The latest advancements in voice assistants have officially landed, but the excitement is somewhat muted—here’s why you might want to think twice before diving in.
After a decade of development, Amazon and Google have unveiled their new-generation voice assistants, yet despite the advancements since their original introductions, they still seem to echo the flaws of their predecessors—think of them as a bit slow on the uptake.
This week, Amazon introduced Alexa+, a revamped version of its classic Alexa, powered by chatbot technology. This upgrade means that anyone with compatible Echo devices can now switch to this enhanced assistant that utilizes a large language model (LLM) akin to the technology behind OpenAI’s ChatGPT. After a year-long early access phase, which one would expect to refine the user experience, the feedback thus far has been rather disappointing.
Reddit users are expressing their frustrations over sluggish response times, lackluster accuracy, and overall dissatisfaction with the experience. Essentially, these complaints mirror the criticisms that were prevalent before the integration of generative AI into these systems. While I haven't extensively tested Alexa+ myself, reviews from outlets like The Verge indicate that the results are less than stellar, to say the least.
This underwhelming performance isn't isolated to Alexa+. Google has also rolled out a new version of its AI-driven voice assistant, but, similarly, it hasn't sparked much enthusiasm. I’ve personally used Gemini for Home—the next iteration of Google's assistant—for a couple of months, and I can confirm the widespread lack of excitement is justified.
Although Gemini for Home may boast a more natural-sounding voice and conversational style, its actual functionality often matches or even falls short of previous versions. A notable issue is the increased time it takes to process commands compared to earlier iterations. This delay could be acceptable if the assistant were more precise and capable of executing tasks accurately, but my experience suggests otherwise. Like Alexa+, Gemini for Home occasionally misinterprets requests, overcomplicates simple tasks, or fails to understand commands altogether—issues that have plagued voice assistants for years.
There are some advantages, such as the ability to command Gemini for Home to perform multiple tasks at once, but I would be misleading you if I claimed these enhancements have fundamentally transformed the smart home or voice assistant experience as Google had anticipated.
In essence, the current state of next-generation voice assistants feels lackluster. I use the term "current" because there’s always potential for growth—perhaps a breakthrough in technology could finally meld the capabilities of voice assistants with LLMs in a way that provides a significant upgrade. However, I have to admit that my optimism for such a turnaround is waning. A contributing factor to my skepticism is the other major player in the voice assistant arena: Siri.
It’s clear that Apple has encountered challenges in making Siri perform as desired. Although a revamped version was announced for 2024, Apple has yet to fully unleash its capabilities, delaying its rollout due to concerns regarding performance and its readiness to handle widespread use. While it seems likely that Apple will soon join the ranks of Alexa+ and Gemini, potentially by this spring, it serves as a reminder that creating effective voice assistants is far more complex than it appears. Even when it does arrive, it’s important to note that Siri will leverage Google’s Gemini technology due to a partnership established earlier this year.
Until someone can unravel the intricacies of perfecting voice assistant technology, we are left with LLM-powered systems that still resemble their non-LLM counterparts. This reality suggests that neither Google, Amazon, nor Apple is achieving the revolutionary progress they envisioned. So, brace yourselves, as you might find yourself repeatedly asking your smart speakers for the foreseeable future.
