Print

[IIO]

This is on the cards, I'm surprised at how much usage it's had over the last three days.

if you´d split it up into two different proxies you could see how much of that is OS9 vs OSX.

not sure about the legal and technical aspects of the pro accounts but i think as long as you only ask for donations to cover the cost noone will care. in the enterprise field they also only buy 2 accounts for 100 clients, so why not in associations.

[Greystash]

This is on the cards, I'm surprised at how much usage it's had over the last three days.

if you´d split it up into two different proxies you could see how much of that is OS9 vs OSX.

not sure about the legal and technical aspects of the pro accounts but i think as long as you only ask for donations to cover the cost noone will care. in the enterprise field they also only buy 2 accounts for 100 clients, so why not in associations.

That's a good idea, usage has dropped off a bit now which is good. If it stays at this level it will be manageable for a while with the donation amounts that have come in so far

[Greystash]

Update: I've just released a 68k version which is compatible with 68020 CPUs and above. The app will now run from a 68k Macintosh on System 7 to El Capitan 

[Knezzen]

Great work! Love it!

[ssp3]

Meanwhile in the news..
https://abcnews.go.com/US/wireStory/chatgpts-chief-testify-congress-concerns-grow-artificial-intelligences-99354486

ChatGPT's chief to testify before Congress as concerns grow about artificial intelligence's risks
...
What started out as a panic among educators about ChatGPT's use to cheat on homework assignments has expanded to broader concerns about the ability of the latest crop of “generative AI” tools to mislead people, spread falsehoods, violate copyright protections and upend some jobs.

[IIO]

dont worry about machines, nothing is more intelligent than the "person, woman, man, camera, tv" guy himself.

[ovalking]

Update: I've just released a 68k version which is compatible with 68020 CPUs and above. The app will now run from a 68k Macintosh on System 7 to El Capitan 

Thanks for the great work.
More people need to tell it about LegacyAI, because it recommends I update my OS if I want a client!

[macarone]

Whatever one thinks about AI in general, and ChatGPT in particular, this application is  truly amazing.

I had it up and running in seconds. It just works and does what is promised and intended on a hurge range of CPUs.

Realizes latest technology in the way that attracted us to Macs. Truly rare these days.

T H A N K S  ! ! !

[Mat]

The last 2 days I am always getting an "Connection Error" after the first answer I recived. My internet connection is ok and as soon as I close the program and open it again I can ask one other question
I am using v. 1.3 at a Ti-Book with 9.2.2, … anybody else with similar behaviours? To be crystal clear, it worked well last week, even v. 1.3.

[Greystash]

I've noticed the same issue Mat. This coincided with OpenAI making lots of changes to the API last week which has resulted in slower response times. The problem is that Heroku (where I host the proxy server) has a 30s request timeout that you can't change . To solve this I could move the server to another service, but I'm not sure whether it's worth the increased cost. If it becomes a big problem I'll think about it more.

For now, you can just resubmit your question and most times it will respond within the 30s timeout. I'll probably implement some retry logic with v2.1 but that's still not a solid fix.

[Greystash]

I've just release version 2.0 which includes a bunch of new features and bug fixes detailed here.

[rvense]

What's Heroku costing you? I think there are cheaper, speedier options if you're OK with doing more work yourself.

[Mat]

The version 2.0 is working great! Thanks for implementing the resizing of the window. I also have no connection problems anymmore and Umlauts are working well. Only very minor issue; "Speak Responses" is activated by default in the Settings, what leaded to a general "Error 3" and crash of the program at my 9.2.2 computer. Perhaps some people might think there is an serious issue, while everything works well after deactivating "Speak Responses". I am using the PPC version, no possibiliy to test the other versions.

A huge "Thank You!" for this nice software.

[Greystash]

Thanks for the feedback Mat! I've addressed all of these issues with the v2.1 release, which comes with many new features. I think there can be issues if your speech setup (extensions etc.) isn't setup as a clean install would be, I haven't really looked into this yet though.

rvense:Heroku is costing $8/month, it's pretty lousy for what It offers. I've just introduced a premium server which is available to registered users in v2.1, which now has a 5 minute request timeout, and has much better performance. This has solved all of the network issues the previous versions were experiencing. It doubles the running costs so that's why I've had to implement licensing/a small fee to use it.

[chrisNova777]

The ai overhype is out of control 

[MacTron]

This allow a great step in to the 21st century to our beloved Mac Os 9 desktops. 
Even today we can't measure the impackt of ChatGPT. Awesome, thank you.

[Greystash]

No worries!

The details in this post are a little out of date now, so as a quick update:
- GPT 3.5 is 100% free to use
- GPT 4 and image generation require a registration
- No monthly usage fees, just a one time registration if you want to use the better AI models

[Protools5LEGuy]

Great!