In 2020, Amazon encouraged users to teach Alexa how to interpret certain requests.

Now, like a Montessori teacher, Amazon wants the AI system to figure some things out on its own.

This approach, which Amazon dubbed “self-learning,” is all about decreasing friction and making human-AI interaction feel more natural. In the last year, the company has begun adding a suite of such features to consumer devices worldwide.

• Self-learning boils down to advancing Alexa’s capacity for common-sense reasoning, which is one of AI’s most notorious obstacles.
• More widely known as “self-supervised learning,” it’s essentially a form of AI learning that’s, well…not supervised. Instead of learning from human-labeled data, the AI analyzes data on its own to draw conclusions.

Zoom out: Self-supervised learning is one reason why natural language processing—a subset of AI that underpins tools from Google Search to auto-complete—has advanced so much in recent years. Meta has used it for image training; Google has used it for medical-image classification; and Amazon has used it for automated speech recognition.

“Instead of a bunch of scientists taking data and trying to update the models and then roll them out, self-learning is something that works out in the field—there’s no human intervention, no human involvement,” Prem Natarajan, Alexa AI’s VP of natural understanding, told us. “It’s directly learning through its interactions with customers, whether it’s implicit customer behavior or otherwise.”

Keep reading to...learn...what self-learning looks like for Alexa.—HF

We don’t like to play favorites here at Emerging Tech Brew, but across July, we did. Last month, we traded hardware for wetware and zoomed in on biotech.

Here’s a quick speedrun through them:

Startups are deploying a thousand-year-old agriculture method to help with carbon removal. Stems, leaves, trunks, roots, flowers, and fruits are the original carbon-removal tech. Now, startups are catching on and trying to develop ways to optimize and accelerate naturally occurring carbon sequestration processes.

This startup wants to invent brand new forms of meat. From juicy ribeyes to sushi-grade salmon, cultivated-meat companies have grand plans to reproduce the most popular proteins in the world. But Vow Food is placing a different bet: It wants to bioengineer brand new, better forms of meat.

How synthetic biology startups plan to revolutionize fertilizer production. A group of startups, including Pivot Bio and Joyn Bio, are leveraging the promise of synthetic biology to try and produce affordable, environmentally friendly fertilizer at scale. The push for these bio-based fertilizers is in its early stages, but experts told us the concept holds promise in the long term.

From infant poop to trance music, here’s a play-by-play of a CRISPR experiment. Much has been made of the technology’s revolutionary potential, but today we’re lifting the veil on what CRISPR looks like in a hands-on setting, via CRISPR co-inventor Jennifer Doudna’s lab.

View all the stories in one place.—DM

The iCar. The Apple car. The Tim Cookmobile.

Whatever you’ve taken to calling the iPhone-maker’s mythical driving machine, one thing is true: You can’t buy one. Despite years of work—and rumors—the project is still closely guarded and no product has been revealed to the public.

Last week, we asked all of you to weigh in on whether Apple will ever bring a vehicle to market. The results are in, courtesy of nearly 3,000 Emerging Tech Brew readers…

• Just under half (47%) of you think Apple will eventually unveil a purchasable car.
• Just over half (53%) think the opposite: Apple won’t ever be an automaker.

The company has been working on the project, dubbed Project Titan, for eight years, and in recent weeks it has been reported that its efforts have been plagued by everything from managerial challenges, like shifting goals and leadership, to technical issues with automated driving software.

• Apple has reportedly spent $1+ billion per year on R&D for the project, and several senior execs—including Tim Cook—are either skeptical or lukewarm about it.

On the other hand: Project Titan continues to move and shake—it recently poached a long-time Lamborghini exec, for instance.

But perhaps more importantly: Apple is extremely profitable and also tends to sit on tens of billions of dollars in cash reserves. It has money to burn on moon-shot ideas—like a self-driving electric vehicle.—DM

Stat: Oh, you thought the US’s ~$52 billion in semiconductor subsidies was big? South Korea (~$260 billion) and China ($150 billion) would like a word.

Quote: “The industry folks don’t really care about the data…They just grab whatever’s easiest. People think it’s all the same and you just need more of it.”—Maarten Sap, an expert in natural language processing, to the Washington Post

Read: The awful, inexorable march of deepfake porn.

Plants in your pants. Through the sustainable power of plants, DUER makes the world’s most comfortable pants—keeping your legs and the environment happy. Get 15% off with code MorningBrew.*

^{*This is sponsored advertising content.}

• Toyota-Panasonic, via a joint venture on battery production, inked an agreement to buy lithium from ioneer’s forthcoming Nevada mine.
• Nikola bought Romeo Power, a battery supplier, for $144 million.
• The US opened bidding for another 5G spectrum auction. Elsewhere in selling air: In India, Jio won the right to $11 billion worth of 5G spectrum.
• NuScale, a startup that makes small, modular nuclear reactors, received a certification for its design from the Nuclear Regulatory Commission.
• Innoviz said VW ordered $4 billion worth of lidar sensors from it.
• China’s virtual influencer industry is booming—but often grueling for the humans that support it.

Snap poll: Have you ever followed or engaged with a virtual influencer (e.g., Lil Miquela)?

Yes
No

Catch up on the top Emerging Tech Brew stories from the past few editions:

• Sodium-based batteries could solve the lithium crunch
• Space investments fell back to earth in Q2
• From infant poop to trance music, here’s a play-by-play of a CRISPR experiment