If you’re not already familiar with lithium, 2022 will probably be the year that changes.

This metal is third on the periodic table, the lightest solid element on earth, and an essential material in the lithium-ion batteries used for electric vehicles and energy-storage systems as the world transitions away from fossil fuels.

While there is plenty of lithium on the planet, it isn’t being extracted and refined quickly enough to keep up with the rapidly growing demand for batteries.

• By 2030, there’s projected to be a lithium deficit between 455,000 and 1.7 million metric tons each year.

Simon Moores, CEO of Benchmark Mineral Intelligence, told Emerging Tech Brew the EV industry is already facing a shortage of battery-grade lithium. Read on for highlights from our conversation, and click here to read the full thing.

On automaker preparedness

“It’s funny, I think OEMs are quite a unique beast in this sense. The OEMs only plan to build battery plants. They don’t think about the raw materials. And that’s because they don’t come from that world. The traditional automotive supply chain is built to serve them, and they’re used to having everything available.

But the difference now with the supply chain for electric vehicles—it’s being built from scratch and doesn’t exist at the scale they need, so they don’t think about the raw materials, and we call this the great EV raw material disconnect. Two years ago, they weren’t even thinking about batteries. They are thinking about batteries now, which is good. But they’re still not thinking about raw materials. And that’s a huge risk to anyone that’s really making EVs beyond 2025 or 2026, which is nearly everyone.”

On the long-term outlook

“Lithium-ion batteries, in general, they’re becoming cheaper. I know the price is going up now, but they’re becoming cheaper in the trend, lower cost to make. They’re at the same time becoming better. Lithium-ion batteries are now probably three times better than they were in 2015. The third thing is they’re becoming abundant through these gigafactories.

So all of those three things combined is this trend that’s underpinning not just EVs, but really the energy-storage revolution and everything we’re trying to do for climate change. I think people only now are starting to realize how important the humble lithium-ion battery is, really.

You’ve got to ride out the volatility. Long term, it will be fine. The price will be fine. There’s no geological shortage. Lithium-ion battery powered EVs will dominate the world.”

Click here to read the full conversation with Simon Moores.—GD

On Wednesday, DeepMind—the Alphabet–owned AI research company—published not one, not two, but three research papers about this year’s most hotly debated AI tool: large language models.

• As a reminder, these models are typically trained on an enormous amount of text and underpin services from Big Tech companies like Google to startups like Grammarly.

The highlights

Bigger isn’t always better: “Gopher” is the name of DeepMind’s new, 280 billion-parameter language model. (Generally, in the NLP world, more parameters = higher performance metrics. For reference, OpenAI’s GPT-3 has 175 billion parameters, while some newer models created by Google and Alibaba range from one to 10 trillion parameters.) With this paper, DeepMind wanted to test when scaling up a model’s size makes it perform better—and when that’s not the case.

• The results: In DeepMind’s tests of different-sized models, relatively bigger models were better at comprehending written text, checking facts, and IDing toxic language. But their larger size didn’t necessarily make them any better at tasks involving common-sense or logic.
• DeepMind also found that no matter a model’s size, it had the tendency to reflect human biases and stereotypes, repeat itself, and confidently spread misinformation.

Smaller isn’t necessarily worse: DeepMind also introduced a new model architecture dubbed RETRO, or Retrieval-Enhanced Transformer. It’s all about training large language models more efficiently (read: faster and cheaper). Think of RETRO as a David and Goliath situation, in that its performance can compete with neural networks 25 times larger, according to DeepMind.

• RETRO’s internet retrieval database is 2 trillion words collected from the web, books, news, Wikipedia, and GitHub, and it reportedly helps with AI explainability—researchers can see the text passages it used to generate a prediction, rather than its decisions being an inexplicable “black box.”

The tech’s risks aren’t going anywhere: In the last paper, DeepMind released a classification of language-model-related risks, categorized into six areas: 1) discrimination, exclusion, and toxicity 2) information hazards 3) misinformation harms 4) malicious uses 5) human-computer interaction harms and 6) automation, access, and environmental harms.

• Risk mitigation is one of the biggest concerns, according to the paper—for instance, figuring out how to address language models learning biases that could harm people.

Click here to view this story on-site.—HF

Last week, we asked all of you if you’d consider harkening back to the days of the Motorola Razr and buy a foldable smartphone.

First, some context: The form-factor has drawn a lot of hype since Samsung unveiled its first prototype in 2018—likely in part because of the nostalgic itch it scratches—but so far, people haven’t really bought ’em. That said, a recent report showed that foldable-smartphone sales grew 480% year over year last quarter to 2.6 million units shipped, per DSCC.

• The research group forecasts total foldable-smartphone sales to hit 17.5 million in 2022.
• There are ~1.5 billion smartphones sold each year.

Back to the survey: Just over half (52%) of you said you wouldn’t consider buying a foldable smartphone, while 48% of you said you would. Apple was the runaway favorite, with 57% of respondents saying they’d be most likely to buy a foldable phone from the tech behemoth.

• Samsung, the actual foldable-smartphone market share leader, was no. 2 with 33% of respondents voting for it.

Click here to take this week’s poll on autonomous driving.—DM

Stat: Motional, the Hyundai-Aptiv autonomous driving joint-venture, released an AV training dataset with nearly five years’ worth of “average” driving footage. It claims this is the biggest public dataset for AV projects.

Quote: “It is essential that investors receive the information they need, when they need it, without misleading hype.”—SEC Chairman Gary Gensler re: SPACs

Read: Inside a DIY, crewed rocket project.

Your stubble’s in trouble. There’s simply no match for the Braun Series 9 pro, the world’s most efficient electric shaver. Seven-day beard? Sheared. This baby has a ProLift trimmer that gently lifts and clips. Powerful. Self-cleaning. Waterproof. 10% off with code MBREW.*

^{*This is sponsored advertising content.}

All week: Apple could become the first company worth $3 trillion this week. As of send time, it had a market cap of ~$2.97 trillion.

Tuesday: The FCC is holding an open commission meeting, and satellite broadband is on the docket. Specifically, the agency is considering changes to its rules around low-earth-orbit satellites, with the hope of sparking greater competition among satellite broadband providers.

Wednesday: Rivian reports its first public company earnings. It filed its S-1 about two months ago, so we may not get a ton of new information this week, but we’re still paying attention to the unproven but extremely buzzed-about EV startup.

A little over two centuries (and three days) ago, Ada Lovelace was born. She’s considered to be the first-ever computer programmer—there’s even a 1980s-era programming language named after her.

Obviously, Lovelace wasn’t bashing if-then statements into a keyboard and scrolling through Stack Overflow in the mid-1840s.

• Instead, her contribution to programming was theoretical—she wrote out instructions for how inventor Charles Babbage’s then also (mostly) theoretical computer could be programmed to perform certain calculations.

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

• Roblox’s metaverse is already here, and it’s wildly popular
• Timnit Gebru debuts DAIR, an AI ethics group that will center marginalized communities
• Batteries, the key component of EVs, keep getting cheaper. But that rate is slowing