Welcome to our twenty-first annual retrospective and predictions Cortex (ten from Intellyx and eleven from ZapThink, Jason Bloomberg’s previous company).
Everyone posts New Years’ predictions, of course, but ours are different, because we take turns and rate each other’s predictions from the previous year.
Here, then, is my take on my colleague Jason Bloomberg’s predictions for 2023.
Collapse of Crypto Due to Mining Lockup
Jason’s first prediction has bitcoin miners going out of business by the end of 2023 due to the lockup issue.
They have twelve days left, so it could still happen. But it doesn’t really seem like it will.
Instead, mining companies are investing in faster hardware and finding ways to work working around mining issues
Hive Digital, a leadking bitcoin mining company, recently announced they are buying 96 powerful NVIDIA H100 GPU units.
Bitcoin miners are also making the shift to staking, which reduces a lot of the effort and energy expended on the previous “proof of work” based mining. It looks like staking reduces the lockup danger to an acceptable point.
As much as we all love to hate crypto miners, I have to say this prediction did not pan out.
WebAssembly and GraphQL will catch fire
Jason is more on target with his second prediction about WebAssembly and GraphQL. Both are indeed gaining adoption.
But I’d say they are not catching fire equally – WebAssembly seems to be catching on more rapidly than GraphQL.
WebAssembly doesn’t require a lot of special coding, while GraphQL depends on having a good schema in place and on developing a set of APIs that pull the right data from that schema. Not quite as simple or easy.
Overall, I would have to say he’s on the crypto with this one, though (so to speak).
Demand for ‘Captcha for all AI’ will drive controversy and regulation
This prediction for an “AI Captcha” to weed out or at least mark AI bots is squarely in line with what happened in 2023: lots of noise about the drawbacks of ChatGPT (including a good column about some of these drawbacks from Jason himself) and increasing demands for regulation.
We covered a startup doing almost exactly this: Originality.AI, which produces an AI score on content it ingests, giving you odds on whether it’s primarily original content, or AI generated.
Their rating system isn’t exactly in the form of a Captcha or a label per se, but it does tell you how much of any given content is likely to be AI generated bullshit.
And AI regulations are being discussed from the White House to capitals all over the globe.
One good thing that’s come out of all of this – people pretty much are aware not to trust AI generated content all that much.
(Note: we at Intellyx are an AI-free zone when it comes to creating content. Yep, we are solely responsible for all our own content. 100% pure and natural – nothing artificial about it at all. Sadly, I predict this will become less and less the case next year – but I’m getting ahead of myself.)
And now here we go for 2024
Database Convergence. I wrote about it earlier in the year – how general purpose database products are adding capabilities from specialized databases, such as analytics and vector support. And it seems like the trend continued for the rest of the year.
So I think the trend will continue into 2024, as database vendors continue to jockey for position in the red hot database and data tools market.
We’ll continue to see time series databases move toward real time analytics, and vice-versa.
We’ll continue to see vector data types added to most database offerings.
We’ll continue to hear how distributed SQL makes NoSQL databases irrelevant.
We’ll continue to hear how you can use the same database for transaction processing and analytics (much like the old SNL “Shimmer” skit – “It’s a floor wax AND a dessert topping!”).
And we’ll keep hearing about how specialized vector databases such as Chroma and Pinecone out perform and are easier to use for AI applications than general purpose databases with vector support bolted on.
(I actually predict vector databases will be an exception to the convergence – the vector data type has distinct characteristics that benefit from specialized storage, sorting, and retrieval mechanisms.)
AI In the Enterprise – Rise of Best of Breed. In 2023 the emergence of generational AI shocked the business world with its potential but very little actual adoption happened.
Unless they felt directly threatened by generational AI, organizations spent a lot of time in 2023 thinking through risks and legal issues and tire-kicking various technologies (which are also evolving so rapidly as to deter quick adoption).
With the announcement of AWS Q and other AI vendors providing do-it-yourself tools, it looks like enterprise adoption is going to really happen in 2024.
The “black box” vs “custom stack” debate will heat up, with some enterprises choosing to run their AI conversations on vendor supplied integrated “black box” stacks while others will select their own best of breed products.
I further predict we’ll see a rise in uptake of best of breed solutions at the expense of the “black box” vendors.
This sounds counterintuitive but I think the best of breed solutions will give enterprises better control over the results they get from their AI bots — better control over the data inputs for the LLMs and control over the context in which the LLMs are trained, transformed, and executed for specific business solutions.
This type of control results in improved safety, security, and better legal coverage.
Conversational AI will Disrupt the Analytics Market. This one isn’t particularly mine, but I heard it at an AI conference and kind of adopted it.
It makes sense that people would rather ask questions of a chat bot than pore over data in a pie chart, bar chart, histogram, or spreadsheet to pull out nuggets of related and significant data.
This is a kind of corollary of the AI in the Enterprise, since it’s likely to be one of the big areas of adoption in an enterprise, and the objective of many enterprise IT projects.
It also goes along with the trend toward database and data product convergence — many analytics databases are likely to be repurposed as AI vector databases to support LLM transformation output searches.
Speech to Speech Translation. I’m going to go out on a limb here and predict that Google’s recent announcement of a language-agnostic translation framework is a major breakthrough in the kind of holy grail quest the industry has been on since Star Trek came out in the 1960s, at least.
The whole concept of automatically translating speech in one language to speech in another language is fascinating and intriguing. But has never become reality despite many attempts.
What Google has accomplished in unsupervised speech to speech translation in Translatron 3 sounds as if it just might work. This will become evident in 2024.
A lot remains to be seen and proven, of course. And we’ve seen a lot of prototypes and partial solutions using highly constrained language subsets.
But the shared encoder for source and target languages sounds like it might really be the foundation for a breakthrough solution.
The Intellyx Take
Anyone can predict the future, of course, and it’s always entertaining to see what people are predicting and go back and review a year later and see how much of it actually came true.
Many predictions for 2024 are, quite frankly, predictable: more data, more cybersecurity, more AI, more hybrid work plans, more “do more with less” with IT budgets, etc.
But here we’ve tried to focus on a few key aspects of the trends beyond the headlines and see if those pan out, or not. I hope I’ve given next year’s analyst some fodder for the 2025 predictions.
We are, as Jason said last year, always willing to stick our necks out and take a dunking, to mix metaphors a bit.
I think the database area is going to be very interesting. Generational AI is basically turning it on its head – especially the analytics capabilities. And vector databases are here to stay.
Combine that with real time speech to speech translation and I think we’ve got a kind of scary trend to think about next year.
Let us know what you think!
Copyright © Intellyx LLC. Intellyx is an industry analysis and advisory firm focused on enterprise digital transformation. Covering every angle of enterprise IT from mainframes to artificial intelligence, our broad focus across technologies allows business executives and IT professionals to connect the dots among disruptive trends. As of the time of writing, none of the organizations mentioned in this article is an Intellyx customer. No AI was used to write this article. Image credit: Tournesol, via Wikimedia Commons (bad translation of Shakespear to Swedish, and Forrest M. Mims III, via Wikimedia Commons (early Russian to English translation computer, 1961).
