14 күн бұрын
21 күн бұрын
28 күн бұрын
Ай бұрын
2 ай бұрын
2 ай бұрын
Пікірлер
@synesthesiam 5 сағат бұрын
Hi Michael, thanks for the review! I'm the main author of Assist in HA. Your points are fair for people expecting an out of the box Alexa-like experience. For people willing to add aliases, prune exposed devices, and explore things like LLM fallback and sentence triggers, things can get much better. The future of open source voice looks bright 🙂
@michaelsleen 5 сағат бұрын
Hey, thank you for all your work on this. Your contribution to voice control is huge. I agree that the experience is much improved when those additional steps are taken. I’m pumped about the potential of what you’re building here.
@patriciahaug1335 13 сағат бұрын
I wish I had seen this before Christmas.
@michaelsleen 13 сағат бұрын
It is still available to order if you want to get ahead for next season! go.michaelsleen.com/apollo-h1-ornament
@gordon861 14 сағат бұрын
Even Google devices have problems with DO THIS AND THEN statements and need you to issue each command individually. I currently have three Google Minis around the house and they are mostly used for listening to the radio as you can sync them together so the sound stays good as you move around the house, and they are rarely used to turn on lights etc so I am not sure how much better these new devices would be. I do wonder though if an over looked feature of these may be that we finally have a real local audio announcer that works native in Home Assistant. I currently use light colours to warn me that the fridge is not closed, being able to have HA tell me that the door is open after a time and keep telling me would be better, but not sure it is £50 better.
@michaelsleen 13 сағат бұрын
You could create an automation like you describe for the fridge and a voice announcement. In fact, I've already been doing this with Echo and Sonos speakers through Home Assistant.
@FoghornWashington 20 сағат бұрын
Fantastic video :) Thanks for the thorough testing
@TatortWeb 20 сағат бұрын
The test is not really usable, you compare things and test things with Alexa hardware that are not really comparable. 1. for using Alexa you also have to take care you have exposed the right devices and having usable names. Assist like google devices can handle aliases to match names of devices easier. 2. Compare LLM with Local for speed is not working. If you want local speed you need beefy hardware as they said, 3. Using LLM and Local separate and giving them same questions is not correct, they are made for different things like LLM can answer things like distance to the sun what Assist can not do 4. An advantage of Assist is you can combine LLM and Assist so Llm can be used as fallback if Assist can not handle it 5. for a full review you should also use Home Assistant cloud for Assist what Is way faster then your local Assist on your used hardware. Don´t get me wrong I am not a fanboy, I watch a lot of videos about this hardware and using different voice assists like google, Siri, Alexa since years. But you testing is giving a wrong view on this devices. Yes it needs more work like handling several timers at same time and other things. But if you use it with Home assistant cloud of really beefy hardware and use an LLM as fallback you test will look way different I think, if you expose your devices and name them as you would do for Alexa.
@michaelsleen 13 сағат бұрын
Hey, thanks for watching. Viewers specifically asked to show how this handles the same requests on local Assist vs. a cloud LLM. People want to see the variances in response times and ability to answer the prompt correctly. It's quite apparent to me that local Assist and a cloud LLM are built to handle different tasks. While it would be interesting to see the performance using HA Cloud, not everyone is going to become an HA Cloud subscriber. I specifically stated at the start of the testing in the video that I wanted to demonstrate the "out-of-the-box" experience. Not everyone will take the time to clean up their device and entity names, room locations, etc., and I didn't do anything to expose more or less entities than what was enabled by default. I call out in the "Tips" section of the video each of these nuances, and how it would improve the results.
@TatortWeb 12 сағат бұрын
@michaelsleen If users want to see LLM vs local you should have used sentences both can handle. Not like distance to sun what assist can not. Then you could show wich one is faster. And having cloud for speed would be an extra comparison. The local assist speed is not that faster but it is private.
@DustinMillerPolyInnovator Күн бұрын
Don't use balenaetcher. it can ruin your flashdrives. Its killed a few of mine tbh. Rufus is much safer
@TvistoProPro Күн бұрын
I like the review overall, but... I think you missed some key things. One of the main problems I see with your review is there are lots of things going on with your personal setup which are influencing the experience. One example? You're exposing TONS of entities to the agent. I'm a tinkerer, and have nearly 70 devices in my home, which most would consider a lot. By default my system started with "only" 103 exposed entities, which I've whittled down to about 80. Most agents/LLM get slower and can be confused by having too many entities exposed, which is why by default only key types were set to auto-expose. Most aren't going to have 350+ exposed entities by default, like you do, which is heavily impacting your experience. There are also lots of things you didn't cover, or missed in your review. For example, you can setup aliases just for voice for devices with non-relatable names. (No need to rename them & screw up automations/scripts.) You are correct in saying to be selective of what you expose, and how things are exposed (lights vs switches), but then fail to consider basic things like device groups (like "bedroom lights"). There's also a tick-box option to have queries sent to both the LLM and the local HA scripting model (which is NOT an LLM). This allows the local handler, which is far quicker and simpler, to handle the request quickly if it can figure out the request before the LLM replies. Kind of a "best of both worlds" option, which you completely skip over. And hardware wise: you note the switch, and complain about the speaker being weak, but then completely skip the audio out jack on the device? (Another missed topic: audio streaming, which is viable with the audio out port). Another option you didn't touch on is running LLM models locally. Yes, it's extra setup, but most folks are likely running HA for LOCAL CONTROLL, which is a key selling point of this device. Models like Ollama 3.1 can run locally (even on a separate local system if you 're on a Pi or Yellow) to handle LLM requests locally and quicker. Even significantly older GPU hardware, which many will already have or could be cheaply obtained (under $50), will run reasonably good LLM models. A good chunk of your delay with Gemini is likely the exposed items and hoisting that all up to the internet every time. This is a brand new product, with no configuration, little optimization, and a team of volunteers putting it out. It isn't going to match or exceed a polished 10+ year old product, with a full time staff of hundreds, and the backing of a billion dollar company. I don't think anyone would expect that, and it's not what it was put out to do. It's put out to be a good "first step" toward a local HA based assistant. The hardware is quite good, and on par with an early Gen1 pod. Echo canceling, an OK speaker, local wake word, and fully programable light ring and input buttons. That's pretty nice for the price. Since 99% of the future features will happen in software, this Gen 1 hardware device will likely remain relevant well into the future. Given that it can be operated via a service (like Gemini) or 100% locally, it has a strong chance at far out living Google and Alexa pucks. Devices which could literally die at the whim of Google deciding to drop support (like DropCams, Stadia, Glass, Nest, Clips, etc). Or Amazon doing the same, or requiring a pay-per-month model in the future, which they've already done for some services, like Ring and other cam devices. Personally: I've been playing with this type of thing since before the year of voice even happened. I have a few ESP devices already setup, doing on device wake word, and am using a local LLM via a small PC with an M2000 GPU. It's not perfect, and I still have an Alexa or two setup for when I want to do something outside the range of the current LLMs capabilities (current news questions mainly). But 90% of what I use a voice assistant for, I can do with a "Hey Jarvis" now, and have it all done locally. And when it gets good enough that I've not called on Alexa for a year? Or a subscription service is required? Or Alexa just stops working? Those all get turned off, and/or scavenged for parts. For sure, any future devices I'll be buying are going to be local controlled assistants, like this or the S3-BOX.
@michaelsleen 12 сағат бұрын
Thanks for watching and your interest here. Not sure if you watched the whole video, because I touched on many of the things you said. I don't see my personal setup with tons of exposed entities as a "problem." In the video, I state that I'm intentionally testing the out-of-box experience since not everyone will take the time to clean up device and entity names, room locations, and exposure to Voice. And then after the test, I state how to improve the results, and note these things.
@TvistoProPro 10 сағат бұрын
@@michaelsleen I did watch the full video, and as I noted, I found it rather reasonable over all (minus the TL;DW conclusion at the start). And yes, on watching a second time I see you did mention some of the items I listed (like the audio jack). But often those were glancing mentions, with little context, which were easily forgettable. Considering the top 3 items most people use personal assistants for are device control, timers, and streaming, the extreme focus on device control and query times seemed a bit unbalanced. Given that focus, not even mentioning parallel handling and down-playing entity count seemed especially odd. I get that you're trying to "be an end user" here, but I feel your attempt is instead presenting a narrative that applies to no one. If someone is new to HA, and prone to thinking a new device like this will be "plug-n-play", they're not likely to have 200+ devices. If they've had HA long enough to have a build-out that large, they're pretty much guaranteed to know something like this will need some level of tweaking to optimize. By exposing 388 entities (as shown in the video), and not doing at least a little mitigation, you're presenting an unrealistic setup. As a result, some claims you make are dubious at best; like having a "typical response time" of several seconds. That's very likely driven by the raw number of devices you have exposed. Simple test: Turn off device exposure (one switch) and time some of the same non-device related queries. With 100 entities, my local (underpowered) LLM took 2 to 5 seconds to reply to a query, about 2 to 3 seconds of which saw high GPU use. When I turned off exposure (0 devices), those same queries took under 2 seconds, with generally under a second of GPU use. At 50 exposures, I'm at 1 to 3 seconds, with half that time seeing GPU usage. Device exposure impacts response time at least linearly, at best. Realize that with EACH QUERY, those 388 exposures are converted to JSON with current state, room, labels, etc, and uploaded to the LLM along with the parameters and query. That entire chunk is then fully parsed by the LLM each time. Just the encode/transmit/parse time alone on that number of entries is likely taking a second or two, even with something as powerful as Gemini. Part of why the local handler is faster is it's just doing regex against the V2TXT string. Enabling that in parallel is a key item, as it can give the best of both worlds. A properly trained user/query can see commands processed on the CPU quickly, while more complex requests can go to the LLM which is generally off-device, or at least more GPU based. (And yes, you may not realize it now, but Alexa/Google have trained you to phrase things in a way it prefers over the past decade.) At a minimum, that's the biggest point I want to convey here: With minor tweaking you can get 90% of the way toward replacing a 10+ year old, well polished product by a major corporation for about the same cost, all locally controlled. That's a HUGE selling point for most HA users, which I feel this review either missed or downplayed at best.
@janetmorris6792 Күн бұрын
I can't find a good easy to read and learn how to set up most of the features of home assistant BOOK, the only listed one is in German, anybody out there with a reference // Thanks
@michaelsleen Күн бұрын
I'm not sure what you mean by Home Assistant "book."
@janetmorris6792 14 сағат бұрын
@@michaelsleen It has a cover and back and full of paper pages with all the help you need in a slow methodical approach to setting up the different parts to build up your own display, I find all the KZbin vid's too fast, Can' find any printed books out there to buy
@michaelsleen 14 сағат бұрын
@ in case it helps, all of my long form videos also have a written article in the video description. That will allow you to go at your own pace.
@Ballissle Күн бұрын
I think what people are missing a lot in regards to being able to answer a lot of questions, especially with a cloud LLM, is that it requires some backend work to expose what you want and also to name things properly and organise the entities well. For example, asking the weather, mine works perfectly when I ask the weather. that is because i exposed the weather entity built into home assistant to Assist. It was not by default exposed.
@michaelsleen Күн бұрын
Yes, that's what I shared in the Tips section of the video: To get the most out of it, you need to take the time to review the exposed entities, organize things in rooms, and write device names in natural language.
@Ballissle Күн бұрын
@ indeed. Most of the other reviewers I’ve seen so far have failed to mention that though and end up putting people off the device. Your review is excellent, although I think that could have been mentioned a bit sooner before the test portion for better clarity. Just an observation. For an early experimental device, I think it’s great already and already serving my particular needs very well when paired with nabu casa cloud processing and ChatGPT LLM (for now until I can get a dedicated LLM machine as my N95 processor takes about a full minute to get a response from Ollama with tinyllama model). I just need to control my entities, speak the main weather stats and answer informational and historical questions, in a timely manner. It’s doing that better than my Alexa’s currently. Plus I love the programmable button and light ring.
@michaelsleen Күн бұрын
@Ballissle Appreciate the feedback and good to hear about your experience with it, thanks.
@222inverter Күн бұрын
Home assistant "Assist"..is not a LLM???..and it shouldn't be compared as such in this video???..Assist in Home Assistant is a smart home assistant for your smart home ONLY!!!.......if the want to link up a LLM..eg chatgpt then of course except slight delay in responding!..I think people who are expecting this to be a drop in replacement for Alexa have got this all wrong!..it will improve over time with work so credit where is due!
@michaelsleen Күн бұрын
Not sure if you watched the whole video. I said it will only get better over time, and people are curious how Assist performs locally vs a cloud LLM.
@Sledge-fu5zr Күн бұрын
I have serious concerns with respect to commercial voice assistant privacy. Home assistant voice running locally is much more secure especially when connected to local llm such as ollama. However this requires a powerful computer for home assistant (example I7) and a very high end computer for llm system.
@michaelsleen Күн бұрын
Yes, I believe many share that concern, and I look forward to local options improving and becoming more accessible.
@Sledge-fu5zr Күн бұрын
@@michaelsleen I look forward to local improvements also. I am sure Home Assistant will improve in the areas you covered. I do believe we will see the need for high end computing power including GPU computing only getting greater in order to be completely local both with local voice and local llm any time soon. LLM;s are only going to improve and have greater capabilities. Using any cloud based llm always opens up security concerns. Home Assistant being open source software should however increase the level of security as developers will identify security risks.
@rachelwooddk2485 Күн бұрын
Not sure how everyone else is getting it to work, as I get the message that the git hub link is invalid.😒
@michaelsleen Күн бұрын
Hey, this is the GitHub link in the video description. It works for me. github.com/brenner-tobias/addon-cloudflared?tab=readme-ov-file
@markp8418 Күн бұрын
I bought 2 pack of Sonoff SNZB-02D sensors with an LCD display - I didn't get as far as connecting the to Home Assistant because it was obvious that they were inaccurate, with them sat next to each other they both showed different temperatures and humidity readings (2C difference, 10% difference)! They belong in the bin - returned as not fit fo purpose
@michaelsleen Күн бұрын
Working great for me. Sorry to hear they didn’t work out for you.
@markp8418 Күн бұрын
@@michaelsleen perhaps I just got dodgy ones ...
@agnieszkasadowska1478 Күн бұрын
Thank you for the great review. The reviews I’ve seen so far were either full of praise or very critical. The device is not that bad. Thanks to you, it’s perfectly clear what it handles well and what it still struggles with.
@michaelsleen Күн бұрын
Thanks for sharing that!
@BadmashKing8 Күн бұрын
Makes sense. Having all speakers connected to their own dedicated bandwidth will most likely keep it working. Also, 2.4ghz is more stable than 5ghz.
@thisaccountiswhat5704 Күн бұрын
Very good review. Thank you.
@michaelsleen Күн бұрын
Thanks for watching!
@tomthompson1198 2 күн бұрын
is the app needed ones and then you can remove it? how many zones can it support? what is the distance and radious of the area it sees?
@michaelsleen 2 күн бұрын
I assume you mean the HLKRadar tool? It’s not necessary to download, and if you do, you can just delete it after configuring your zones.
@StjepanPuljko 2 күн бұрын
Testing results for Alexa and Google home for same sentence wold be great.
@michaelsleen 2 күн бұрын
Thanks for the feedback. I considered but the video was already getting quite long, and figured more people may already have an Alexa or Google Nest to test as well.
2 күн бұрын
Great review! What also helps a lot with device recognition is adding aliases, as sometimes you might refer to them by different names.
@michaelsleen 2 күн бұрын
Thanks!
@xStellar 2 күн бұрын
Thanks for your tutorial. I've switched from my old Raspberry pi3+ to Bmax B4 plus, and everything went well from first try. Now i'll be able to get frigate and more
@michaelsleen 2 күн бұрын
Glad it went well!
@harrismagnum04 2 күн бұрын
"PREVIEW EDITION"
@michaelsleen 2 күн бұрын
YES
@simonsayshomeassistant 2 күн бұрын
Cool you got it!
@michaelsleen 2 күн бұрын
It’s been fun to try out. I also just like saying, “Hey, Jarvis.”
@simonsayshomeassistant 2 күн бұрын
@@michaelsleen Yeh I am in the next shipment 20 January
@StjepanPuljko 2 күн бұрын
I guess most of the people having this device hates you saying "Hey Jarvis"
@simonsayshomeassistant Күн бұрын
@@StjepanPuljko I use OK Nabu on mine
@jhonattansouza 2 күн бұрын
nice work michael!
@michaelsleen 2 күн бұрын
Thanks!
@michaelsleen 2 күн бұрын
👉 Let me know your take on Home Assistant Voice Preview Edition, and anything else you’d like to know about it!
@Alexx.Designer 2 күн бұрын
Michael your reviews are amazing 🤩 Whenever you review, you do it with complete honesty so thank you but don't you think if you improve your thumbnails you can grab more attention I noticed that your thumbnail needs some improvement and I would love to help you to improve them I can give you a FREE thumbnail trial and if you like it I can make more for you Waiting for your response 😊
@michaelsleen 2 күн бұрын
Thanks, but I’m good for now.
@88Snipi88 2 күн бұрын
is it possible to send only a massage when the postman arrived?
@michaelsleen 2 күн бұрын
In the LLM description you can tell it to only comment if the mail carrier is seen (it correctly identifies ours every time). Then you would need some kind of condition to only proceed with the rest of the automation if the response variable for the LLM contains a response. I haven’t tried this exactly.
@MarkGhali 2 күн бұрын
Nice video Michael! Love the honesty up front! Also check out Brad futureproof homes his Sat1 is an awesome looking voice device as well
@AlanJones-vc9yr 2 күн бұрын
Hi. Great video and thank you. I am using a Ring doorbell camera. All works fine except the notifications via HA app on my iPhone are truncated, so only get the first 100 ish characters of the description. Help please TIA
@michaelsleen 2 күн бұрын
What you’re describing may just be the expected behavior. You can try telling the LLM in your prompt to limit the description to a certain number of characters. But if it’s too long, long pressing on the notification on iOS at least should reveal the full message and show a larger image of the snapshot photo. To my knowledge, there is nothing on the HA side that would limit the amount of text shown - this is enforced by Apple. See here for an example: drive.google.com/drive/folders/1KwET0jCuqH9LbVqm8VUQTJprmTYaZtus
@AlanJones-vc9yr 2 күн бұрын
@@michaelsleen Hi Thanks for the reply. The message is actually truncated prior to receipt by the HA/iPhone. It stops mid mid word. I have tried requesting a shorter response which reduces the characters but still truncates the message ??? Flummuxed :-(
@Dazdigo 2 күн бұрын
I wish home assistant still supported portainer. It is a pity that the Raspi ended up being nearly as expensive as a mini pc. The main requirment for me is that I want Home Assistant to run on an SSD. I am running bare metal, but should install debian with the supervised version of HA. I ended up with the GMKtec G3 with 8 gb with a 256 gb nvme.
@michaelsleen 2 күн бұрын
Yeah it’s too bad the Pi isn’t much cheaper by comparison.
@oxxide 3 күн бұрын
This is amazing, but my issue is it doesn't show the full resposne. I can see full responses for other notification with other apps, so i don't think its a phone issue. It cuts off, for example my last noti "The image shows a front porch with a brick wall, a door with a securit...." which is fixed by just making 2 notifications, for some reason though, the date and time isn't working for me, also have a g4 door bell and last_changed doesn't get a date on the automation but does when check it with templates.
@michaelsleen 2 күн бұрын
What you’re describing may just be the expected behavior. You can try telling the LLM in your prompt to limit the description to a certain number of characters. But if it’s too long, long pressing on the notification on iOS at least should reveal the full message and show a larger image of the snapshot photo. To my knowledge, there is nothing on the HA side that would limit the amount of text shown - this is enforced by Apple. See here for an example: drive.google.com/drive/folders/1KwET0jCuqH9LbVqm8VUQTJprmTYaZtus
@michaelsleen 3 күн бұрын
The official release of HA 2025.1 is now out and I wanted to highlight this note from the developers: "If you are using any custom solution for backups, they will continue to work today. Even with everything new, we’ve made sure to keep everything backward compatible."
@leightonevans1071 3 күн бұрын
How do I know what file path to use for a snap shot image? Can someone give a step by step guide? On a side note I find most HA videos assume learners know the steps in between!
@michaelsleen 3 күн бұрын
It's really up to you to decide. As an example, you could use the path /media/images/doorbell_snapshot.jpg for the camera.snapshot action, and then the path /media/local/images/doorbell_snapshot.jpg for the action that sends the image in a push notification to your phone. You can use something other than 'images' or 'doorbell_snapshot.jpg' if you prefer another name or organization structure.
@kendrickdsouza8460 3 күн бұрын
Hey, there is no add-on option for HA on container, so we cannot add anything from add-ons. What is the way out with cloudflare add on?
@michaelsleen 3 күн бұрын
I don’t run Docker, but this might help you: community.home-assistant.io/t/installing-ha-addon-cloudflare-on-docker-container/540039/4?u=msl2424
@ANTIH3RO_ 3 күн бұрын
The Beelink S12 pro is waiting at my house to begin this process over the weekend. I literally chose the exact same mini PC so I can follow your guide. Really appreciate all of the amazing content and help you provide the community! Looking forward to following your journey and trying to keep up :)
@michaelsleen 3 күн бұрын
Exciting, and good luck!
@masterthiefsr 3 күн бұрын
The Code for plotly Seems to bei deleted. Any Idea how i can geht this Code? Trying to configure my msr2 with it
@michaelsleen 3 күн бұрын
Got a new link for you here: go.michaelsleen.com/plotly-graph.
@TheRealDanielsan 4 күн бұрын
I can't beleive you pay walled the yaml...
@michaelsleen 3 күн бұрын
I share everything you need to know both in my video and in a written article on my website, all for free. It is not necessary to pay a small fee for the code, but for those who want it to be as quick and easy as possible, I make that option available, and I'm not the only one to do so. Thanks for watching.
@asrr62 4 күн бұрын
Just tell me how to pair the thermostats together troll.
@michaelsleen 3 күн бұрын
Pairing a Z-Wave thermostat should follow very similar steps to what I showed in the video.
@mdavidf 4 күн бұрын
I know Aqara "gatekeeps." Aqara end devices I've purchased from China typically have more features, using Home Assistant, than the same devices purchased in the U.S. For example, every Window/Door Sensor I've purchased from China shows me the humidity and temperature of the room, closet, etc. Whereas, the two I've purchased in the US only show Opened/Closed.
@michaelsleen 4 күн бұрын
Super interesting. Thanks for confirming my suspicion!
@Ben_Saffer 4 күн бұрын
Absolutely wild that we’re in 2025 and this is so hard to do
@michaelsleen 4 күн бұрын
100%
@ckbowen 4 күн бұрын
Finally! I’ve always said how nice it would be to get a smart alert for boiling water vs having to check it (and interrupt what I’m doing) every few minutes. Will definitely be looking into this.
@mutley247365 5 күн бұрын
hay, Ive followed this to the letter, first off Im not getting any image through to my phone (Samsung) but I am getting the AI generated response. is this possible for android devises?
@michaelsleen 5 күн бұрын
Yes, it's possible. The only Android nuance I'm aware of is the AI-generated response may be cut-off if it's too long. Check to see if the camera image is being saved to the folder path you put in your automation for generating the camera snapshot.
@DragonXVI 6 күн бұрын
Thank you so much … had been messing with this for a while … didn’t know about the add-on portion, thought it was the cloud flare integration.
@stepmback 6 күн бұрын
I got the mtr-1 and can connect to it via bluetooth but it only connects for about 30 seconds then disconnects. Tried rebooting and unplug and replug but keeps disconnecting. I am using an iphone with the hlkradar tool
@michaelsleen 6 күн бұрын
Hey, this should not be the case. My MTR-1 has been rock solid since I installed it 7 months ago. Did you connect it to Home Assistant using ESPHome over WiFi as shown in the video? The HLKRadar tool is to make zone creation easier, but should not impact the device connection.
@stepmback 6 күн бұрын
@@michaelsleen I have. I it connected to my IOT network and that is working just fine. It is only when using my iPhone and hlkradar tool for creating zones that there is a problem. I have the bluetooth enabled on the device and comes up in the list of devices and connects, but then within 15-30 seconds it disconnects. Wondering if there is some baud setting or something like that I need to tweak.
@RWPKRPLRSVIDS 6 күн бұрын
This is great, but this will not work over cellular data if you don't have an apple hub.
@michaelsleen 6 күн бұрын
A home hub such as an Apple TV is needed for remote access and device control when you are away from home.
@h3ld3rk1d 6 күн бұрын
Hello, Im trying send to telegram but no luck, can you send the response to telegram? How? Thnkx
@michaelsleen 6 күн бұрын
I do not use Telegram so cannot comment on it.
@h3ld3rk1d 6 күн бұрын
@@michaelsleen thnkx my friend.. i wil try find solution
@ElMono19585 7 күн бұрын
Why ratgdo over tailwind?
@michaelsleen 6 күн бұрын
Tailwind is certainly an option as mentioned in the video. I chose ratgdo since it was created by an independent developer in the Home Assistant community, which I wanted to support. Thanks for watching.
@KarmenD-ce5re 7 күн бұрын
I don't think that having an external way of power a lock if the battery goes dead is a smart idea because again when the lock fails you're not expecting it and what's a chance of you having an external power supply to power it that is why I want a physical key just keep it in your wallet or on your keychain and you're in guaranteed
@michaelsleen 6 күн бұрын
There are lots of smart lock options that include a mechanical key slot if that is your preference. It would not be too difficult to obtain a 9 V battery in the event of a need to power it externally to get back in your home. At the same time, it really should not be a surprise if the battery is low. The device itself alerts you, or you can use a home automation platform like Home Assistant to always have access to information about your battery level and send custom alerts at different percentages. Alexa can even notify you well in advance that it’s time to re-order batteries. In four years, I have never had a concern of being locked out unexpectedly from a battery failure.
@KarmenD-ce5re 6 күн бұрын
@michaelsleen I've been working with electronics long enough for many years they're not always reliable you can't always rely on battery status either apparently you're a newbie
@BrendanMacWade 7 күн бұрын
Amazing brick. But BYOC. Bring your own 240-Watt USB C cable.
@marcusagren2838 7 күн бұрын
On my Oneplus 9Pro Android, the LLM-text doesn't really fit in the notification if i also attach the snapshot. I only see "Here is a description of the image in one sentence..." No way to expand the notification so i can see the ful description.
@michaelsleen 7 күн бұрын
I’ve seen others with Android say something similar. This is not an issue on iOS, and I don’t have an Android phone.
@marcusagren2838 6 күн бұрын
@@michaelsleen Yeah, maybe a bug in the Android companion or os. For now, I split up the notification in two. One for the snapshot and one for the LLM.