The Future of LLMs and Spatial Computing with @mad.marchy: TDI 17

Threads Dev Interviews

I am finding developers on Threads and interviewing them, right on Threads.

Note: The views in these interviews are personal views and do not represent the interviewee’s employer.

“I think this kind of technology will be essential for expanding how we interact with the world around us…”
— @mad.marchy

Let’s get this week off to a good start. @mad.marchy What is one area of technology or development that you are most passionate about right now, positive or negative?

Picking the tech I’m most passionate about is a difficult choice! But I’d say in the near future I’m most excited about large language models, after that would be spatial computing. There have been some pretty big advancements in both recently, and I think we’ll be seeing the results of that pretty soon in commercial tech.

OK, let’s talk about Large Language Models (LLMs). For those who don’t know, what are LLMs? And what do you see as the best usage of LLMs over the next 1-3 years?

Large language models are basically AI trained to understand and generate human-like text. In the near future I think the most use we’ll get out of it will be in research and discovery. Having the ability to “load in” vast amounts of information and data, then being able to speak directly to the data, is invaluable for progress. Outside of research, I think we’ll see LLMs as a form of interface for computers. Instead of clicking through windows, we’ll be able to speak directly to the computer!

Highly recommend this talk about GPT if you’re curious about the future of AI and how quickly things are progressing. youtu.be/bhYw…

Spatial Computing – not a topic we have touched on yet in these interviews What is it and why do you see it as impactful?

Spatial computing, in a nutshell, is providing a way for devices to communicate and share data directly with an intuitive interface. A great current example of this is from FluidOne who is using the U1 chip in iPhones to expand functionality and convenience for IOT devices. I think this kind of technology will be essential for expanding how we interact with the world around us and make simple conveniences much more tangible.youtu.be/DHBu1…

There’s also a lot of challenges though, like privacy and security. It opens the door for not knowing who is watching and listening to these interactions, and what kind of data they’re able to collect especially in public places. But once we figure these out, I think we’ll start seeing this everywhere. Imagine walking into a store and getting a notification for a sale along with AR directions to the product. Or being at a party and seeing all the profiles of the people around you from your phone.

What type of development do you currently do?

Right now I work in logistics as a software engineer, focusing mostly on 3rd party integrations and front-end for a large transportation management system. It’s been nice so far 😄 Previously I’ve been a bit all over the place, working in theme parks and ticketing, access control systems with facial recognition, crowd monitoring systems and IT/Engineering management.

What is your development language and environment of choice right now?

I’ve been using JavaScript for around 20 years, then node for the past ~10. I rarely find that I need to reach for anything else, aside from image processing with python. Framework-wise, I’m a big fan of nextjs especially with the latest React18 updates. Outside of building applications though, I’m a big fan of building shell scripts for support utilities and infrastructure management. Any repetitive task is worth automating, especially in the IT realm, and it’s all available in the cli.

What are some of your favorite developer resources? sites? videos? people? podcasts?

freecodecamp.org is a great way to get started. They have multiple courses and certifications, as well as project ideas that are great for learning more advanced workflows when trying out a new language. Aside from that, the best way to learn is to do! Not every piece of code you write needs to be enterprise-ready. Creative coding, finding ways to make your job or your life easier with code, and learning how to solve the right problems allow you to learn a lot more than you’d think!

If you’re already a developer and want to get ahead of latest trends in tech, some great blog posts about the future of Serverless and distributed computing.

How can people find you elsewhere online?

Thanks for having me!

I mostly hang out around here now, but I’m on pretty much every platform under the same handle. If you’re interested in some projects I’ve been working on, feel free to check them out! (I’ll be updating readme’s later 😅)