{"id":162711,"date":"2022-12-20T00:14:04","date_gmt":"2022-12-20T00:14:04","guid":{"rendered":"https:\/\/harchi90.com\/after-retreating-in-2022-what-is-assistant-to-google\/"},"modified":"2022-12-20T00:14:04","modified_gmt":"2022-12-20T00:14:04","slug":"after-retreating-in-2022-what-is-assistant-to-google","status":"publish","type":"post","link":"https:\/\/harchi90.com\/after-retreating-in-2022-what-is-assistant-to-google\/","title":{"rendered":"After retreating in 2022, what is Assistant to Google?"},"content":{"rendered":"
\n

The height of \u2014 what I deemed \u2014 an Assistant takeover was when \u201cGoogle Assistant\u201d got listed after Android in the \u201cOperating System\u201d section of the Pixel 3’s tech specs. A few years later, things have changed and the clear trend for Google Assistant in 2022 is one of retreat. <\/p>\n

Assistant in 2023 <\/h2>\n

This retrenchment was made clear this year after Assistant already had a quiet 2021 (as foundational advancements remain in the wings). The first warning sign was Google shutting down Assistant Snapshot, which had traces of what made Google Now so promising as a personalized, centralized feed that could have unsiloed information out of apps. <\/p>\n

The second was the removal of Assistant’s Driving Mode \u201cDashboard\u201d that \u2014 in May of 2019 \u2014 felt like a major coup by Assistant over Android (Auto). It took a long while to even launch and has since been discarded for a Google Maps-first experience.<\/p>\n

\n
Pixel 3 specs in 2018<\/figcaption><\/figure>\n<\/div>\n

Lastly, after over a year of work, Google stopped development of Assistant Memory in June. It looked to supercharge saving content and reminders on Android. This was a good idea as the Google app’s Collections feature is underwhelming (with a poor UI) and would have been a welcome successor. <\/p>\n

Besides all three being feeds, the important commonality among Dashboard, Snapshot, and Memories was how they were all attempts to extend Assistant from its voice roots. They were rife with Assistant branding, but ultimately had nothing to do with Assistant’s core and original competency. Their primary claim to being helpful was centralizing relevant information, but there’s nothing inherently Assistant about that.<\/p>\n

Meanwhile, a report from The Information<\/em> in October laid out how Google was going to \u201cinvest less in developing its Google Assistant voice-assisted search for cars and for devices not made by Google.\u201d This was framed as focusing and improving the experience on first-party Google products, while it comes as the company is cost-cutting by consolidating various efforts.<\/p>\n

The most significant result of this was the Sense 2 and Versa 4 not featuring Google Assistant even after it was offered on the previous generation of devices. That’s also the case on non-Pixel and Samsung wearables at launch. It’s wild that Wear OS 3 devices with only Amazon Alexa exist. Another \u201cless important\u201d area of \u200b\u200binvestment was said to be Assistant for Chromebooks. <\/p>\n

\n
\n
\n
\n
\"\"<\/figure>\n<\/div>\n<\/div>\n<\/div>\n<\/div>\n

Assistant is no longer\u2026 <\/h2>\n

At its height, Assistant looked to be the connective tissue between all of Google’s form factors, if not the main way you interact with those devices. It being the primary interface on smart speakers and headphones was straightforward enough, but Google then tried to give Assistant UIs, starting with the Smart Display. That still felt like a good idea, but \u2014 in hindsight \u2014 Assistant replacing Android Auto for Phone Screens did not make a great deal of sense. Yes, voice input is a key part of getting help when driving, but Google could have left the touch-based UX to Android.<\/p>\n

Another prime example of that is how the original Reminders experience, which is set to be replaced by Tasks with Assistant integration, had a horrible phone-based UI for years. Interacting by voice was fine, but it’s not clear if the Assistant ever had visual expertise, and it’s better to let existing teams continue that work.<\/p>\n

At I\/O 2019, when the new Google Assistant that worked on-device was also announced alongside Driving Mode, it seemed that the company wanted Assistant to be another way to control your phone. There are times when that is very helpful, but it will never be the primary interaction method.<\/p>\n

\n
\n
\n
\n
\"\"<\/figure>\n<\/div>\n<\/div>\n<\/div>\n<\/div>\n

Having lived with that next-gen Assistant for a few years now, which has yet to expand beyond Pixel phones, it has not changed the way I navigated my device, which is something Google was pushing on-stage at one point. There’s a touchscreen for a reason and taps will likely be the fastest way of doing things for the foreseeable future (until smart glasses). Another reason could be the lack of buy-in from apps and letting them be controlled by voice.<\/p>\n

If anything, the takeaway from on-device Assistant was how that work led to Assistant voice typing on Pixel phones, and how that technology is better applied to focused experiences, like transcription and editing in Gboard.<\/p>\n

Where Assistant is going <\/h2>\n

Of the Assistant announcements Google made this year, the clear direction is improving the core voice experience. On the Nest Hub Max, you can now invoke the Assistant by just gazing at the Smart Display thanks to the camera-powered Look and Talk, while Quick Phrases let you skip \u201cHey Google\u201d for pre-selected commands. <\/p>\n

At I\/O 2022, Google also announced how Assistant in early 2023 will ignore \u201cumm,\u201d natural pauses, and self-corrections as you’re issuing a command. Like with the two new ways to activate Assistant, the goal is to make the experience more natural:<\/p>\n

\n

To make this happen, we’re building new, more powerful speech and language models that can understand the nuances of human speech \u2014 like when someone is pausing, but not finished speaking. And we’re getting closer to the fluidity of real-time conversation with the Tensor chip, which is custom-engineered to handle on-device machine learning tasks super fast.<\/p>\n<\/blockquote>\n

Meanwhile, Google is using AI to improve the accuracy of Assistant smart home commands across three areas.<\/p>\n

\n
\n
\n
\n
\"Nest<\/figure>\n<\/div>\n<\/div>\n<\/div>\n<\/div>\n

Natural language understanding is what drives Assistant. It’s gotten a great deal better in the past decade or so, and it’s easy to forget those advancements. Of course, that’s because our expectations are ever-growing and it seemed that the original Google Assistant was trying to do the same with its reach. There were of course improvements to the core voice experience as Assistant spread its tendrils across all of Google’s consumer-facing products, but this looks to be the case of trying to do too much, too fast.<\/p>\n

Google’s forever challenge is taking R&D from the lab and giving it a good end user experience. Assistant can be at the forefront of that for Google, especially as new form factors arrive. However, trying to replace what works has not worked well, but the company is at least now finally aware of that.<\/p>\n

\n

FTC: We use income earning auto affiliate links.<\/em> more.<\/p>\n

<\/div>\n\n

Check out 9to5Google on YouTube for more news:<\/i><\/p>\n

<\/p>\n