• Yet another incredible AI from DeepMind…
  • Coinbase urges customers to switch to USDC…
  • Uber’s robotaxis are coming to Las Vegas…

Dear Reader,

Today will be the last “normal” issue of The Bleeding Edge for the year. 

Though “normal” isn’t a word I’d use to describe The Bleeding Edge and all of the incredible topics, developments, and breakthroughs that we revel in throughout the year.

But fear not! We’ll still publish The Bleeding Edge daily for the next two weeks. I’ve been working feverishly in the background with my team the last few weeks to put together my thoughts on the most important topics of this year and what they mean as we look ahead into 2023.

We refer to this as my annual Prediction Series. We look back at 2022 and see what I got right from my previous predictions, and I make some pretty exciting calls about what’s to come in 2023. I assure you this isn’t dull editorial – nor is it repurposed. 

The truth is, I don’t get much sleep at all during the first couple weeks of December as I work to put additional research together for the holiday break.

This allows me to give my team, and myself, a break at the end of the year after 12 months of incredibly hard work. And I’d like for you, my dear subscribers, to have something interesting to ponder over a nice cup of coffee – or a stiff drink – as we mentally prepare for what’s to come in the new year.

2022 was a tough one. I’ve never seen such a confluence of bad government policies and geopolitics happening at the same time. 

And it’s frustrating to me because I know that it could have all been avoided. It weighs heavily on me because I know what’s happening will negatively impact the next generation.

Tough markets like we saw this year aren’t “downtime” for me and my team. We end up working even harder to stay on top of developments and deeply understand what’s going on – and we look for ways to find investment opportunities that have limited downside risk while still maintaining upside potential. 

We worked harder than ever this year and had the opportunity to explore different investment strategies for the market that we found ourselves in.

However, all of the insanity of this year stands in stark contrast to the remarkable number of breakthroughs that we saw. It almost didn’t matter what the sector was, be it genetic editing, life sciences, robotics, artificial intelligence (AI), machine learning, quantum computing, nuclear fusion, autonomous driving technology, etc. 

We saw incredible breakthroughs in all of these areas that will lead to further commercialization and progress in 2023.

And while we’re still in for a rough first half of the year in 2023, I believe we’re going to see a strong recovery in the markets that will likely start as soon as early summer. Institutional capital will start to flow back into growth stocks and investment levels will remain high, driving an even larger number of breakthroughs in 2023.

I want to thank all of you for sticking around this year as we worked together through a tough environment. I’m grateful to have you as readers, and I always appreciate the high levels of engagement and feedback that we receive week in and week out. 

I’d also like to extend a deep thank you to my team that worked tirelessly this year in a tough market and continued to partner with me to produce fantastic research and editorial for the Brownstone Research universe. I simply couldn’t do it all without you.

While painful and frustrating, it was still a fun year. And I hope that all of you were able to still find some enjoyment here with me in The Bleeding Edge. I’d like to wish all of you a wonderful holiday season with friends and family, and I look forward to spending more time with you in the new year.

We’ll publish our usual mailbag issue tomorrow and, of course, we’ll monitor all model portfolio positions throughout the holiday and will send out alerts to our subscribers if anything actionable pops up.

Best wishes for a wonderful holiday season.

DeepMind’s latest release is a sign of what’s to come…

Yesterday, we had a look at DeepMind’s release of a computer programming AI called AlphaCode. Well, as if that weren’t enough, the division just revealed something called Dramatron.

Dramatron is built on a similar large language model as AlphaCode. But instead of writing software code, Dramatron is designed to write scripts for movies or theatre.

DeepMind trained the AI on how to generate plots and characters, dialogue between different actors, and even location descriptions. It generates text in the same way that a scriptwriter does. The output can be used to produce plays, full-feature films, and even a television series.

And it’s very much an iterative process.

The team at DeepMind worked with professional scriptwriters to develop Dramatron. What they found is that the AI is very formulaic… it’s not perfect on the first attempt, but it does add a lot of value to the process.

Often, the scriptwriters would get a lot of ideas from Dramatron’s initial output. Then they would modify the ideas a little bit and feed them back to the AI… and it would spit out even more ideas to improve the script.

I typically talk about AI from the perspective of it being used to augment human labor. This is a perfect example. Scriptwriters can use Dramatron to get ideas and produce full-length scripts much faster than they otherwise could. This doesn’t replace human efforts; it augments and makes our work process more efficient.

And what I most love about this is that it’s a sign of what’s to come.

With Dramatron, DeepMind demonstrated how we can take these large language models and optimize them for a very specific task. This could be done for any application. And it will be.

For example, I think customer service is an area that’s ripe for this kind of thing. A corporation could train an AI like this on its specific products and processes… and that AI could then interact with individual customers in a knowledgeable manner. 

Literally, the specificity for training an AI will be customized down to individual companies and optimized for very specific tasks.

So when I see something like Dramatron, I can envision how easy it would be to replicate this technology for thousands of different applications. We’ll see a lot of that happening in 2023.

The blockchain industry is self-regulating…

Coinbase just made a big announcement.

The giant cryptocurrency exchange is recommending that anyone holding U.S. dollar stablecoin Tether (USDT) switches to U.S. Dollar Coin (USDC). And Coinbase is even offering to make this switch for its customers at no cost. That’s how serious it is about this.

I’ve been sharing my concerns about Tether this year in The Bleeding Edge. In fact, just last month, I warned that it could be the next to collapse.

That’s because Tether doesn’t maintain a one-to-one backing for USDT. Instead, it only holds 82% of its assets in cash and cash equivalents.

This leaves 18% of USDTs potentially “backed” with risk assets. But Tether hasn’t disclosed exactly what those assets are.

Meanwhile, USDC is fully backed one-to-one with cash and cash equivalents. There’s no leverage or risk assets at all.

So the fact that Coinbase is launching a zero-fee service to convert USDT to USDC on behalf of its customers means that it must be concerned about USDT’s structure and underlying assets. Keep in mind that Coinbase will incur some costs by doing this… so it must feel strongly about USDT’s risk.

To me, this is one of the biggest players in the blockchain industry working in a self-regulating manner. That’s healthy.

And we should also note that Coinbase is helping facilitate this migration out of USDT in an orderly fashion. This helps ensure that there’s no moment of collapse where everybody rushes to liquidate their USDT at the same time.

This is a great move. While Coinbase is incurring a cost to do this, it will benefit the entire industry and hopefully help to avoid another collapse. I think we’ve had more than enough excitement this year.

Additionally, this bodes well for my prediction that USDC will overtake Tether as the world’s No. 1 U.S. dollar stablecoin.

Tether currently has a market cap of just under $66 billion. USDC’s market cap is just over $45 billion. So, the gap’s already narrow. Given the current direction, I suspect USDC will overtake USDT before next year is out.

Uber’s surprise announcement…

We’ll wrap up today with a surprise move from Uber.

Last month, we talked about how the ride-hailing giant has a 10-year deal in place with Motional to launch autonomous robotaxi services. If we remember, Motional is the joint venture between auto tech company Aptiv and South Korean conglomerate Hyundai.

At the time, we talked about how we would find out which cities Uber is targeting for its self-driving services in 2023. As it turns out, we didn’t have to wait that long.

Uber just announced that its first robotaxi service will launch in Las Vegas, Nevada. This makes perfect sense. Lyft already has its own autonomous driving service running there as well in partnership with Motional.

And like Lyft, Uber’s customers will get to choose whether they want a robotaxi or a regular driver. When people open the app, it will present that choice and the pricing differential between them.

At first, Uber wants to have a safety driver in the front seat of its robotaxis. Then it will eventually remove the safety driver as it gets comfortable with the service that Motional provides.

Now that the two largest ride-hailing services in the U.S. have launched their autonomous riding-hailing services, it’s all about scaling now to other metropolitan areas. 

With each successful deployment of the technology, the industry will be building a track record that will be useful for gaining approvals in new metropolitan areas.

2023 is the year in which these services go mainstream in many cities across the country.


Jeff Brown
Editor, The Bleeding Edge