Predictions for 2024: AI, AR and Satellites | ✉️ #36
Hey! 👋
We are almost 3 weeks into 2024, so it’s almost too late to give any predictions - but I’ll still try and I’ve got three.
Firstly, I think 2024 is the year we will see the true mass adoption of LLMs and related AI technologies. Just now Samsung showed all the ways Google Gemini is integrated into their new phone, and I am pretty sure that Apple will come back with it’s own answer on the next WWDC. You might be thinking that ChatGPT and alike are already massively adopted, but I want to challenge that assumption. When I look around, I don’t see that many people using it just yet. There are many who do, but most still don’t. 2024 is going to change that.
Secondly, I am extremely excited to see that the hardware and software around AR finally got to the stage, where quite some new use cases are being unlocked almost every day. I don’t remember seeing that many really cool examples of how the headsets like Quest 3 can be used. The line between the outside world and what the headset projects on top of it is finally close to none-existent. It seems to be one of those things, that won’t be used by everyone right now, but it will be a whole new level for many professional use-cases. Just look, at what kind of Iron Man Jarvis-like work we can do now. Or what learning to play the piano looks like with AR . Or how deeply immersive the work of interior designers and architects becomes, or what kind of demos those professionals can offer to their customers to really see how a certain space would look like. And then, of course, there is Vision Pro, coming out any minute. VR is amazing for playing games and watching videos. AR, which requires more complex hardware and software, is going to be amazing for a plethora of professional use cases.
And finally, satellites. I just can’t stop thinking, where companies like StarLink and Kuiper will get us. High-speed internet, anywhere, with just a small dish device? Whole office buildings connected to the web without a single cable in sight? Imagine what bringing the internet to a country that's unfortunate to not have any decent alternative would be - how much simpler the rollout becomes. In the mountains, on the sea, and even in a Bavarian village. And rumors are, we will eventually get access to that internet from our smartphones, without an extra device. iPhones already got some satellite emergency features last year, which is obviously just a little first step towards a much bigger change in how we connect to the internet in places, where it was previously impossible.
One bonus item that got me excited: compact nuclear batteries, which, apparently, were in development for many years, and are almost there. Or, at least, one Chinese company claims this to be the case, with first mass-produced, 50-year-lasting, zero outgoing radiation batteries coming out.. any time now? What a time to be alive.
2024 is going to be exciting technology-wise. And hopefully, all this great new technology will result in a big positive change for all of us. Now let’s get back to work and do our part in this journey.
What We've Shared
The biggest cloud scam. Each instance of communication with a machine across various availability zones incurs a cost, as does every instance of data egress and each time you exit a cloud network. These constant charges represent what could be considered the greatest deception in the history of cloud computing:
AWS re:Invent 2023 Top Announcements. AWS re:Invent 2023 is over and it's time to look at the most important releases, from Amazon Q and Bedrock, to AI in CloudWatch and slightly ridiculous new Serverless offerings:
DevOps Accents #28: What Data Protection means in 2024? In this episode of DevOps Accents, Pablo and Leo talk to Filip Verloy, a Field CTO at Rubrik:
What We've Discovered
It’s About Time! Brilliant explanations about why exactly time in compute systems matters - and why AWS adding microsecond-accurate time is kind of a big deal for distributed systems.
Speed Up Container Image Builds in Tekton Pipelines. Hot take: real speed up will come from not trying to bind together 90 CRDs to get a a simple pipeline running.
LLM Inference Performance Engineering: Best Practices. Many good tips on running open source LLM models in production.
Incident Response and DevOps in the Age of Generative AI. Experts from multiple companies share there experiences with using AI for handling incidents.
Aurora Servelress v1 is being shutdown. On one side, it’s sad to lose the only DB that actually scales to 0. On the other side, it’s good that the confusion around two simultaneously developed versions with different feature sets is finally gone. And v2 is actually great anyway!
A random reminder
Our webinars continue in 2024! Check out individual webinar pages for dates. Do you want a specific webinar with a practical explanation? Reply with topics you're interested in and we'll see what we can do :)
The 37th mkdev dispatch will arrive on Friday, February 1st. See you next time!