At the start of 2022, most of the US still had yet to experience the promise of 5G.

5G coverage from the major carriers extends nearly nationwide, but not many can access the connection. As of January 2022, T-Mobile customers had an active 5G connection 35.4% of the time, per a report from OpenSignal, while AT&T and Verizon customers had one 16.5% and 9.5% of the time, respectively.

But, but, but...Even beyond infrastructural issues, like limited availability, 5G faces another challenge: It still lacks a killer app, an essential and definitive application of the technology that could show consumers and businesses alike that 5G is worth embracing.

• For example, the advent of 4G technology enabled live video streaming, refined the accuracy of geolocation-based apps like Google Maps and Uber, and allowed online mobile gaming, all from the comfort of a cell phone.

“We’re all trying to figure out, you know, that killer app,” Brian Danfield, vice president of commercial and 5G strategic planning at Verizon, told Emerging Tech Brew. “No one knew these other apps were going to exist when we were launching LTE.”

With low latency and high speeds that 5G boosters have promised, a number of use cases could, in theory, be vastly improved upon in the future:

• Augmented reality and virtual reality from gaming companies like Sony, which could use 5G to support immersive and persistent universes for players to meet, socialize, and play with each other. This could also apply to work-based metaverses, like those proposed by Meta and Microsoft.
• Smart city and mobility applications with low margins of error could be improved, like robotaxi fleets, while smart city tech, like digital twins, could be updated and referenced in real time.
• Remote operations, like robotic surgery, or tele-operation of heavy machinery, could allow doctors to perform operations remotely.

Danfield said that while there’s no definitive killer app yet, “it’s not all vapor and, ‘We’re gonna.’” He pointed to 5G-enabled smart-home tech and faster home broadband as examples of potentially game-changing tech that is available today.

Big picture: “The important thing to recognize here is the 4G experience was smartphone applications looking for spectrum,” Tom Wheeler, a former FCC chairman, told Emerging Tech Brew. “The 5G experience, thus far, has been spectrum looking for applications. I think that they’re coming. There is not the kind of demand-pull economics there was in the 4G environment."

Click here to read the full story.—JM

On Wednesday, OpenAI went public with a big project: It overhauled GPT-3, its signature large language model, and introduced a new default tool—a set of language models called “InstructGPT.”

Quick recap: GPT-3, like other large language models, was created in part to generate human-like text in a convincing way. Researchers, technologists, and companies—from startups to Microsoft—have used it to generate summaries, change text style and tone, and more. But since the model’s 2020 debut, it has also been criticized for producing racist and sexist outputs, as well as other biased behaviors.

• When researchers asked GPT-3 to complete a sentence containing the word “Muslims,” it turned to violent language in over 60% of cases—choosing words like “bomb,” “murder,” and “terrorism.”

What’s new

Compared to GPT-3, the new InstructGPT models are A+ students, according to the AI research and deployment company. They’re better at following instructions in English, less inclined to produce misinformation, and at least slightly less likely to produce “toxic” results.

How it’s trained: OpenAI had 40 people rate GPT-3’s responses to a series of prompts, like, “Write a creative ad for the following product to run on Facebook,” MIT Technology Review reported. They downvoted nonsensical, violent, or clearly biased responses—and the well-rated ones were used to train InstructGPT via a reinforcement learning algorithm.

• The results: The majority of OpenAI’s data labelers preferred the new models’ responses to GPT-3’s, even though InstructGPT is 100 times smaller (1.3 billion parameters vs. GPT-3’s 175 billion). It’s one example of a better-trained language model performing better, in some cases, than a larger one.
• GPT-3 will still be available, though OpenAI will recommend users choose InstructGPT instead.

But, but, but: InstructGPT is far from infallible. At its core, like any natural language processing tool, it’s a large-scale pattern imitator. It can still makes simple mistakes, fail to follow certain instructions, and understand user prompts to be true even if those prompts contain misinformation. And it can still promote dangerous stereotypes—according to OpenAI, “InstructGPT shows small improvements in toxicity over GPT-3, but not bias.”

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

Coworking is a weekly segment where we spotlight Emerging Tech Brew readers who work with emerging technologies. Click here if you’d like a chance to be featured.

How would you describe your job to someone who doesn’t work in tech?

I use satellite images and manually collected information about forests and trees to track their growth and health without having to send people back out into the field. We try to predict forest biomass, disturbance, and potential from images, lidar, and radar.

What’s your favorite emerging tech project you’ve worked on?

Machine learning applied to remote sensing. There are so many cool things we’ll be able to do with the terabytes of satellite data monitoring the entire Earth. We can learn so much and improve the health of natural ecosystems with this knowledge. I truly can’t believe I get to work on something this incredibly interesting.

What emerging tech are you most optimistic about? Least? And why?

I’m most optimistic about battery and energy storage innovations and least optimistic about carbon air capture.

Both of these are related to humanity’s climate change response, but one seems way more likely to succeed. We have a lot of great tools available today to reduce carbon output, and carbon air capture seems very far away from prime time. I hope I’m wrong. But I’m very worried that we’ll design our climate-change response plans assuming it’s much more mature than it is.

One thing we can’t guess from your LinkedIn profile?

In addition to all my tech work, I am also an author. I’ve written two books: Platformed, about a disenchanted software engineer in near-future Silicon Valley, and Pretty Deadly, a dark retelling of Cinderella.

Click here to view on-site.—GD

Stat: In 2021, 68% of US electric vehicle buyers received some form of purchase incentive, like a federal tax rebate.

Quote: “Everything can be a computer. We’re just finding a way to make the hardware physics do what we want.”—Logan Wright, a physicist at Cornell University who co-led a study about building AI systems with everyday objects instead of microchips

Read: What if quantum computing is a bust?

Learn: Sign up for the Morning Brew Quantitative (MB/Q) accelerator program. Over seven weeks, you’ll learn how to use numbers to drive performance and how to analyze financial risk—without hating it the whole time.

All week: Another packed earnings slate. NXP, Google, AMD, Facebook, Qualcomm, Sony, Amazon, Snap, GM, and Unity all report this week.

Monday: Metro Connect, a conference on digital infrastructure, begins and runs until Wednesday. It will cover everything from planning for the $65 billion of infrastructure bill money earmarked for broadband to sustainability.

Wednesday: Groundhog Day.

In this week’s installment of our biweekly reader poll, we examine the dietary preferences of you all.

Kidding, sorta. Two weeks ago, we asked all of you to weigh in on whether you’d try cultivated (aka lab-grown, aka cellular, aka cultured) meat, given the chance. Here are the results, from the 718 of you who responded:

• Most of you responded positively: More than one-third (36%) of respondents said they’re very likely, and 28% said they’re likely to try it.
• As for the nay-sayers, 14.5% answered “very unlikely,” while 12% said they’re unlikely.
• ¯\_(ツ)_/¯ = the remaining 9%’s answer.

As for everyone else: This isn’t a perfect comparison, as the survey methodologies and respondent groups vary quite a bit, but a recent Piplsay survey found that 19% of US and 18% of UK consumers are “eager” to try lab-grown meat when available.

• And while around two-thirds of US and UK respondents are concerned about the environmental impact of meat production, nearly 80% say they would prefer traditional meat over other alternatives in the long run.

Anyway, for now, lab-growth meat is only legal to sell in Singapore. But China did just include it in its five-year agricultural plan for the first time, and some industry analysts think 2022 could be the year the US green-lights the products.

Click here to take this week’s reader poll, on autonomous vehicle safety.

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

• The global race for battery production is underway
• How the ‘world's first’ AI-managed ETF stacks up, almost five years later
• Ready to electrify your vehicle fleet? EV fleet tech providers want to help