- Like
- SHARE
- Digg
- Del
- Tumblr
- VKontakte
- Flattr
- Buffer
- Love This
- Save
- Odnoklassniki
- Meneame
- Blogger
- Amazon
- Yahoo Mail
- Gmail
- AOL
- Newsvine
- HackerNews
- Evernote
- MySpace
- Mail.ru
- Viadeo
- Line
- Comments
- Yummly
- SMS
- Viber
- Telegram
- JOIN
- Skype
- Facebook Messenger
- Kakao
- LiveJournal
- Yammer
- Edgar
- Fintel
- Mix
- Instapaper
- Copy Link
Google Assistant provides a powerful voice assistant to many smart speakers, Android phones, smart displays, and other devices around the world. Even in this new world of Bard and Bing, it’s unlikely that Assistant is going away.
Google is currently testing Bard, an AI that can accept questions in natural language format and provide text in response. It’s intended as a competitor against ChatGPT and Microsoft’s Bing Chat, and it will soon be available in Google Docs, Gmail, and other Google services. It does have some functional overlap with the existing Google Assistant, since both technologies are designed to answer questions and perform actions based on natural language questions, but it’s not a true replacement for Assistant.
Last month, CNBC reported that Google was shifting some staff currently responsible for developing Google Assistant to work on Bard, citing a memo sent to Google employees. The shift reportedly includes Google Assistant’s vice president for engineering, Amar Subramanya, moving to the Bard team. There was also a report from The Information in October that stated Google was investing less in Assistant on headphones, smart glasses, smartwatches, and other form factors, which may have been a cost-cutting measure. More recently, Google is ending support for third-party Google Assistant smart displays.
Now, all that put together might sound like Google is phasing out Assistant, possibly in favor of Bard or something based on it. There have been at least a few news articles recently making that assumption. The more likely future is a bit more complicated, and it almost certainly doesn’t involve Assistant going away completely.
Assistant is Feature-Complete
Google Assistant was officially released in May 2016, first as a chatbot in the Allo messaging app, then later appearing on smart speakers, Android devices, Chromebooks, Wear OS watches, Android Auto, and Google’s other software platforms. Assistant was an evolution from the company’s previous search and voice assistant tool, Google Now, which itself was an upgraded version of normal web searches. That pushes back the timeline for Google’s voice assistant efforts to around 2012, over a decade ago.
It doesn’t feel like a stretch to describe Google Assistant as “feature complete,” and it has arguably been in that state for a few years. It can give you search results, control your device, connect to streaming services, and even give you preemptive information like weather reports and shipment information before you ask. It’s also intertwined with Google’s smart home ecosystem, which seems like a success — it’s hard to find a new smart light bulb or appliance that doesn’t connect to Google Home or Assistant in some way. In my own experience, I can’t think of anything that feels broken or incomplete with Assistant, except for some occasional issues with multi-step commands (like turning down the volume and start playing a podcast).
For Google Assistant, there are no more worlds to conquer. Google said in October 2022 that Assistant has more than 700 million monthly active users, covering over 95 countries and 29 languages. That’s nearly twice the population of the United States. Assistant seems to be in a similar position as Android, where the focus is now on smaller improvements and security upgrades, rather than groundbreaking new features.
It’s easy to assume that a software project is dead or abandoned if it’s not recieving a constant stream of new features, which is how many products end up with feature bloat or other similar problems. Google Assistant is fine. I would much rather Assistant stays in its current state than become something like Microsoft Edge.
The other evidence for Assistant winding down seems… far-fetched. Moving staff from Assistant to Bard makes sense, because Assistant is already a successful product, and Google would obviously like to replicate that success with a project that is just getting started. The end of support for third-party displays isn’t great, but it’s not clear Google or the display manufacturers cared to keep the effort going — there aren’t many third-party smart displays running Alexa, either.
Bard Is Not a Replacement
Google Assistant, Bard, and other generative AI solutions all share a broad goal of answering questions in normal spoken or written language. Bard and Bing Chat even have a similar messaging chat-style interface as Google Assistant on phones and tablets. It’s easy to see how Bard could replace Google Assistant in that regard, but that’s a long ways off, if it ever happens at all.
The first problem is computing power. Google Assistant is primarily powered by cloud servers, which is why Google’s original 2016 smart speaker still works today. Pixel phones have an on-device version that can answer some responses without any servers. Bard and other similar AI technologies are large language models (LLMs) that require far more computing power. A chairman at Alphabet (Google’s parent company) told Reuters in February that talking to an LLM is ten times as expensive for servers than a standard keyword search.
Google and other tech companies are working to improve the efficiency of large language models, but they will likely remain far more costly to operate than usual keyword searches or voice assistants for the foreseeable future. Google said that over 700 million people use Assistant at least once a month, which already requires many servers spread across the globe, and switching to Bard would require duplicating that several times over. Bard would also be impossible to run locally on most phones or tablets — Alpaca, a chatbot like Bard and ChatGPT, needs 16 GB RAM and 20 GB of storage for the best-available model. Google could build a more limited version for offline use, but it’s not clear what advantages that might have over the current Assistant.
There are other problems with Bard that make it a poor replacement for Assistant. Google would need add new integrations for smart home devices, starting navigation, device controls, and other functions, like ChatGPT plugins. Google Assistant also provides information in a completely different (and more helpful way) than Bard, in the form of summary cards and links to the source. In its current form, Bard usually doesn’t provide sources, and likes to make up information.
The Hybrid Approach
It’s possible Google could develop a hybrid solution, where most Google Assistant queries are handled by Assistant’s current technology, but more advanced questions are handed off to Bard for processing. That would keep Assistant functional with all of its current features and advantages, while giving it some of the capabilities that people enjoy in tools like Bing Chat and ChatGPT.
Google I/O, the yearly event where Google reveals new products and developer tools, is happening again next month. The event will likely be filed with AI news and updates, and it’s possible we might get a sneak peek at future updates for Bard and Assistant. In any case, Assistant going away doesn’t seem any more likely right now than Siri or Android going away, though it may get a few LLM-powered upgrades at some point.