The New Dawn of Machine Learning

Over the past five years, we have seen neural network technology take off. GPT-3 can create human-like text on demand, and DALL-E, a machine learning model that generates images from text prompts, has exploded in popularity on social media, answering the most world's most pressing such as "what does Darth Vader look like ice fishing? or "what would Walter White look like if he were in Animal Crossing?"

We wanted to know what was going on with this wave, so we asked our Director of Machine Learning, Fergal Reid, if we could choose his brain for today's episode. Although the work is always a balance between what is possible and what is doable, things, it seems, are just starting to evolve. The technology landscape is changing, business applications are (potentially) changing the game and, spoiler alert, Fergal is a firm believer in the hype.

On today's episode of Intercom on Product, Paul Adams, our Product Manager, and I sat down with Fergal Reid to talk about the recent craze around neural networks, how machine learning is powering business and what we can expect from technology in the next few years.

Here are some of our favorite takeaways from the conversation:

Neural networks have made significant progress over the past five years and are now the best way to process unstructured data such as text, images or sound at large scale. In CX, neural networks will likely be used alongside more traditional machine learning methods to choose actions that provide the best possible interaction with the customer. Building ML products requires balance - there's no point starting with the problem if the solution is unattainable, but you shouldn't start with the technology if it can't meet the real needs of customers. AI has been pretty overhyped in the past. While more realistic claims are likely to close fewer accounts, it pays off in terms of customer retention. ML teams tend to invest a fair share of resources in research that never gets delivered. Associate it as much as possible with projects that have a real impact on the customer experience. If you want to invest in ML, hire someone with both technical and operational experience so they can start working with the product team from day one.

If you like our discussion, check out other episodes of our podcast. You can follow on iTunes, Spotify, YouTube or grab the RSS feed in the reader of your choice. The following is a slightly edited transcript of the episode.

The hype strikes back

Des Traynor: Welcome to Intercom On Product Episode 18. Today we have an interesting topic to discuss. It's all about artificial intelligence and machine learning. I am accompanied, as always, by the product manager of Intercom, Mr. Paul Adams. Paul, how are you?

Paul Adams: I'm fine, Des. Thank you.

Des Traynor: And today we have a special guest, Mr. Fergal Reid, who is our Director of Machine Learning. Fergal, how are you?

Fergal Reid: That's good, Des. I'm really excited to be on the podcast today. I can't wait to get started.

Des Traynor: Excellent. I think you are our first or second guest, so you should feel very, very grateful.

Fergal Reid: I feel very privileged, indeed.

"We've seen steady progress of something new and exciting - neural network-based technology - which is really starting to come into its own and be useful"

Des Traynor: Well, let's start at the end, in a way. It feels like the AI ​​hype machine is in overdrive again. It happens every few years from my perspective, but what I can really see happening is people creating a lot of art. The DALL-E generation has kicked off and some of the images created are breathtaking. The other day I saw that there was a marketplace for DALL-E prompts where you can literally buy prompts that create images for you, which is as meta as it gets. In a more practical sense, the GitHub co-pilot can now augment your code as you write, which is pretty amazing; I played around with OpenAI's GPT-3, asked the questions, and let it create little paragraphs and stories for me, and it was pretty awesome. If we back up a bit, what's really going on? Did something happen in...