turns out that a lot of the time the best solution is just having a dude spend an afternoon hardcoding common interactions, so you know it works and how it works
Sure, they’ve invested a lot, but frankly, that doesn’t mean they’ve integrated any of it into assistant.
I haven’t seen any change in what assistant can do for years, it’s good for controlling your Google connected smart home things, asking fairly basic Google queries (the kind of stuff that will prompt a dedicated reply from Google search - like, “what’s 5 times 5”, or “how tall is mount Everest”), and… end of list.
Pretty much anything else it can/could do, it either doesn’t do, or it does so poorly that it might as well be incapable of it. I used to play a riddle game on my Google homes, which was quite good, it involved being asked a riddle, and it would wait for you to answer, then evaluate your answer to see if you were right, and provide appropriate feedback about that. Like all good things, the riddle game has gone away. I can’t seem to give my Google assistant a prompt that will cause it to work. There’s other games, sure, but the little riddle one had a little story to it and it was rather enjoyable. When you got to the end, it said something about the story that it would be expanded later (the story was always the same)… and now here we are. No riddle game.
The other stuff I know it can do is to prompt bedtime stories (good for all you parents out there), and play trivia games… with multiple participants even.
For actually useful stuff, like setting reminders, it’s not great. The reminders don’t really work, they don’t tie into your phone… There’s just a whole host of issues with it. Even the voice identification stuff is pretty much garbage… Even if it recognizes your voice, you’re basically screwed either way, because you’re not getting access to much more data than you would without it.
In OP’s case, I’m not sure it understands the location, or location based alerting is off… Since it would need to coordinate that with your phone. Which isn’t something that works with Google in my experience; I have a pixel and I’ve never gotten an alert or reminder to work from my assistant to my phone. The assistant stuff is largely built around the idea of the Google (now nest) home speakers. So functionality sucks across the board.
Honestly, the AI/LLM stuff coming out with chatGPT and others is only really exciting to me if it pairs into something like Google assistant. I can ask Gemini some extremely complex questions and get good answers, far better than assistants “here’s a result from the web” garbage. I’ve briefly tried both chatGPT and Gemini, and they both do a pretty good job. Pairing that up with a speech to text system (of which, Google’s isn’t half bad), and a good natural language text to speech system, would make it very useful. Tying that into the existing functions of assistant would be great; and it would be even better if that functionality can be extended into your apps on your phone (or at least those synced to your phone, like text chatting apps: eg: the app formerly known as hangouts, Facebook Messenger, WhatsApp, etc.). IMO, that would be game changing.
I know of some in the DIY/homebrew communities that have experimented with tying an LLM to a program that can interface with online apps, as well as do speech to text and text to speech, all locally. It’s hell to get set up, but the examples I’ve seen of it have been spectacular. I’m considering setting one up to replace my old Google/nest home units because right now, Google assistant is only really used for turning on and off my hue lights.
Home labbers have proven that it works, but as far as I can see, no commercial entity has pushed that into any consumer products yet. Including Google.
Don’t get me wrong, Google’s actual AI is great. As good as any other out there that I’ve tried, but that tech is not present in assistant from what I can tell. It should be.
Has Google assistant improved at all? I swear it came out and it was the best assistant on the market… and then it just never got better. I haven’t even had it turned on the last couple of years because it just wasn’t useful to me.
I haven’t noticed any improvements, if anything, things have been taken away, or simply stopped working as expected.
IMO, it will only improve if they tie in the features of the Gemini AI and add functionality along with it.
IMHO, most companies are hesitant to tie modern generative AI/LLMs into products that have control over anything. So adding Gemini to Google assistant which can control your home, send messages and interact with you Google account by making calls and sending emails and similar things, is possible but companies are hesitant to deploy that; likely because of the resistance a lot of people appear to have to giving AI any control (or in many cases, any access at all), to their lives. They don’t want a machine to decide what to do without being able to know exactly what it is doing and how. AI is seen by many as a kind of “loose cannon” so to speak. I understand that, but I don’t necessarily believe that.
I believe that fear is why there hasn’t been a larger rollout of an LLM to any major assistant type device.
Home Assistant is currently working hard on assistants. I’ve not used it much yet but their text to speech offers so much more than any of the larger companies in just customisation alone, plus it all runs locally.
I have Google Home devices all over but they currently mostly act as a dumb speaker and I just get HA to do all of the heavy lifting. The most Google does is set timers and even that just goes into HA for most of the processing.
It’s not hugely complicated but instead of me having to ask an assistant everything, I let HA tell me everything through various speakers based on the state of sensors around the house at appropriate times.
When I wake up in the morning and go downstairs it’ll detect my presence, and if it’s a work day it’ll inform me of weather, traffic (as well as a suggested time to aim to leave by) and a basic schedule of my day, then it’ll stick some music on.
As it gets closer to the time to leave it’ll chime up again telling me I have x minutes left to get ready, but only if it detects me in a room so I definitely hear it.
All that is controlled by HA automatically and isn’t something you’d ever get from any of the big players, because they don’t have the sort of information and stats that HA does.
If I set a timer in Google Home then it’ll become available to HA through it’s integration and I’ll pop up a timer bar on some of the displays I have dotted around so I can track the time left without having to talk to the assistant, and as any timer gets close to expiring then it’ll even show a message on the TV saying which timer is about to activate.
There’s a few smaller things that just make life a bit easier too, like turning speakers off in rooms that aren’t active, or integrating my dumb doorbell into HA using an RF receiver so I can automate doorbell presses.
And this is the company that has invested the most in AI.
Honestly, it seems like Assistant was way better at this stuff before the AI craze.
turns out that a lot of the time the best solution is just having a dude spend an afternoon hardcoding common interactions, so you know it works and how it works
Ya, just a dude and an afternoon and you’ve got it all covered, no prob.
It hasn’t been able to do location reminders for as long as I’ve used it.
Sure, they’ve invested a lot, but frankly, that doesn’t mean they’ve integrated any of it into assistant.
I haven’t seen any change in what assistant can do for years, it’s good for controlling your Google connected smart home things, asking fairly basic Google queries (the kind of stuff that will prompt a dedicated reply from Google search - like, “what’s 5 times 5”, or “how tall is mount Everest”), and… end of list.
Pretty much anything else it can/could do, it either doesn’t do, or it does so poorly that it might as well be incapable of it. I used to play a riddle game on my Google homes, which was quite good, it involved being asked a riddle, and it would wait for you to answer, then evaluate your answer to see if you were right, and provide appropriate feedback about that. Like all good things, the riddle game has gone away. I can’t seem to give my Google assistant a prompt that will cause it to work. There’s other games, sure, but the little riddle one had a little story to it and it was rather enjoyable. When you got to the end, it said something about the story that it would be expanded later (the story was always the same)… and now here we are. No riddle game.
The other stuff I know it can do is to prompt bedtime stories (good for all you parents out there), and play trivia games… with multiple participants even.
For actually useful stuff, like setting reminders, it’s not great. The reminders don’t really work, they don’t tie into your phone… There’s just a whole host of issues with it. Even the voice identification stuff is pretty much garbage… Even if it recognizes your voice, you’re basically screwed either way, because you’re not getting access to much more data than you would without it.
In OP’s case, I’m not sure it understands the location, or location based alerting is off… Since it would need to coordinate that with your phone. Which isn’t something that works with Google in my experience; I have a pixel and I’ve never gotten an alert or reminder to work from my assistant to my phone. The assistant stuff is largely built around the idea of the Google (now nest) home speakers. So functionality sucks across the board.
Honestly, the AI/LLM stuff coming out with chatGPT and others is only really exciting to me if it pairs into something like Google assistant. I can ask Gemini some extremely complex questions and get good answers, far better than assistants “here’s a result from the web” garbage. I’ve briefly tried both chatGPT and Gemini, and they both do a pretty good job. Pairing that up with a speech to text system (of which, Google’s isn’t half bad), and a good natural language text to speech system, would make it very useful. Tying that into the existing functions of assistant would be great; and it would be even better if that functionality can be extended into your apps on your phone (or at least those synced to your phone, like text chatting apps: eg: the app formerly known as hangouts, Facebook Messenger, WhatsApp, etc.). IMO, that would be game changing.
I know of some in the DIY/homebrew communities that have experimented with tying an LLM to a program that can interface with online apps, as well as do speech to text and text to speech, all locally. It’s hell to get set up, but the examples I’ve seen of it have been spectacular. I’m considering setting one up to replace my old Google/nest home units because right now, Google assistant is only really used for turning on and off my hue lights.
Home labbers have proven that it works, but as far as I can see, no commercial entity has pushed that into any consumer products yet. Including Google.
Don’t get me wrong, Google’s actual AI is great. As good as any other out there that I’ve tried, but that tech is not present in assistant from what I can tell. It should be.
Has Google assistant improved at all? I swear it came out and it was the best assistant on the market… and then it just never got better. I haven’t even had it turned on the last couple of years because it just wasn’t useful to me.
I haven’t noticed any improvements, if anything, things have been taken away, or simply stopped working as expected.
IMO, it will only improve if they tie in the features of the Gemini AI and add functionality along with it.
IMHO, most companies are hesitant to tie modern generative AI/LLMs into products that have control over anything. So adding Gemini to Google assistant which can control your home, send messages and interact with you Google account by making calls and sending emails and similar things, is possible but companies are hesitant to deploy that; likely because of the resistance a lot of people appear to have to giving AI any control (or in many cases, any access at all), to their lives. They don’t want a machine to decide what to do without being able to know exactly what it is doing and how. AI is seen by many as a kind of “loose cannon” so to speak. I understand that, but I don’t necessarily believe that.
I believe that fear is why there hasn’t been a larger rollout of an LLM to any major assistant type device.
Home Assistant is currently working hard on assistants. I’ve not used it much yet but their text to speech offers so much more than any of the larger companies in just customisation alone, plus it all runs locally.
I have Google Home devices all over but they currently mostly act as a dumb speaker and I just get HA to do all of the heavy lifting. The most Google does is set timers and even that just goes into HA for most of the processing.
I would like to know more about your setup. Sounds interesting
It’s not hugely complicated but instead of me having to ask an assistant everything, I let HA tell me everything through various speakers based on the state of sensors around the house at appropriate times.
When I wake up in the morning and go downstairs it’ll detect my presence, and if it’s a work day it’ll inform me of weather, traffic (as well as a suggested time to aim to leave by) and a basic schedule of my day, then it’ll stick some music on.
As it gets closer to the time to leave it’ll chime up again telling me I have x minutes left to get ready, but only if it detects me in a room so I definitely hear it.
All that is controlled by HA automatically and isn’t something you’d ever get from any of the big players, because they don’t have the sort of information and stats that HA does.
If I set a timer in Google Home then it’ll become available to HA through it’s integration and I’ll pop up a timer bar on some of the displays I have dotted around so I can track the time left without having to talk to the assistant, and as any timer gets close to expiring then it’ll even show a message on the TV saying which timer is about to activate.
There’s a few smaller things that just make life a bit easier too, like turning speakers off in rooms that aren’t active, or integrating my dumb doorbell into HA using an RF receiver so I can automate doorbell presses.