Technology trends for product managers
So far, 2022 has been an incredible year for technology–from breakthroughs in AI to 5G connectivity.
Each of these technologies has the power to improve a product’s functionality and value if implemented successfully. To keep you up to speed on these technological developments, we’ve created a quick overview below, sourced from our engineering leaders as they contribute and innovate on the world’s leading products.
Here are some of the most interesting technology trends and advancements that product managers should consider in today’s landscape.
Natural language processing (NLP)
Advances in sentiment analysis are changing the way users communicate with machines
Single-page web apps
These are finally dying, but you have some good alternatives
Machine learning as a service
Enables you to analyze more data, with less risk
Commercial and industrial have made huge strides, helping teams break into this space
Let’s dive into these predictions, their projections, and nuances.
Natural language processing
Altering the way we communicate with machines
NLP has been around for decades but has advanced significantly in the past few years.
As of today, there are some insanely futuristic language models that are ready to be unleashed. Certain industries need to understand the potential of this technology or risk becoming obsolete.
In one example, an NLP concept called “semantic search” (see the sentence similarity Huggingface models) challenges traditional keyword-based search technologies. Semantic search engines (aka, Vector DBs) like Jina, Haystack, and Weaviate threaten 800-lb gorillas like Elastic Search – that is, if the latter doesn’t embrace the new tech.
By using semantic search, product teams will be able to solve the problem of language ambiguity, which will enable machines to understand humans on a deeper level. Here’s an example of how these models could be used:
A user will be able to search Netflix for,
“a scary movie that is safe for my 12-year-old”
and their search engine will be able to serve up films that fit these abstract parameters.
A more serious example might be a case for robo therapists. Users can describe how they feel–complexities, conflicting emotions, and all–and a machine can recommend useful resources.
The ability for machines to understand the ambiguity of language is officially upon us.
Death to SPAs
SPAs were never great for users
SPAs, or single-page web apps, have been incredibly popular over the last several years. They’re characterized by:
– Fairly long initial load times
– Different parts of the page change at different times e.g., compare slack (a SPA) to Wikipedia (a server-rendered site)
– Data is served up by APIs, usually after the page loads something
The main benefit SPAs have provided is a clearer separation of responsibilities between frontend and backend developers. That’s useful for internal teams, but not for consumers. In fact, SPAs are often a worse experience for users unless your team is spending a lot of time and energy to get things right.
Think about when you click a link and a new page loads, but the page doesn’t automatically scroll to the top and you have to do it by hand. That only happens with SPAs because they’re trying to avoid full-page reloads.
This has caused a growing backlash against SPAs.
Luckily, if you’re a product manager of a SPA, you have some good alternatives.
Today, teams are starting to use a new generation of frameworks to get the advantages of both SPAs and server-rendered sites. Next.js, Remix, and SvelteKit are all examples of this.
Machine learning as a service
Enable products to analyze more data with less risk
Today, it’s difficult to find machine learning engineers and data scientists. The individuals and teams that have this specialized knowledge are scarce, making machine learning cost-prohibitive to many companies.
However, new tools and services are starting to allow pretty much any full-stack software engineer to use general-purpose machine learning models, or easily train models themself.
An example of this is computer vision. Computer vision modeling is so advanced, that it’s often wasteful to hire a machine learning expert to train a custom image classifier. Instead, they might use a solution like AWS Recognition. This kind of out-of-the-box solution is spreading to other facets of machine learning.
What does this mean for you? If you have software engineers on your team, you will be able to utilize machine learning for many of your automation and data needs. These solutions will be more available and approachable for everyone, even small companies and startups.
View these ML services here
Amazon Web Services ML
Google Cloud AI
Azure ML Services
IoT development standardization
Both commercial and industrial advancements see risk reduced
Don’t expect much until 2023, but knowing that standardization is coming may affect your product roadmap.
Eventual standardization of IoT development is necessary to make systems cheaper to produce and maintain. Think AA batteries and USBs. If every product had its own batteries and chargers, it would be a nightmare for customers.
And while IoT products have been around for a while, there is no standardization among them. However, we think this year will begin to change, with impactful effects in 2023. There are two noteworthy IoT trends.
Matter, a unified protocol for IoT devices, is in development now. Matter’s goal is for any standard IoT device to be able to connect with other devices, regardless of brand. For example, you could cordlessly connect your Roku with your Samsung Smart TV and program your Phillip’s Smart Lights to dim during movies.
According to Silicon Labs, “Matter simplifies both product development and the end-user experience by providing a unified connectivity standard for a wide range of smart home and commercial applications.”
Matter will help streamline the development of smart home IoT devices because developers will no longer have to provide specific interfaces for Google, Amazon Alexa, etc. The network effect should take over as the market puts additional pressure on IoT device manufacturers to support Matter. However, we don’t think Matter will be a reality until the end of 2022.
Recently, Amazon Web Services launched AWS Private 5G for manufacturing sites. This program encourages businesses to deploy their own private 5G network which will allow businesses to:
- Connect thousands of devices/machines with high bandwidth and a private 5G network
- Get a network up and running in days
- Secure a network with access controls for all connected devices
- Scale network capacity on-demand or add devices with a few clicks
- Only pay for the capacity and throughput used
Each of the technology trends above has serious power–to engage users, automate analysis, and create new data sets. To explore how to leverage these technologies, contact our product team.
Personalize your experience