2023 Predictions

With that settled up, let's take a look at what I think will happen across calendar year 2023.

First off, the pandemic hasn't left us, and CDC is starting to talk about masking up again. Despite desperate efforts to pretend we're clear of COVID (and related viruses), we're not, and it's going to continue to rear its head in ugly ways going forward. I don't think we're going back to mandatory stay-at-home policies, mind you, but it's going to be a couple of years of really bad winters, medically speaking. The "taper down" I talked about last year definitely seems to be in full swing (or slope, as the case may be).

Secondly, even though the Ukraine conflict continues, and inflation reared its ugly head during the end of 2022, the economy doesn't look as bad as we thought it would six months ago, and I think things are going to "perk up" once we get through the holidays. 2022 sucked, but there's a lot of reasons to think 2023 won't be nearly as bad. (Of course, we thought that about 2022 compared to 2021, and about 2021 compared to 2020, so....)

Specifically, though, I think 2023 will show us the following:

• Blockchain-related losses are going to be huge tax write-offs. It's already starting, and other companies will follow suit as soon as they see others doing it. In a related note...

• Blockchain-related businesses are likely to be the target of lawsuits. (Probability 0.8) Individuals lost a ton of money in NFTs and other such schemes, and they'll likely look for lawyers to help them go after the companies that sold these worthless things--and the various personalities that helped sell them. (Note that one such lawsuit was dismissed, but the judge left it open to be re-filed with new evidence, which the plaintiffs have already declared they will do.)

• The fallout from the cryptocurrency nightmare is going to stain 'Web3' for a good long time. (Probability 0.8) The criminal trials are just beginning, more cryptocurrencies and brokerages and stores are collapsing every day, and the various governments are just starting to stir on the subject; the crypto stories are nowhere close to being over, and a lot of investors are going to want to steer very clear of this whole space until things settle in. Same thing happened in the early "recreational cannabis" days--and still isn't entire clear, while the feds take their time trying to figure out where they want to sit on the issue. "Web3" is not going to be a fun place to be for a while.

• Distributed systems start to embrace contract-driven development cycles. (Probability 0.7) HTTP-based APIs have become a staple of integration and partnership now for several years, but as companies find that they need to evolve their public-facing APIs, they're running into the core problems of doing so: Namely, that partners and clients get really upset when they have to rebuild API clients from scratch to match all the parameters correctly. JSON-over-HTTP was widely chosen because it wasn't as complicated or complex as something like IIOP (the old CORBA protocol) or XML-serialization-over-SOAP, but we're now starting to feel the pain of doing something "simple" but "by hand", as opposed to have a specification document (aka an interface definition file) that we can code-generate from. That's going to cause another half-spin on the Ferris Wheel of Technology History, and things like OpenAPI/Swagger and AsyncAPI are going to start to become much more popular and much more necessary in production-quality API consumption.

• Microservices start to give way back to deliberate monoliths. (Probability 0.6) The microservices hype train has just about run out of steam, as companies that embraced the microservice ideal saw some negative ramifications from having a distributed system spread out across 1700 different nodes exhibit less-than-stellar characteristics. As influential articles emerged last year, we also saw anecdotal evidence that microservices were less an architecture and more a fad--nobody could ever define exactly how granular a "microservice" was intended to be, for example. Given that the shine has come off (starting as far back as 2019), we're going to see some organizations start to compose multiple existing microservices together back into a single node, and new project will be less aggressive about trying to "microservice all the things". This could in turn have some interesting implications for the single-function services (a la Azure Functions, AWS Lambdas, etc), since they were a natural tool to use when "microservicing", since you really can't get more micro than a single function. Right? (Don't answer that--I'm sure somebody out there has figured out a way to make four services out of a single function invocation.)

• Security companies are going to need to OSS their offerings. (Probability 0.5) Those companies who sell security-related software or cloud offerings sell their wares based on the premise of trust--namely, that you can trust that they will keep their secrets safe, and therefore by extension your secrets. Sadly, it's becoming more and more clear that those secrets (such as passwords or code) are hard to keep safe. Therefore, these companies are going to need to OSS their stacks, slimming down the surface area of the data they need to keep secret, in order to rebuild some of that trust. (Sadly, I kinda expect that this won't happen, because too many industry executives still believe that secrets are in the code, when most security and cryptography experts say the opposite--spread the code far and wide, but keep the keys secret. Considering cryptography experts have about twenty centuries' worth of greater experience, I'm inclined to agree with the cryptography experts, myself.)

• Work-From-Home (WFH) continues to normalize and become "just another location". (Probability 0.7) As much as companies would like to get people back into their (very expensive) offices, there's a lot of companies that did a lot of remote hiring during the white-hot labor market during the pandemic, and it would be a pretty big bait-and-switch (not to mention terrible PR!) to have promised "remote" employment back then and go back on it now. Not to say that a lot of companies won't, but doing so would be essentially a huge black mark against that company within the tech sector, and make it that much more difficult to hire technologists for the rest of the decade. As a result, while some companies are going to continue to be "remote-always" (because they most likely don't have an office lease somewhere that they need to justify, or they've never built a corporate HQ someplace), and some companies are going to be "office-always" (because they do!), many, if not most, companies are going to simply make "remote" another location in their "Locations" dropdown on the application web form.

• Somebody is going to sue over state income tax. (Probability 0.7) One of the things that those companies who offer remote options is the fact that the legal/tax situation around remote work is murky at best. Some states have interesting laws on the books regarding employees who work within their borders, and other states have no state income tax at all. When I was at Rocket Mortgage, I discovered that there were two states in the US that Rocket simply did NOT want anybody working from, because of the byzantine tax implications it would cause us. (Don't ask me for details, I don't have them.) At some point, this is going to culminate in a legal action--against an employee, against a company, and maybe even against one or more of the state governments--because somebody didn't understand the rules, or found the rules to be prohibitive, or is facing a LOT of money lost for one reason or another. That's going to start in 2023, take many years to wind its way through the courts, and gather a lot of interesting "takes" from the business community, ending in who-knows-what kind of solution. But it'll start in 2023.

• Low-code/no-code is going to accelerate further. (Probability 0.7) With an economic contraction of any form, getting your Amazing Innovation Project(tm) funded so you can hire software developers to build it will be that much harder. You'll start looking for ways around that obstacle, and as a result, interest in low-code/no-code tools will really begin to accelerate even further. Additional prediction: In 2025, interest in integrating those low-code/no-code solutions into the larger fabric of the company's IT will skyrocket. (Personally, I think this will bleed into new language development, too, but seeing the results of that will probably take longer.)

• Internal tech audits will become relatively popular. (Probability 0.4) During times like this, when we're not quite so "move fast and break things", we have a tendency to slump in the chair and look around the room at the collateral damage that the previous "go-go-go" era caused. At a company level, that means starting to take some deep audits of what we have, and some efforts to "clean house" will result. External purchases will be paused while execs look to integrate existing things rather than building new things. Regardless of what comes out of the analysis, it starts with the IT audit, though, and as a result, it's likely to be a pretty high-priority project on everybody's road map for 2023.

• Muskian-style management will be studied, cargo-culted, and lead to the failure of more than one company. (Probability 0.6) Love him or hate him, Elon Musk has propelled himself into the center of all things "management academia" with his handling of the Twitter purchase and takeover. Personally, I think he's flailing and thrashing and in general way out of his depth, but one thing is for sure: His rampant (and some suggest ill-advised) decisions about Return-To-Office and refusal to honor certain debts (like employee severances and expenses) are being watched very carefully by other CEOs around the industry, some of whom already are signaling that they're thinking about following suit.

  At the very least, there's going to be a book or two that will be published on all this, probably due to hit shelves somewhere at the end of the calendar year, and it will garner all sorts of reviews from established periodicals. Some will be critical of Elon's approach, some will seek to find some sort of "underlying meaning" behind his actions that we can carry away insights from, but all will be looking, pointing, whispering, and taking notes.

• 2023 will be the year we try to figure out what to do with all this AI stuff. (Probability 0.8) ChatGPT and OpenAI have completely taken the media and social media world by storm, but it isn't clear yet how any of that actually helps. Like most applications of AI, it will likely be a mix of human creativity and AI computing power, but how? With all the image-recognition stuff that was the darling of social media, we ended up doing.... what... with it? It certainly hasn't changed anybody's life that I know. That's what we'll spend 2023 trying to figure out.

  Oh, and the whole "GitHub Co-Pilot is going to take your job, you silly programmers" hype? Hot air. Whipped-up hot air. With all the nutrition and substance of cotton candy. Look, CoPilot might be able to take certain things on its own hands, but (1) it's proprietary to GitHub, (2) it's built on top of open-source projects whose contents are not acknowledged, violating license, and (3) it's not going to be able to bring the creative element that humans bring. It's a legal mess, and given the deep pockets GitHub is now blooming from (that is, Microsoft), it's ripe for some major legal action before it can safely take anyone's job. (Not to mention, somebody still has to drive CoPilot.) Worst case? CoPilot elevates the abstraction level, bringing a more natural-language element to programming languages--because somebody still has to figure out which algorithms and how to "codify" business rules. (FWIW, people said programmers were obsolete with the release of Visual Basic 3 and other high-level 4GLs; how'd that turn out?)

  BONUS PREDICTION: 2024 will be the year that we start putting AI into some of our apps--like using ChatGPT for chatbots--and 2025 will be the year that we start ripping it all back out because we completely got it wrong. (We always do this. And I see no signs that we'll stop in my lifetime.)

• Companies will accelerate their drive to become 'tech companies'. (Probability 0.6) For quite a few years now, there's been a drive among a number of firms to be seen as a "tech company", with all the perks and benefits of being one, but with many (if not most) of the companies seeking to make this transformation unsure of what that exactly means. Particularly with the projected "soft" IT market next year to start, it will be a good time to take a moment to take stock, reshape the thought process, and then jump into it, particularly because...

• Hiring will accelerate at the end of 1Q/2023. (Probability 0.8) Even as companies are engaging in layoffs, these aren't the savage cuts that we saw back in the Dot-Bomb era (circa 2001) or the Great Recession (circa 2008) era. Both of those were about the industry figuring out that whole lines of business were suddenly seen for what they were (unprofitable--Pets.com, anyone?) and making some fundamental adjustments to our thinking. This time around feels more like a "correction", more of a "We know there's something in this line of business, but we didn't quite nail it and we need to re-examine and re-allocate". People I've talked to at both Microsoft and Amazon feel like as soon as this "organizational refactoring" is finished (which shouldn't take too long after the holidays), things will start to spin back up again. Which will also mean that...

• Entry-level positions are going to be easier to come by. (Probability 0.5) When the market gets soft like this one appears to be, companies often start looking for ways to do take smaller risk in their hiring by hiring people newer to the workforce--a la junior and associate developers (and their kin), who they can hire at half of the cost of a senior software developer, thus risking less against their budget in the event the hire doesn't work out. (It doesn't mean companies will make it any easier for those early-stage developers to get hired, though, because they'll still be trying to put candidates through seventeen rounds of interviews over a six-month period before offering...)

• Platform-oriented development is going to begin making some waves. (Probability 0.6) As companies start looking for ways to create new revenue opportunities for themselves, or at making the transition into becoming a "tech company", they're going to start looking very hard (once again) at how to build platforms out of their business, and the software they seek to build will follow suit--in other words, developers will be asked to build platforms rather than just hooking into one.

  In fact, there is a reasonable chance (and maybe this is more of a 2024 thing) that libraries/frameworks/languages will emerge to make it easier to integrate two platforms together more seamlessly and easily, rather than the current crop of tools designed to consume a platform's offering.

• Cloud will begin to shift. (Probability 0.5) TikTok is now an enemy of the US state, because it appears to be a tool of the Chinese state. Russian hackers and bots are shadowy figures looking to destabilize just about anything they can in the West, or so the pundits would have us believe (and which is probably at least partly true). And massive data breaches continue to plague comapnies left and right (lookin' at you, Twitter....). Cloud is not responsible for these things, and in many cases will be a powerful tool to fight against them, but the US government is signaling that it is going to start cracking down on some aspects of how the virtual world operates--and that could mean some guidance and/or restrictions around where data gets to live and where it gets to be processed. Even before the Feds tell CEOs where to put it, though, CEOs are going to start asking themselves the same questions, and this will lead to some "reallocation" of assets in and around various clouds. It's not likely cloud usage will shrink, but how we use cloud and where the cloud servers that we use are located will become more important.

• BOLD TWITTER TAKE: Elon will sell Twitter (perhaps involuntarily) at the end of 2023, to another tech firm (like Microsoft or Oracle). (Probability 0.6) The current model is unsustainable. Twitter has no source of income: advertisers are bailing, and his "verified checkmarks" income is nowhere close to what he needs to keep things going even under the new staffing levels. (And let's just leave aside the existing debts and/or expenses from the pending lawsuits that are likely to slam into him in 2023--I'm sure he can get Tesla or SpaceX to cover those, no matter how illegal that may be....) Twitter's losing people in droves, there's no new signups to replace the losses, and it's now become clear that where Twitter might have had a "cultural monopoly" on the idea of "short-form mass-widespread posting" (as opposed to some of the other forms of social media), that monopoly has now been shattered as non-technical folks start to pick up on terms like "Mastodon" and "Post". Twitter needs users, and users are very quickly learning they don't need Twitter. Without stable leadership and a massive infusion of cash, the platform will plunge in value until one of the other tech giants just shrugs and says, "We can pick it up out of petty cash--how can we not buy it now?"

  And that's how the grand MuskTwitter experiment will end. Not with a bang, but with a pathetic whimper.

  Meanwhile, alternatives to Twitter will continue to sprout up, but none will actually be able to be everything Twitter was, partly because Twitter was (at the time) unique in its approach to social media, and as a result it created a "perfect storm" that led the platform to its current (well, pre-Muskian) level of success. That will be nearly impossible to recreate.

• BOLDER SOCIAL MEDIA TAKE: RSS will make a comeback. (Probability 0.4) RSS, the protocol that make blog-aggregators work to consolidate/federate all the blogs you wanted to read into one place, will be adapted or adjusted to work with all the differnet social media sites you want to read, so that social media aggregators can now bring all the social media you want to read into a single place. And if it's not RSS, it'll be a new protocol that does basically the same thing (because our industry loves building a new protocol to do the same thing an old protocol does, only "better"). Maybe ActivityPub but I won't hold my breath.

Talk to you all next year... if not sooner.

Last modified 29 December 2022