I can’t put this off any longer, so here are the tech predictions I made a year ago for 2018. We have to see how well or poorly I did before we can move on to my predictions for 2019 and beyond. These old predictions have been edited for length, but not to avoid embarrassment.

I try to never avoid embarrassment.

One thing I’ve noticed over the years is that my predictions get longer and longer (this column, alone, is 4329 words — my second longest, ever) as they have drifted from new products to explaining new strategies. This sometimes works against the prediction since it is often easier to claim success if your goal is vague, but I see it more as a tribute to my readers. Many of you have been with me for decades and the very fact that we are both still here has as much to do with the work as with its results. How the future fits together is just as important as where it is heading.

My first prediction a year ago: The rise of 5G networking will lead to a crash in broadcast TV. All the mobile carriers will begin rolling-out 5G wireless networking in 2018, though only about 22 percent of the country will be covered by the end of this year. 5G is interesting because it will bring a huge increase in wireless bandwidth that we don’t really need… So what will all this 5G bandwidth be used for? I predict it will replace legacy broadband providers like telco DSL and cable ISPs. The telcos will feel this effect first. Every DSL vendor is ultimately doomed. There’s no fighting this trend so there will be a mad rush to 5G by these companies. Look for some of them to try to dump their wireline services entirely.

Okay, this clearly didn’t happen, though both network and cable TV total viewership in the USA have peaked, with total subscriber numbers declining in 2018. But do you have 5G? Neither do I. The prediction still stands, but it has been pushed back by the slow 5G roll-out. So I was wrong, but it’s only a matter of time before this happens exactly as predicted. My family and I were homeless for months after the fire, staying with friends, in AirBnBs, and in our 1978 GMC Royale RV. Throughout those months we used our T-Mobile unlimited 4G Internet for EVERYTHING — VoIP, Internet, Netflix, Hulu, Amazon Prime, YouTube TV. All it took was a USB Channel 12 LTE modem, a MIMO antenna stuck to a window with suction cups, and a $40 WiFi router. And now that we are finally back in a house, we haven’t changed at all our mix of programming. It may be the golden age of television, but TV is dying. Still, my over-optimism about timing made me zero for one.

My second prediction a year ago: The end of Windows supremacy. This is NOT a prediction that MacOS or Linux will take over from Windows. It’s more complex than that. What I am seeing is that Windows is becoming less and less important to Microsoft and as Microsoft’s focus changes so will our focus as consumers of personal computing. It’s not surprising that Microsoft is changing because desktop PC sales are down and Redmond can’t get much money anymore for Windows upgrades… Under Satya Nadella, Windows Phone is gone and Microsoft’s concentration is on (in this order): 1) Azure (Microsoft’s public cloud); 2) Azure services like storage and — to some extent — Office 365; 3) Microsoft Office, and: 4) Windows. That’s Windows going from first to fourth and Microsoft Office going from second to third. That’s huge.

And it happened exactly as I said. One for two.

My third prediction a year ago (sorry, this one is long): 2018 foreign profit repatriation (part of Trump’s tax reform) is a $591.8 BILLION taxpayer ripoff. Multinational U.S. companies have, since 2005, squirreled away about $2.5 TRILLION in profits overseas because U.S. tax law allowed those profits to go untaxed until they are returned to the USA. Understand that this $2.5 trillion is more than just the profit made on overseas business: many companies changed their ways of doing business to divert what would have been U.S. domestic profits, sending them overseas for parking to await a business-friendly U.S. Administration that would cut them a sweetheart deal to bring back all that moolah. The amount of U.S. corporate income tax that went unpaid during this 12 year period was 35 percent of $2.5 trillion or about $875 BILLION. That’s taxes of $73 billion annually for 12 years that went unpaid — about five percent of the federal budget for those 12 years.

Admittedly U.S. tax law was out of step with most of the rest of the world and the new law changed much of that, cutting the corporate rate to a more competitive 21 percent (from 35) and ending U.S. taxation of foreign profits entirely from 2018-on (foreign taxes still apply). But it is also clear that the companies keeping profits overseas won big in the repatriation deal.

The new tax law says companies must bring those profits home but instead of paying the 35 percent tax they’d been working so hard to avoid or even the new 21 percent corporate rate, those profits  will be taxed at a special 15.5 percent rate, saving the companies about $500 BILLION they would have paid under the old rules. That’s the $500 billion President Tump says will be converted mainly into new jobs, pay raises, and aggressive R&D. Yeah, right.

This is a $591.8 billion taxpayer ripoff, rather than just a $500 billion ripoff, because the law allows companies to pay those back taxes over EIGHT YEARS. They may owe $375 billion, but will only have to pay this year $30 billion, leaving $345 billion in essentially free money to play with — money that wouldn’t otherwise be available. Beyond the simple fact that you and I could never get a deal like this if we had gone 12 years without paying our taxes, economists are pretty much in agreement that the majority of the $2.2043 trillion in net repatriated profits will be used for share buy-backs, propping-up both earnings and stock prices. The present combined market cap of all U.S. public companies is about $30 trillion with repatriating companies likely to buy back at least $1.5 trillion (five percent) of that in the near future.

Again, what’s wrong with that? What’s wrong is we have already gone eight years without a recession and all these foreign profits going into stock buy-backs will create a weird effect once the inevitable recession is upon us, probably in 2019. Sales will drop, the economy will contract (that’s the definition of a recession) but earnings and stock prices will remain stable or even rise as CEOs and CFOs pit buy-backs against the recession, their hope being to somehow bridge the earnings gap and get to the next period of expansion without a market drop. American CEOs and CFOs do such things because their average tenure is just four years, all of which is spent trying to ensure their own comfortable retirements. With that golden parachute always an average of only two years away, why not try something — anything — to keep share prices up? And so they will. And maybe — just maybe — it will work. But more likely it won’t work. The next recession will start, markets will at first appear to defy gravity through buy-backs, but eventually and inevitably what went up will come down… hard. It will be like 2008 all over again, just in time for the next Presidential election…

Yes, the multinationals got their windfalls. No they didn’t roll the money into new jobs, raises, factories and R&D: that was simply a lie. And they are absolutely using share buy-backs to extend the apparent expansion. I was (and continue to be) right on this one, which will continue in 2019. Two for three.

My fourth prediction a year ago: Bitcoin stays crazy until traders learn it is not a currency. 2017 was a wild ride for cryptocurrencies and for Bitcoin in particular, rising in price at one point above $19,000 only to drop back to a bit over half of that number now (January, 2018). But which number is correct? If only the market can tell for sure — and these numbers are coming straight from the market, remember — what the heck does it all mean? It means Bitcoin isn’t a currency at all but traders are pretending that it is. 2018 will see investors finally figure this out.

The problem with Cryptocurrencies, and the reason why their prices (and implied underlying value) are so volatile, is because the basis of that value isn’t the same as the basis for a real currency. It’s not backed by gold. It’s not backed by “the full faith and credit of the United States of America.” It’s not backed by, well, anything. In that sense of innate value cryptocurrencies are worthless. But this is not to say that Bitcoin has no value. It’s just that the value is misunderstood. And here’s probably the most important point I am trying to make here: as long as the value of Bitcoin is misunderstood there will be hucksters taking advantage of suckers, stealing their crypto lunch money.

EVERYTHING in the cryptocurrency world is about the exchange rate, which is to say nothing is about the actual cryptocurrency itself. You see there is nothing holding the price of any legitimate cryptocurrency up or down except supply and demand and both of those are constantly changing. So the smart trader knows never to hold Bitcoins — or at least not to hold them for very long — because what went up will inevitably go down.

This is what makes Bitcoin a horrible investment but a wonderful trade, because there is money to be made in that inevitable volatility. Not only will what goes up come down again: it will also go back up again if you wait long enough. This is all because Bitcoin and similar currencies aren’t currencies at all but financial instruments. They are tools just like options or derivatives. And like good tools, they perform a specific function very well — reliably transferring value between parties without those parties having to meet or even know each other. But once that transfer is complete, then the true value of the Bitcoin has effectively been used and is therefore subject to change, which again explains the volatility.

The value of Bitcoin goes up with demand on the part of those who would use it to transfer value. Some Russian oligarch, for example, has heard a rumor that Putin is after every Ruble in his piggybank, so the oligarch exchanges those Rubles for Bitcoins as he boards his private jet in Moscow, then exchanges Bitcoins for U.S. dollars before he lands in some safer country. He doesn’t care much about the exchange rate because the oligarch only intends to hold the Bitcoins for at most a few hours. But turning 10 billion Rubles into Bitcoins will inevitably drive up the Bitcoin spot price. This is key — the spot price.

Now let’s say we have that day buying Bitcoins in addition to the Russian oligarch a Chinese industrialist, a Saudi prince, and a Nigerian con man, each with cops pounding on the door and $1 billion to move ASAP. With all that demand the Bitcoin price goes higher and higher and none of these parties actually gives a damn because over the course of their trades the price only goes up a bit. But to Bitcoin traders and (shudder) Bitcoin investors — this rise in price looks like a rise in value so of course they jump-in driving the price higher still. What happens, though, when the big trades are all finished and the Bitcoins reconverted to real currencies? The price of Bitcoin then drops because its utility as a financial instrument is not so urgently needed. There is nothing bad or good about this price change which is perhaps most analogous to breathing. In-out, up-down, that’s just how cryptocurrencies operate. Bitcoin is worth more than the others only because it has been mined longer and there are more of them in circulation, meaning greater liquidity to support larger transactions.

Until everyone comes to understand these underlying truths there will be huge Bitcoin price volatility with fortunes made and lost every day. But once we all get onboard, Bitcoin promises to become the little trading engine that could. That’s because for every time we know Bitcoin will go down we know that it will inevitably go back up again.

If you notice Bitcoin is down, buy it. When you’ve made as much profit as you need, sell it, then wait for Bitcoin to inevitably go back down again. Those who have a very high risk tolerance or think they can time the market will hold on longer and be more likely to lose their shirts. It’s a much better Bitcoin trading strategy to not be greedy, living instead on the crumbs of oligarchs and conmen.This is an understanding that I predict will broadly emerge in 2018.

But did it actually happen as I predicted? I think so. The price of Bitcoins has trended downward for the last year but within that there has also been a fair amount of volatility. The overall downward trend represents the market starting to figure out that Bitcoins priced in the thousands of dollars is nonsense. That the price remains as high as it is comes down to the fact that not everyone has this figured out and the traders are really enjoying sticking it to the rubes. The volatility within the downward trend is the real Bitcoin going up and down as big transactions are made. Eventually Bitcoin will settle on a price and stay near there, just riding up and down on those demand waves. We’re not there yet but it is coming. So I claim this one was correct. Three for four.

My fourth prediction a year ago: The H-1B visa problem will NOT go away. Immigration reform will have little actual effect on H-1B visa abuse… Immigration reform is coming, we’re told, and the Trump campaign said a lot about H-1Bs, especially about eliminating the lottery system that is used to allocate those 65,000 yearly slots. About half of those slots have gone to foreign outsourcing companies, especially from India. Once the H-1B reform legislation appears I think we can expect the lottery to go away in favor of a system based strictly on worker qualifications and the dire need to fill the position, which sounds egalitarian and terrific, eh? Probably not. You see the point of this H-1B reform will be mainly to get rid of the Indian outsourcers so their H-1B slots can be used by American companies hiring directly. It’s for this reason that the Indian companies are hiring U.S. citizens as fast as they can. And the U.S. companies are preparing, too, by practicing all those techniques that will allow them to appear to seek qualified domestic candidates and yet not find them. Toward that end, for example, I’m hearing that one big company that rhymes with IBM is starting to use recruiters whose native language is not English. This is not to say that non-native speakers can’t learn wonderful English or be consummate H.R. professionals, but the track record of big companies that rhyme with IBM is not good in this area. For all their billions in profits (and billions more in repatriated profits) these companies, which include biggies like Apple and Google, just can’t seem to bring themselves to pay market rates for labor if they can wriggle out of doing so. The Trump Administration sure isn’t going to make them do it, either. And for these reasons I predict that the H-1B visa program may change in 2018 but its problems will remain pretty much the same.

Well, what happened? Almost nothing. We didn’t get immigration reform, we got troops sent to the border to protect us from marauding women and children. We got a long discussion about walls and who would pay for them. We got some administrative changes that made life harder for some folks and easier for others, but the H-1B system remains pretty much as it was a year ago. As I predicted, the H-1B problem did not go away. Four for five.

My sixth prediction a year ago: AI comes of age, this time asking the questions, too

Paul Saffo says that communication technologies historically take 30 years or more to find their true purpose. Just look at how the Internet today is different than it was back in 1988. I am beginning to think this idea applies also to new computing technologies like artificial intelligence (AI). We’re reading a lot lately about AI and I think 2018 is the year when AI becomes recognized for its much deeper purpose of asking questions, not just finding answers.

Some older readers may remember the AI bubble of the mid-1980s. Sand Hill Road venture capitalists invested (and lost) about $1 billion in AI startups that were generally touted as expert systems. Alas, it didn’t work for two reasons: 1) figuring-out how experts make decisions was way harder than the AI researchers expected, and; 2) even if you could fully explain the decision-making process it required a LOT more computing power than originally expected. Circa 1985 it probably was cheaper to hire a doctor than to run a program to replace one.

But now, approximately 30 years later, AI has come back to life. Part of this is simply Moore’s Law. One million dollars worth of 1985 computational power costs less than a buck today, making those software experts way cheaper to run. A second reason for AI’s resurgence is the availability of huge online data sets. Over the past 30 years nearly all the information that formerly resided on paper was reduced to electrons making true machine learning possible. This can’t be over-emphasized. Anyone my age remembers when early search engines indexed thousands of web pages, not billions. Highly technical data generally wasn’t available online in any volume but now it is. The third major reason for AI’s resurgence is today we don’t even try to pick some doctor’s brain to build an expert system, instead empirically deriving skills directly from the data. Taking this one step further, we are moving to a system where we don’t even start with a question, just the data, allowing cloud-based systems to find what’s learnable from the data. That’s allowing AI to come up with its own questions and it’s the emerging trend on which I am trying to focus this prediction.

We’re getting to the point where AI should begin to organically suggest approaches that will help us improve medical outcomes. The big low-buck solutions like immunizations have for the most part already happened. Future gains will have to come incrementally, often one gene at a time. But that’s just where AI should shine, mining medical gems from all that data in our smart phones, fitness trackers, and home DNA tests.

Thirty years into AI, it’s time to start seeing significant rewards from this approach in many fields, not just medicine.

But did it happen as predicted? Is AI coming-up with its own questions? It is starting to and part of the reason why is simply a lack of data scientists, which I didn’t anticipate. Just to be sure about this, I passed the question to Kai-Fu Lee, who is ex-Apple, Silicon Graphics, Microsoft, Google, and quite the AI pioneer in his own right. Now a VC based in China, Kai-Fu published last year a popular book about Artificial Intelligence that prompted a column you can read here.  And below you’ll find his opinion on whether I was correct about prediction five:

I think it would be fair to say that “Yes it did begin to happen”.  Today, using AI to help on drug discovery is one of the most popular areas of research as well as VC-funded work.  Also, AI diagnosis on lung cancer, skin cancer, by reading MRI or CT have shown that they can do better than average doctors.

These promising early results and influx of money, of course, do not guarantee solutions.  Solutions may still take a long time.  But given your question is “did it begin to happen”, I think the answer is “yes”.

Kai-Fu

Five for six.

My prediction #7 for 2018: There will appear the first Alexa virus. This one was courtesy of my son Fallon, then 11, who is a pretty good strategic thinker. Fallon’s idea for a computer virus is that it should be possible to create otherwise benign Alexa skills that, when used together, can make trouble. Think about it. There are presently more than 15,000 Alexa skills that have been officially approved by Amazon and are available for download. These skills do everything from launching programs to gathering data to setting reminders. Though relatively simple, each is still a cloud app that can connect tens of millions of Echo products to Amazon Web Services (AWS). Each Alexa skill is tested by Amazon before being approved, but are they tested together? They don’t appear to be.

One skill, for example, could open a communication session while another could gather audio or video data for spying. One skill could take control of the Echo while another could put the resulting bot to terrible use in a local network or on the Internet as a whole. I’m sure you can imagine any number of clever combinations. Remember the combinations don’t have to be operating on the same Echo device to function cooperatively. This makes them even harder to detect.

Did it happen? Were Amazon Echo devices and similar hacked in 2018? Well there is a so-called Alexa Virus loose in the world, though that apparently has nothing to do with Alexa devices. But a group of Chinese hackers demonstrated last summer their ability to turn your Echo into a listening device at the DefCon security conference in Las Vegas. So yes, Alexa devices have been hacked.

Six for seven.

My prediction #8 for 2018: Apple will buy one or more of the many startups helping shift desktop computing loads to the cloud. Cupertino will compete with Amazon, Google, and Microsoft offering virtual cloud PCs.

Just as with prediction one, this is wrong but only because Apple is moving slower than I expected. With Cupertino’s ever-stronger orientation on services, don’t be surprised to see this still happen… sooner rather than later. But for now, it’s still wrong.

Six for eight.

My prediction #9 for 2018: Ginni Rometty will this year be replaced as IBM CEO. This should have been a no-brainer, yet Ginni is still running IBM, so this is wrong. But I’d be willing to bet that this is the first of my IBM predictions that Ms. Rometty has ever wanted to be true.

Here’s the deal. The rule at IBM has been for more than 100 years that unless your last name was Watson all IBM CEOs retired at 60. Only Tom Watson Senior and Junior defied this trend… until now. Ginni Rometty turned 60 last summer and I am sure she would like to be on a golf course somewhere, but what’s keeping her on the job is IBM’s still listless financial performance, it’s many legal problems (age discrimination suits, anyone?), and the fact that Ginni hasn’t been able to yet jack her retirement package up into the stratospheric numbers pioneered by her predecessor, Sam Palmisano. Sam lucked-out and Ginni is paying the price.

I don’t know if Ginni will retire in 2019 or not. It really depends on that golden parachute, which will require a few more quarters of profit growth (well at least one more, going for two or three out of 25 or so consecutive quarters) before Ginni pulls that rip cord. Buying Red Hat helps, though that purchase won’t close until the second half of this year. Maybe Ginni’s career will close with it. Thousands of current and former IBMers certainly hope so. Still, the prediction was wrong for 2018.

Six for nine.

My final prediction for 2018: Facebook’s Mark Zuckerberg will give up his political ambitions. What, you didn’t know Zuck wanted to be President? I’m pretty sure he once did have such a goal and, heck, if Trump could do it why not a younger, smarter guy whose wealth is real? Zuckerberg’s recent listening tours were all a part of his very deliberate transformation from geek to statesman. Alas, it won’t work.

And it didn’t, for several reasons — enough reason that I think some explanation is required. Zuckerberg’s political ambitions would have been nuked in the short term by Facebooks incredibly-stupid customer privacy blunders that are apparently about to cost the company a multi-billion-dollar fine from the Federal Trade Commission. You can’t assist in the manipulation of a national election — knowingly or unknowingly — and expect that to be quickly forgiven by voters. But even beyond Facebook’s obvious screw-ups, Zuckerberg’s experiences testifying before Congress last year showed everyone (including Zuck) that he just isn’t cut out for political office. His skin is way too thin.

The problem for Zuckerberg isn’t that he couldn’t do the job, it’s that the personal cost is too high. Presidents have no privacy, their every move is criticized, and for a guy like Zuckerberg who likes to be in control, there’s damned little control actually available. He’s started to realize this just as Facebook needed more attention during the Fake News and Russian election tampering crises, which have been brutal for Facebook. I got this one right.

So I ended-up the year seven for 10 or 70 percent correct. Historically I’ve average about seventy percent correct over the last 20 or so years, so 2018 was about average. But I also think this was the first year when every one of my failed predictions is still likely to come true.

So I may come back at some point, asking for a recount.

But enough of 2018. In my next column I’ll turn to my startling and depressing predictions for 2019 and beyond.

Yum.