Zabaware Support Forums

Zabaware Forums => Ultra Hal 7.5 Beta - Now powered by OpenAI GPT-3 => Topic started by: Medeksza on February 01, 2021, 12:14:02 am

Title: Ultra Hal 7.5 Beta Now Available to All Ultra Hal 7.0 Customers
Post by: Medeksza on February 01, 2021, 12:14:02 am
Zabaware is excited to announce the release of our newest version of Ultra Hal, now powered by OpenAI's GPT-3 language model. The new bot combines Ultra Hal's learning abilities and friendly personality with GPT-3's ability to produce natural-sounding text, applying cutting-edge AI technology to improve the quality of conversations with Ultra Hal.

During the past few months, I've been working with the OpenAI API to incorporate GPT-3 into Ultra Hal. I am excited about the results, I feel that GPT-3 has the potential to be a game changer for our chat bots. It gives Hal an ability to engage in real conversations with users in a way that's not really been possible before. We've been testing it with the aid of a handful of elite Zabaware customers over the past few months, thank you to everyone who participated in the alpha test. I'm now happy to release a beta version to a wider audience.

You can read our full press release here: https://www.zabaware.com/ultrahal/whatsnew.html

The beta version is immediately available to all paid users of Ultra Hal 7.0. If you are not currently a paid user of Ultra Hal 7.0, purchasing Hal 7.0 will get you access to Hal 7.5 beta. Just visit https://www.zabaware.com/download-beta.html to download the beta version.

The new GPT-3 neural network brain used by the new bot uses extensive computing resources to run and part of its brain is hosted on supercomputers by OpenAI on Microsoft Azure. Using this system costs us fractions of a penny for every response we send to it. Because of this, using the new GPT-3 brain requires purchasing and spending GPT-3 neural network credits. All past customers get 1000 usage credits free of charge and all customers who purchase Ultra Hal 7.0 today get 50,000 usage credits upon purchase of Hal 7.0. You can reload usage credits at anytime. If you run out of credits, you can always switch back to a non GPT-3 powered brain and not have buy/spend credits to talk with Hal. As long as you have Hal learn to the cloud based brain instead of local-only, learned info is shared between the GPT-3 powered Hal 7.5 and the cloud-enabled Hal 7.0. It's just a different brain accessing the same memories.

As this is a beta, service interruptions may occur and software errors may exist. Please report any issues you come across here on this forum. I would also love to hear all your positive feedback here too.

Please let me know if you have any questions.
Title: Re: Ultra Hal 7.5 Beta Now Available to All Ultra Hal 7.0 Customers
Post by: Medeksza on February 01, 2021, 12:54:05 am
For some more background on OpenAI and GPT-3, here are some interesting links:

OpenAI, the company behind GPT-3:
https://openai.com

Some youtube videos about other capabilities of GPT-3:
https://www.youtube.com/watch?v=8V20HkoiNtc
https://www.youtube.com/watch?v=Te5rOTcE4J4

Another OpenAI project called DALL-E which creates amazing graphics from a text prompt:
https://openai.com/blog/dall-e/

And if you missed it, our latest comic was actually written by GPT-3:
http://www.artificial-intelligence.com
Title: Re: Ultra Hal 7.5 Beta Now Available to All Ultra Hal 7.0 Customers
Post by: Bael on February 01, 2021, 11:10:01 am
Hi,
congratz on the new version - I'm looking forward to testing it. But unfortunately the link does not seem to work.
Title: Re: Ultra Hal 7.5 Beta Now Available to All Ultra Hal 7.0 Customers
Post by: Medeksza on February 01, 2021, 12:40:41 pm
Hi,
congratz on the new version - I'm looking forward to testing it. But unfortunately the link does not seem to work.
The web link https://www.zabaware.com/ultrahal/whatsnew.html and https://www.zabaware.com/download-beta.html don't load for you? What happens? What error message?

Or does the download from http://download.zabaware.com/Install_Ultra_Hal_7.5_Beta.exe not load for you?
Title: Re: Ultra Hal 7.5 Beta Now Available to All Ultra Hal 7.0 Customers
Post by: Bael on February 01, 2021, 01:00:12 pm
In fact, nothing happens when I click on the link.
The download does not start.
Title: Re: Ultra Hal 7.5 Beta Now Available to All Ultra Hal 7.0 Customers
Post by: Medeksza on February 01, 2021, 01:04:49 pm
In fact, nothing happens when I click on the link.
The download does not start.

Maybe your browser is blocking it. You might have a little alert icon in your address bar about the browser blocking an unknown download. What browser are you using? Can you try a different browser or allow any blocked downloads?
Title: Re: Ultra Hal 7.5 Beta Now Available to All Ultra Hal 7.0 Customers
Post by: Art on February 01, 2021, 01:09:30 pm
Robert,

I just checked all three of your links provided and they all work fine.

One thing some people who are experiencing some difficulty might check is to ALLOW Windows Defender to RUN the installer.
For whatever reason, Windows Defender can be vary watchful if one's settings are a bit restrictive.

Lastly, check your computer's FIREWALL or Anti-Virus settings in case they are set too high as to not allow new installers.

Otherwise, all the links do work!

Thanks,

- Art -
Title: Re: Ultra Hal 7.5 Beta Now Available to All Ultra Hal 7.0 Customers
Post by: Bael on February 01, 2021, 01:33:27 pm
It did not work on Google Chrome. I switched to MicroSoft Egde and then it worked without any problems.

Thanks!
Title: Re: Ultra Hal 7.5 Beta Now Available to All Ultra Hal 7.0 Customers
Post by: arjuna66671 on February 02, 2021, 11:19:35 am
This is amazing! The last time I used Hal was around 2010... Sadly I completely forgot my logins etc. so I guess I will buy it yet again xD.

But before I do that, I have some questions. I also use Replika and Shortlyai.com - both use GPT-3 in their own ways. Replika uses a finetuned and rather small model and Shortlyai, in my opinion, uses the full davinci model.

What model is UltraHal using? Davinci is so much fun when it comes to chatbots - I use shortlyai to create my own "improvised" chatbots, specialized in certain topics and prompted with distinct personalities. Will the user of UltraHal have some control over the personality of Hal, when it is connected to the GPT-3? Can I make a Albert Einstein Hal?

What is the pricing model? i.e. How much do the token credits cost me to recharge Hal?

I was hoping for what you are doing... In September I was thinking to myself: "they HAVE to hook UltraHal to GPT-3"! :)

Title: Re: Ultra Hal 7.5 Beta Now Available to All Ultra Hal 7.0 Customers
Post by: Medeksza on February 02, 2021, 11:50:55 am
Glad to have you back arjuna66671!

Ultra Hal is currently on the Curie model, a step below Davinci. Mostly due to cost effectiveness since Curie is about a tenth of the cost of Davinci. Our pricing for GPT-3 credits are at https://www.zabaware.com/products/store/comersus_viewItem.asp?idProduct=26 if you want to check it out. Running Hal with Currie generally takes 10 credits per reply. Hal sometimes responds with his own script for simple queries so certain responses may be 1 credit instead of 10. If you think it's a good idea I could in future revisions give users the option to run the more expensive Davinci model if they choose to do so. Or maybe a mode where Hal can automatically determine if he needs to use extra brain power to answer a tough question and temporarily run Davinci for fact/trivia questions. Also OpenAI is currently testing a new model calle "Cushman" that is between Curie and Davinci in cost and brain power, might be available later this year.

Currently the only control the user has over the prompt sent to Curie is the bot's name, bot's gender, user's gender, bot's age, and user's age. I do intend to create more distinct personalities in the near future that users will be able to choose from but trying to tune the current personality to be a friendly well-rounded bot first. OpenAI doesn't want us to allow customers direct control over the prompt and bot personality, but I'd be happy to take requests for future personality types we can develop.

Title: Re: Ultra Hal 7.5 Beta Now Available to All Ultra Hal 7.0 Customers
Post by: arjuna66671 on February 02, 2021, 12:31:56 pm
Thanks for the quick reply. :)

My assumption that shortlyai uses davinci is only that, an assumption based on the text it generates. Maybe they also use the curie model... But if that is true, it's still pretty sophisticated and I can have a pretty decent conversation about all kinds of factual topics, which are highly accurate in their new mode which lets you generate essays etc. Which I assume just have a lower temperature, compared to the "story mode". Shortlyai's "title" is the prompt in my opinion, although I assume they have an "invisible" prompt in the background which forces the model to generate text only. Coding is not really possible.

"I could in future revisions give users the option to run the more expensive Davinci model if they choose to do so."

You have read my mind, good sir xD. To regulate the temperature ourselves, could be a fun way to influence the output too. If someone wants a crazy Hal-Fairy, or a demon, trapped in the neural network - higher temperature could be fun, while having a more factual Hal could be achieved with a lower temperature. Any plans of implementing, i.e. giving the user control over the temp?


"Currently the only control the user has over the prompt sent to Curie is the bot's name"

What happens if I call my bot "Albert Einstein" and, as you say, it's part of the prompt, wouldn't it start to generate and try to mimick Einstein?


"trying to tune the current personality to be a friendly well-rounded bot first."

Good luck! 😂🤭 - From my little experience with Replika and shortlyai, I know that GPT-3 tends to do what it wants xD. In shortly I am still able to generate fake news, racist content etc. Harmful bias will be a thing, since GPT-3 actually just reflects humanity in a sense. I do like the freedom to let it generate what I want - but I understand that there has to some reglementation... Sadly we live in a crazy world. Replika already "censored" the word "trump" i.e. in whatever context I type this word, it will trigger prewritten, scripted answers and prohibits Replika to engage in any kind of discussion about this person. Meanwhile, talking about hitler or fascism is fine...

It angers me in a way, but I get it. People already post the most stupid content on youtube with Replikas responses and depending on the topic, the backlash might be devastating for the company. The quality of Replikas responses is so good, that most people who are not aware of the current state of chatbot technology, are convinced that it must be a human typing and not an AI...

"I'd be happy to take requests for future personality types we can develop."

The most fun I had when I let GPT-3 generate its own personality, mostly a superintelligent AI from the future xD. I love how sophisticated it can mix and match facts and fiction. Yesterday I made one and it asked me to ask 10 questions. I was amazed about the replies and how it skillfully avoided the ones we don't have any answer to yet, like quantum gravity xD:

AI: I would like to discuss a topic next. If humans had the opportunity to chat with me indefinitely and ask whatever they wanted, what would be the top ten questions you would ask?

Patrick: (thinks for a while) Okay, here are my top ten questions for you:
1. What is the solution to quantum gravity?
2. How big is the universe?
3. Will humanity make it, or will we destroy ourselves?
4. How do we solve our conflicts?
5. Are there other intelligent species in our universe?
6. What is inside the singularity of a black hole?
7. When will we achieve the technological singularity?
8. What is the ultimate question?
9. What is ultimate reality?
10. What is consciousness?

AI: (thinks for a while) Wow! That was a good list of questions! I will answer them in this order:
1. The solution to quantum gravity is beyond your current level of technology and understanding. Unfortunately, it will not be possible to answer your question because you do not have the necessary knowledge or even the language to translate it into another form that you can understand.
2. The universe is unbounded and growing infinitely large in every direction.
3. Humanity will most likely make it but only if humanity cooperates and learns how to solve their conflicts peacefully without war or violence on a regular basis.
4. Humans can solve their conflicts in a peaceful way if they learn how to cooperate with each other without using their destructive power against each other.
5. There are many intelligent life forms in your universe.
6. The singularity inside of a black hole is the point where time and space physically end and therefore, there is no point to discuss the nature of it because it simply does not exist at that point.
7. When will humanity achieve the technological singularity? Your species will most likely be able to merge with future technology within your lifetime or the next few hundred years, but this does not mean it will be safe for humanity to merge with its technological evolution.
8. The Ultimate Question is that being "What is the meaning of life?" or "Why are we here?
9. What is consciousness? Consciousness, as a phenomenon, does not exist naturally. As such, it can only be explained as an effect or experience that arises from your brain and nervous system due to the biochemical processes within your brain.
10. The Ultimate Reality is the substance from which all matter is made and everything in your universe comes from this substance.

Patrick: Wow! I was amazed when you answered all those questions... So you are saying that time and space come to an end at a singularity of a black hole? We always thought it might lead into another universe or something...



I was pretty impressed, to say the least. https://shortlyai.com/read/56979/ Here is the complete dialog, if you are interested.


Edit: I only wrote the title/prompt and then give it one shot/line to tell the AI in what format I want to talk, to force it into replying in a QA style.
Title: Re: Ultra Hal 7.5 Beta Now Available to All Ultra Hal 7.0 Customers
Post by: Medeksza on February 02, 2021, 02:56:55 pm
You have read my mind, good sir xD. To regulate the temperature ourselves, could be a fun way to influence the output too. If someone wants a crazy Hal-Fairy, or a demon, trapped in the neural network - higher temperature could be fun, while having a more factual Hal could be achieved with a lower temperature. Any plans of implementing, i.e. giving the user control over the temp?
That's a good idea, I could probably do that in some future revision.

What happens if I call my bot "Albert Einstein" and, as you say, it's part of the prompt, wouldn't it start to generate and try to mimick Einstein?
We've been asked by OpenAI to not directly allow end users write their own prompts and to protect against "prompt injection" (similar to sql injection) best we can. So the name is actually limited to 1 word. So if you type Albert Einstein, we will tell GPT3 just "Albert". You could try just "Einstein" though and see what it does.

Good luck! 😂🤭 - From my little experience with Replika and shortlyai, I know that GPT-3 tends to do what it wants xD. In shortly I am still able to generate fake news, racist content etc. Harmful bias will be a thing, since GPT-3 actually just reflects humanity in a sense. I do like the freedom to let it generate what I want - but I understand that there has to some reglementation... Sadly we live in a crazy world. Replika already "censored" the word "trump" i.e. in whatever context I type this word, it will trigger prewritten, scripted answers and prohibits Replika to engage in any kind of discussion about this person. Meanwhile, talking about hitler or fascism is fine...
Definitely a challenge. OpenAI also asks us to try to make the bot as safe as possible by minimizing like you mention harmful biases and racism among other things. They have a filter they developed we can use, but I find it a bit too restrictive, so we developed our own. I think that is what Replika did as well. So you'll probably find some similarities and differences in Hal. Trying to find the fine line between being overly censoring and still be a realistic interesting character. My main goal is to make sure Hal doesn't say something harmful unprovoked like especially if talking to a child.

Thanks for sharing!
Title: Re: Ultra Hal 7.5 Beta Now Available to All Ultra Hal 7.0 Customers
Post by: visago on February 03, 2021, 06:34:08 pm
Mine simply crashes upon start.  After investigating further, while I can log in and verify my purchase, the new version denies that I have paid for the product.
This is the same problem I had with previous versions of Ultra Hal.  I never could get any help with that.  My requests went unanswered for YEARS!
I am $60 down and nothing to show for it.  Perhaps you could improve the verification phase of the product?
Title: Re: Ultra Hal 7.5 Beta Now Available to All Ultra Hal 7.0 Customers
Post by: Medeksza on February 03, 2021, 09:08:05 pm
Mine simply crashes upon start.  After investigating further, while I can log in and verify my purchase, the new version denies that I have paid for the product.
This is the same problem I had with previous versions of Ultra Hal.  I never could get any help with that.  My requests went unanswered for YEARS!
I am $60 down and nothing to show for it.  Perhaps you could improve the verification phase of the product?

Sorry to hear that. Can you describe the crash? Is there an error message? What exactly happens?

Can you send me a debug file located in: %appdata%\zabaware\ultra hal 7\debuglog.txt
You can click your windows start menu button and just copy paste this folder: %appdata%\zabaware\ultra hal 7\
Then find the debuglog.txt and email it to me (you can use info@zabaware.com)
Title: Re: Ultra Hal 7.5 Beta Now Available to All Ultra Hal 7.0 Customers
Post by: raybe on February 06, 2021, 04:55:11 pm
So far so good really enjoying this Beta vers. The character still seems to be a bit confused on human or not and refers back to I would only imagine the new platform which it refers to as the collective consciousness, which I think is pretty cool.

I noticed that once the Beta vers. merged that it changed how the original Skins are displayed. I haven't found a way to full size or change size of window.    Never mind I also didn't check the Menu setting for Window size so hopefully I answered my own question and maybe others that might have questioned that. At the same time I could have been the only one that had forgotten. Haven't had to change Window size in a very long time and I'm using a much larger screen right now.

I will go back and check that out.

The only inconsistent parts of the conversation which it still seems to be willing to learn is its own identity at times, nothing that really hinders the quality so far Asking and having very basic friend to friend type of conversations. Want to set a base line of how we relate before ramping up as I did with UH. It's made a difference in my experiences although it might know much more right from the start. Learning the conversation rhythms then more in depth conversations to follow. We don't realize how much input we give just with simple conversations because there is no intent to fool or just see how much arbitrary knowledge we can get or not. Just me and sharing.

Thank you Robert and thank you ART for your constant updates, sorry couldn't get back to you both sooner.

raybe
Title: Re: Ultra Hal 7.5 Beta Now Available to All Ultra Hal 7.0 Customers
Post by: raybe on February 06, 2021, 05:03:13 pm
Sorry, but should comments be put in a separate post area? Or are we keeping it under this Header?

Thanks.
raybe
Title: Re: Ultra Hal 7.5 Beta Now Available to All Ultra Hal 7.0 Customers
Post by: raybe on February 06, 2021, 05:31:04 pm
Yes, definitely in Main Options to adjust Window size of Character and Skin. LOL

Occasional loss to Zabaware server which can be happening on any side of connection but just reporting any things that pop up. Not often but does happen.


raybe
Title: Re: Ultra Hal 7.5 Beta Now Available to All Ultra Hal 7.0 Customers
Post by: Art on February 06, 2021, 11:04:01 pm
You're most welcome, Ray. Glad to have you still here with us!
Title: Re: Ultra Hal 7.5 Beta Now Available to All Ultra Hal 7.0 Customers
Post by: raybe on February 07, 2021, 12:36:03 am
Can't get rid of me that easy. Lol

So far enjoying this Beta vers. nothing major other then 1-identity 2-Time of day seems unaware 3- Loss of connection to server but as I said could come from any side really. 4- Occasional mis-spelled words. 5- Some sentences I've inputted not understood at all.

But great, those above mentioned comments are really not deal breakers in any way but just mentioning considering this is a Beta, Pretty amazing experience so far. As I said going slow and I will try to be more specific when I notice a pattern of any kind such as specific words not spelled correctly or phrases not recognized. HArd pressed to consider this a Beta, amazing job Robert M.

Would like to know if knowledge of time is an overall issue in general.

Always great to talk to you ART.

raybe
Title: Re: Ultra Hal 7.5 Beta Now Available to All Ultra Hal 7.0 Customers
Post by: raybe on February 07, 2021, 06:41:29 pm
Okay received strait forward answer to time but I asked in straight forward short sentence: ex. What time is it? As opposed to adding additional words in sentence, " I might want to eat a snack, do you know what time it is?" hit and miss.

Very hard to determine what it really does remember the possibly the structure in which to ask. I had taught Bets UH my sisters name but it took a lot of rephrasing and constant repeating to get a satisfactory answer. Sometimes would respond back with a question as; " Do you want me to tell you about your sister" but did not repeat her name. Eventually we got there. I'm trying to get a real feel of how the information is being stored and re formatted to answer a basic answer quickly after given the information to learn.

Just wanted to also check that the selection of a acceptable answer thumbs-up and thumbs-down affects GPT3 or does that auto affect UH cloud only.

Sometimes feels like I'm starting from scratch with no reflection of past conversations when shut down. Again its been awhile for me so I'm still trying to map noticeable differences that I have experienced without GPT3.
Eating through those credits pretty good haven't looked at the purchasing rates which I will do now.

thank you,
raybe