo AI isn’t one of them.
Interestingly, Ford’s CEO Jim Farley made a boneheaded comment when making the Argo announcement:
We’re optimistic about a future for L4 ADAS, but profitable, fully autonomous vehicles at scale are a long way off…
A long way off? As we know, autonomous vehicles are on the road right now.
Cruise has self-driving cars running in San Francisco, California. Waymo has cars in Phoenix, San Francisco, and now Los Angeles. Then there’s Motional, which is running ride-hailing pilots with both Lyft and Uber. And of course, we have self-driving Teslas all over the road these days.
So, Farley’s comments were way off. They were nothing but posturing in a poor attempt to spin his decision to shut down Argo AI.
However, the CEO did later clarify that Ford is still interested in self-driving tech. It just plans to partner with other companies for it.
That’s a smarter approach. Ford doesn’t need to spend several billions more to develop the technology in-house. It will be able to license the technology from one of the other major players.
Either way, we’re starting to get some clarity in the industry. I’m excited for the wave of new self-driving deployments coming next year.
Precision medicine is about to go mainstream…
Pacific Biosciences, known as PacBio in the industry, just announced a new DNA sequencer. It’s called Onso.
This news is flying under the radar right now… but it shouldn’t be. Onso will help propel precision medicine into the mainstream.
That’s because Onso is PacBio’s first short-read sequencer. Here it is:
PacBio’s Onso DNA Sequencer
Source: PacBio
The significance here is that short-read machines can sequence DNA faster and cheaper than long-read sequencers.
And that’s critical to getting costs down to the point where regular people can afford to have their DNA sequenced – and eventually even covered by insurance.
Up to this point, Illumina has been the only company with a mass-market short-read sequencer. We just talked about its NovaSeq X product a few weeks ago.
NovaSeq can sequence up to 20,000 human genomes a year. And the price point is about $200 per sequencing. As we discussed, that puts sequencing in a place where many health insurance companies may cover it.
Well, the launch of Onso will only drive sequencing costs lower. That’s the beauty of competition.
I believe we’ll get down to $100 per sequencing in the very near future. That’s when all insurance companies will offer coverage.
And that would bring precision medicine to the masses.
If we think about it – only a fraction of the population has had their genome sequenced. That number is likely less than 2%.
So Onso will open up a massive untapped market. That’s good for both PacBio and Illumina. And it’s great for consumers as well.
Once it’s covered by insurance companies, genome sequencing would become a doctor’s first recommendation. Then, if the patient consents, a geneticist can analyze their genome and pinpoint specific genetic mutations that may be the root cause of the problem.
At that point, it’s just a matter of prescribing a therapy designed specifically to correct those mutations.
It could be a specific drug that’s determined to be most effective for that specific genetic condition. Or, in time, it could be a genetic editing therapy to “correct” an unwanted mutation. The solution can even be as simple as some changes to fitness and nutrition. No more guesswork required.
Ultimately, this dynamic will save the industry tens, maybe even hundreds of billions of dollars on unnecessary tests and treatments that simply wouldn’t be effective for individual patients.
What’s more, this will vastly improve healthcare outcomes for everybody.
The generative AI that creates humans…
A project called Generated Photos just announced that its generative AI created 100,000 full-body images of humans… all from scratch.
These full-body images aren’t modified from human photographs. These are all original creations from the AI.
Here’s a look at some of them:
Full-Body Generated Images
Source: Generated.Photos
Notice how life-like these images are? Everything from the facial expressions… to the posture… to the clothing. Most would never know these are AI-generated images.
Generated Photos originally started developing headshots for use in media applications but has since upgraded its AI to create full-body images.
As a result, the project now has a feature where it will create images of humans upon demand. Users can feed the AI prompts like height, weight, skin tone, hair color, eye color, and body type, and the AI will generate a human image accordingly.
As quirky as this sounds, there are wild implications here.
I instantly think about marketing and advertising. Companies spend billions of dollars every year on actors for their commercials and marketing messages.
Well, with this generative AI technology, those same companies could simply create their own human model from scratch. And if we imagine taking this one step further, this technology can be paired with deepfake technology. That would enable the generated human to have its own voice, mannerisms, and gait.
So companies can create their own spokespersons with this. And they can dress them in whatever style or brand of clothing they want. All while shaving billions from their marketing budgets.
As strange as this may sound, this is a powerful technology that could get adopted very quickly.
And it’s a heads up for us… Pretty soon, we won’t know if the people we see in videos and commercials are real or AI-generated. And in a few years, it’s going to get hard to tell who, or what, we’re speaking to.
Regards,
Jeff Brown
Editor, The Bleeding Edge