• These animations are bringing history to life
  • Realistic “deepfake” videos just got scarier…
  • AT&T has finally admitted defeat…

Dear Reader,

Just like that…

Daily new COVID-19 cases began a rapid slide down to levels we haven’t seen since June of last year. Incredible news, right?

We haven’t reviewed these numbers in a while. It was intentional. I wanted to wait for a couple of months until we had a better picture of the bigger trend. And it was worth the wait.

But what was it? What was the significance of the peak in early January?

That peak was on January 8, as indicated in the graph above. It was the day immediately after the formal electoral college ballot count for the U.S. presidential election.

And as we can see in the graph, the number of COVID-19 cases has been on a slippery slope downward. It’s as if the fog of COVID-19 has suddenly been lifted.

But did anything really change? We’re still in the middle of a pandemic, aren’t we?

I’ll let readers draw their own conclusions. But I will share the dramatic numbers that demonstrate such a remarkable “improvement” in daily new COVID-19 cases. Since January 8, daily new cases have dropped an incredible 82%.

Wow. What’s going on? How is that possible? As we can see above, we did not see the same kinds of dramatic slopes after the first and second waves of the pandemic.

What makes this difficult to quantify is that there are a few different dynamics at play. Yes, the fact that vaccines began to be administered in the weeks prior certainly had some positive impact on the reduction of daily new cases.

That said, through January and February, the vaccine was only being distributed to a small subset of the population. Thus the impact would be relatively small.

The larger dynamic at play is the dramatic drop in daily COVID-19 tests. And almost no one is talking about it.

The peak was on January 15 at 2,326,802 daily tests. Yesterday, that number dropped to 1,349,529. That’s a decline of nearly 1 million COVID-19 tests a day.

And as we’ve discussed previously, the majority of the PCR-based COVID-19 tests produce false positives. This is when someone tests positive but is not carrying a live COVID-19 virus. These people are not infectious at all.

Should we be surprised then that daily new cases have dropped 82% when tests declined by almost 1 million a day? Absolutely not. It’s a perfectly logical and predictable effect.

But the dramatic decline in testing doesn’t explain the entire 82% decline in daily new COVID-19 cases. There is something much larger going on in the background with diagnostics testing that is almost entirely unknown to the general public.

And I look forward to sharing that with you tomorrow.

Now let’s turn to today’s insights…

Bringing historical figures to life…

Genealogy is something that has always interested me. Many years ago, I kicked off a study on my own family heritage and quickly discovered how difficult it was. There are some great resources available online, but what I discovered is that so many of the ancestral records are not readily available.

It is still very much a manual, time-consuming process. And that’s especially true if we want to verify and document each and every family connection. A few generations back is easy, but after that it quickly becomes difficult.

Ultimately, I came to the conclusion that I needed professional help – researchers that specialized in accessing and documenting my family tree.

My interest over the years is probably why a fascinating development at MyHeritage, an online genealogical company, caught my eye. The firm is using artificial intelligence (AI) and computer vision (CV) technology to bring historical figures – famous or otherwise – back to life.

MyHeritage built a platform that allows users to upload old photographs of individuals so they can have them digitally reanimated. These could be pictures of deceased family members or famous historical figures. Once uploaded, the AI generates a modern avatar of the person, complete with realistic facial expressions.

From there, users can create videos of the deceased person smiling, nodding, and even winking. These videos are realistic, and the technology will only improve. They allow us to envision deceased individuals as they might have been.

MyHeritage’s “Deep Nostalgia” Animation

Source: MyHeritage

This is amazing technology, and it’s not hard to imagine some interesting applications for it.

At the simplest level, this platform will allow those of us interested in our ancestry to see lifelike videos of deceased family members that we’ve never met. We can imagine what it would have been like to interact with our deceased relatives.

The technology could also be used to produce videos of historical figures engaged in action and perhaps giving speeches. That could help bring history to life for us. This could be an incredible resource for educational and entertainment purposes.

At the same time, this creates an interesting ethical dilemma.

Obviously, there’s no way to know if the digital representations we create using AI and CV are accurate. This becomes especially problematic if bad actors seek to intentionally portray certain historical figures in a negative light.

This is especially relevant as the incredibly toxic and intolerant cancel culture is so prevalent and empowered by social media platforms actively censoring and imposing their own bias on society at large.

So this is something we’ll need to keep an eye on going forward. MyHeritage will include a watermark on each video created through its platform to indicate that the video was artificially created. That will certainly help.

Still, we are in the very early stages with this technology. It’s going to quickly become prevalent, as will the discussion around how to ethically use it.

And that brings me to our next topic…

The stunning rise of deepfakes…

Longtime readers of The Bleeding Edge will be familiar with the rise of “deepfake” technology and its use to foment mistruths and falsehoods.

The technology is enabled by a form of AI called deep learning that can superimpose a face over that of another person.

But as we have discussed before, when these deepfake videos aren’t labeled, they are nearly impossible to detect by the untrained eye.

And what I have seen in the last few days indicates that there has been a breakthrough in the technology, which is both impressive and frightening.

Oddly, the technology can be seen in an explosion of Tom Cruise deepfakes on social media released in the last few days. Check this out:

Tom Cruise Deepfake

Source: YouTube

This is just one of several deepfake videos of what appears to be Tom Cruise that have surfaced recently. And these videos are unbelievable. They look and sound just like Tom Cruise, but they were generated by AI.

There are no tears or imperfections on the face. There is a lot of motion in the videos. The voice is exactly like that of Tom Cruise, and it just looks and feels like it is him. I’m amazed at how good these are. The only real tell is that Cruise appears much younger than he actually is today.

Other than that, it is impossible for someone watching to tell that the videos are fake.

Clearly, the deep learning AI technology behind deepfake videos has progressed tremendously since we last talked about it.

While these Tom Cruise deepfakes weren’t malicious, they still weren’t Tom Cruise. And Cruise certainly didn’t authorize their production.

It’s very easy to see how deepfakes could be used to defame public figures, destroy reputations, and scare or mislead the public into behaving irrationally. And that’s a frightening thought.

The worst part is that we don’t have a perfect solution to deepfakes.

A partial solution is to always check to make sure the videos we watch on social media were posted by verified accounts. At this point, we should assume that a video is fake if it is coming from an unverified source.

But the reality is that most people won’t think to do this. If they see a video that looks real, they will assume it is real.

So we’re in trouble here. This is going to exacerbate the problem we already have of knowing what information is and isn’t accurate.

My hope is that the technology community will develop a format for verifying video content to confirm real videos vs. deepfakes.

This is something that we desperately need to focus on.

AT&T is finally cutting the cord…

We will wrap up today with another incredible story of corporate strategy going wrong. Case studies like these are always interesting to learn from.

We talked before about Verizon’s bumbling foray into media with the acquisitions of dead internet companies America Online (AOL) and Yahoo back in 2015 and 2017, respectively.

Verizon spent nearly $9 billion on these two transactions… and it ended up taking over a 50% haircut after it had to write down $4.6 billion worth of its media business just 18 months after the Yahoo deal. Talk about a quick way to lose billions.

Well, AT&T wasn’t immune from catching the dumb acquisitions bug back then either.

AT&T acquired DirecTV for an incredible $49 billion in 2015. Yet it was already becoming clear that multichannel pay TV services were on the decline.

The trend of consumers “cutting the cord” and ditching cable and satellite TV for streaming services was already well underway. And DirecTV’s customer base was shrinking as a result.

AT&T followed this up by acquiring Time Warner for a whopping $85 billion in 2018, accumulating a mountain of debt in the process.

As I’m sure many readers know, these two acquisitions have been a disaster. Both DirecTV and Time Warner are dying slow deaths as more consumers cut the cord each year.

And AT&T finally admitted defeat. The company just announced that it is teaming up with private equity giant TPG to spin out DirecTV along with its own internet-based TV services, U-Verse and AT&T TV. These three TV services will become part of a new standalone company “creatively” named New DirecTV.

AT&T will retain 70% ownership of the new entity. And in exchange for a cash investment of $1.8 billion, TPG will own 30% of the new entity. The two hope to value the business at $16.25 billion – roughly one third of what AT&T bought DirecTV for in 2014.

Given that AT&T’s ownership stake will be worth less than $11.4 billion, the company is effectively writing down $37.6 billion on the DirecTV deal. Ouch.

And this begs the question – what about Time Warner?

Well, New DirecTV plans to borrow roughly $7.8 billion from banks to pay AT&T back for the assets. AT&T announced that it will pool that money with the $1.8 billion received from TPG. Then AT&T will use this $7.8 billion to pay down some of the debt it accumulated from the Time Warner acquisition.

No surprise there. Time Warner is still an albatross around AT&T’s neck three years later. The company is still trying to dig out of the hole created by its bad deals.

Fortunately, AT&T finally figured out that it needs to stick to its core business of operating wireless networks. That is what it should have been doing all along.

But it seems these massive legacy incumbents always need to learn their lesson the hard way. It’s fairly obvious, but companies like Verizon and AT&T are not media companies, nor should they try to emulate them.

The good news is that AT&T can focus more intensely on its 5G build-out now that its TV services are gone.

And hopefully the company will figure out how to off-load the Time Warner assets as well in order to focus entirely on its core competencies. That’s a strategy that I could get behind.


Jeff Brown
Editor, The Bleeding Edge

Like what you’re reading? Send your thoughts to [email protected].