"Hey Neon, Talk to ChatGPT" - Mark II Voice Assistant - Update May 2023

New Neon AI Voice Assistant Update For Your Mycroft Mark II Includes Pre-Release ChatGPT Skills!

Neon AI has released the latest Neon AI Voice Assistant for the Mycroft Mark II (and other platforms), and in the newest pre-release version a skill which enables Neon AI users to converse with ChatGPT, simply by saying “Hey Neon, I want to talk to ChatGPT”.

To get started, tell your Mycroft Mark II device:

  1. “Hey Neon, enable pre-release updates”
  2. “Hey Neon, check for updates” and “Yes” to update
  3. “Hey Neon, update configuration”
  4. Once it’s finished updating, you’ll be able to say “Hey Neon, talk to ChatGPT”

    Neon AI is the most advanced Open Source conversational personal assistant.
    The Neon AI private voice assistant now responds to hundreds of natural language commands (see below).

Upgrade Offer - Serious Storage! Free to Mark II Community Member Contributors
Mark II’s have their best performance with an add-on 256GB SSD memory. Visit the Neon Square.com shop and use coupon code “MycroftForumReader” to order one for the discount price of $29 with free shipping*. Active community contributors can receive one for free - email me for a code ~Clary.
*Limited offer. Orders will be mailed out in about 4 weeks.


“The speed increase of running Neon AI OS off the SSD instead of a USB stick is immediately noticeable.” ~Mike S.

Hints

  • “Hey Neon, Increase the volume”
  • Change the homescreen wallpaper in the system menu.
  • Swipe down for the menu, choose Settings → Display → Wallpaper Settings
  • “Hey Neon, Talk to me faster”
  • “Hey Neon, Change my units to metric”
  • To leave the pre-release updates track, say: “Hey Neon, Enable stable update”.
  • “Hey Neon, Create a support ticket” to generate and email a log of recent activity.
  • If you need to, you can say: “Hey Neon, Restart"

Resources
*Quick Reference & Quick Start guides.
*GitHub documentation at: About - Neon.AI® Documentation
*Advanced Neon developer updates are posted on our GitHub Discussions page

Please visit our Matrix community to continue the conversation and read more about the Neon OS.

Wishing you success with your private personal digital assistant,

Clary Gasper
Neon AI Projects
Clary@Neon.AI

====================================
Community Collaboration

Community Testers wanted: Visit the new Matrix chat room for testing requests. Developers working on advanced skills for the Neon OS will benefit from your testing and commenting on new features before release. To run advanced versions you can say: “Hey Neon, Enable pre-release updates”.

Community Developers wanted: Bounties available for skills, applications and documentation development. The Neon AI team is available to assist community developers.

Contact clary@neon.ai with your ideas and suggestions.

====================================
Need a Mark II?
We’ve confirmed this morning with Michael Lewis that MycroftAI is still fulfilling all the orders placed here on their website. He’ll be updating the order page soon, and you can also order one from us here on Square.
We welcome proposals from content creators, reviewers, and developers for projects that you would like to take on with a Mark II as a “bounty” from us.

3 Likes

Thanks this looks great. It co-incided with the Neon SSD arriving in the post. Also great - it just works. I’m definitely up for tinkering and experimenting but later when I have time.

Looking forward to trying the Chat GPT skill out properly. I’ve tried it a couple of times and it’s said ‘chat GPT is unavailable’

1 Like

So glad you’re happy! Did you remember to say “Hey Neon, update configuration”? :slight_smile:

I did “Hey Neon, update configuration” again to make sure.

Still getting the same

  • “Hey Neon, talk to chat GPT”
  • screen notification says ‘you are now talking to an LLM’
  • “Hey Neon … some random question …?” and the response is “ChatGP is unavailable right now”

I left it for a few minutes and after 10 mins or so it said “goodbye”

Is there something else I can try?

Thanks

It’s set to time out after 5 minutes, so that people don’t accidentally leave ChatGPT running.

Not working in the first place is strange. Maybe an internet connection issue of some sort? But if it was, I’d expect other API-based skills to fail as well, and red lights to appear on top of your Mark II.

Can you please reboot, then if trying to talk to ChatGPT still fails, please have Neon “Create a troubleshooting package” and for the subject say something like “ChatGPT not available report by Fromport on Mycroft forum”. That will go to @NeonDaniel to look at, and we’ll be able to connect it to your question successfully.

do I need a ChatGPT account for this?

1 Like

You don’t need an account :slight_smile: we are using our own Neon AI API to access ChatGPT. Your queries are passed to ChatGPT by Neon, and the answers returned to you. Though the actual content of what you say to ChatGPT should be treated as “public” because we have no control over what OpenAI does with it, we don’t pass any of your other information along. It is similar to the process we use to query WolframAlpha.

1 Like

that is good news, so at least there is not any profiling due to an specific account. Of course, how the content is used by OpenAI is another story.

1 Like

Thanks and sorry for not responding sooner.

It says it’s emailed a report to support.

1 Like

This sounds just great, I’m eager to try it out to speak with ChatGPT :slight_smile: Right now I am having trouble with the update to the pre-release version though. I try to update from version 23.5.11 to version 23.5.19 alpha13. At the end of the update process, when he is verifying the update, I get an “Update failed!” message on the display. Any idea @NeonClary ?

I have tried to “Hey Neon, update configuration” also at version 23.5.11 to see if that made a difference, but it didn’t.

I’m so pleased to hear you are wanting to try it out, and that’s very frustrating that it isn’t working for you.

I suggest updating your current stable version, even though it doesn’t need it, because that may fix any errors that have cropped up. Alternately, if you’re comfortable re-imaging your drive you can try that instead. Then retry switching to pre-release updates, installing that update and updating your configuration. Daniel put a new beta release out, so it will be 23.05.25a17 now. Please let me know if it still doesn’t work.

Are you on an SSD or a USB, and is it one I sent?

Thanks for the quick guidance, @NeonClary !

I came further this time. I downloaded the latest image from Download the Neon AI Personal Assistant and Network Services and Applications | Neon AI, flashed it with the Raspberry Pi Imager (I use the Neon USB-stick that came with the Mark II), and proceeded with the upgrade process as you suggested. And I got an “Update Success!” message this time. But then it got stuck after the automatic reboot showing the Neon AI logo (I gave it an hour to try to complete the reboot). The same happened after a new “hard” reboot attempt (cut power). Can I get out some error log or similar to help with the debugging?

1 Like

Nice work!
If you tell Neon your email address - “My email address is ____” and then “create a troubleshooting package” or “create a support ticket” it will email you a log. It doesn’t have everything, but it’s a good place to start.
If you need more than that, if you’re able to come visit Matrix someone may be able to help you with accessing your device via SSH or other more technical options. https://matrix.to/#/#NeonMycroft:matrix.org

Trying the same things on a different boot drive would be a good way to make sure it isn’t a hardware issue with the boot drive.

1 Like

Can you share what kind of SSD is used?

Adding international shipping costs I’m close to premium Samsung SSDs?

We haven’t been brand loyal. I’ve been testing various SSD drives, both their in-action performance and with Crystal DiskMark.

The most interesting thing I’ve learned is that the main focus of SSD development for general sales has been on raising the highest read/write numbers, which are driven more by large file processing than small. Our Neon OS’s operation, as well as OVOS’, is reading and writing small and very small files for the most part. So while we see a great improvement in performance moving from USB to SSD, we haven’t seen a big payoff in moving to SSD with higher max read/write speeds.

We have a supplier who has found some potential chips for us which are better optimized for small files. A sample is on the way to me now. I’m very interested to try them out. They would cost more than what we’ve ordered previously, but less than the premium SSDs, and we hope will perform as well or better for our specific use case than those SSDs that boast the highest max read/write speeds.

I have a Samsung T7 on my desk right now for testing, and my first timed tests indicate the Neon OS might be answering queries a half second faster than the ones we’ve been shipping. So you’ve got an excellent SSD, and I don’t have anything better to offer you at this moment, but perhaps in a couple of months I will.

1 Like

I have bought several differen aftermarket ssd’s that I am using with my MK2.
If you tell me how you test, more than happy to share my test results.

1 Like

thx for trying it out, so the conversation could feel a bit more natural :slight_smile: but I need to solve link to matrix thread first:

looking through the logs and I see some unexpected warnings/errors around file locking. "

1 Like

Thanks! We’re using CrystalDiskMark, and I set the parameters to these:
image
Input is always welcome :slight_smile:

I don’t run/use windows so I can not provide you with CDM results.
I can provide you the results I ran on the MK2 itself with the “fio” (flexible I/O tester) package installed

https://termbin.com/716d

1 Like