Bob, impressive beta! I talked with it about a few things I know pretty well and have had to explain to others. I noticed it sometimes reads the markdown literally ("hash"). Other verbal tics will be more tricky. Some technical terms have slang pronunciation the speech engine wouldn't know about. It got SCSI correct about half the time, and started off with a great deconstruction of the term. C64 "SID" and "VIC" chips it always initialized. TRS-80 was "eight-zero", etc. Maybe a table? The voice and inflection are very good. I noticed that for longer responses it starts speaking slowly and sped up as it went until it was quite fast. Speaking of fast, the responses to questions began very quickly after my speech stopped. This is great, as many voice models have a noticeable processing delay before responding. This felt more naturally conversational. It handled interruptions well. A simple prompt injection failed, great! I didn't try hard. I asked for exhibit information responses in a few languages which it did well. The speech model had a noticeable American accent! Curiously it responded immediately in Spanish, but took about 20 seconds to start in Japanese. I asked a question in Dutch but it didn't understand so English inputs only at this time? I'm wondering if you've got the LLM to prefer local knowledge you've given it before going to its training or the web. A couple of times it stopped speaking in the middle of a sentence. I don't know if it thought it detected a speech interruption. Made me consider when this is used on a phone speaker in the museum with other unrelated speakers around it. I think this will be a great addition to the museum.   Regards, Ralph Broom       On 2026-01-02 07:00, vcf-midatlantic-request@lists.vcfed.org wrote:

Send vcf-midatlantic mailing list submissions to vcf-midatlantic@lists.vcfed.org

To subscribe or unsubscribe via the World Wide Web, visit https://lists.vcfed.org/mailman/listinfo/vcf-midatlantic or, via email, send a message with subject or body 'help' to vcf-midatlantic-request@lists.vcfed.org

You can reach the person managing the list at vcf-midatlantic-owner@lists.vcfed.org

When replying, please edit your Subject line so it is more specific than "Re: Contents of vcf-midatlantic digest..."

Today's Topics:

1. AI Computer Museum Docent - BETA - DEMO (Bob Roswell) 2. who was first ... (Jeffrey Jonas)

----------------------------------------------------------------------

Message: 1 Date: Fri, 2 Jan 2026 02:28:06 +0000 From: Bob Roswell <broswell@syssrc.com> To: vcf-midatlantic <vcf-midatlantic@lists.vcfed.org> Subject: [vcf-midatlantic] AI Computer Museum Docent - BETA - DEMO Message-ID: <CY4PR0401MB3586D6F293D1175602B08EF7B0BBA@CY4PR0401MB3586.namprd04.prod.outlook.com>

Content-Type: text/plain; charset="us-ascii"

I wrote a voice chatbot that is somewhat guided to be a knowledgeable docent at our computer museum. This is available for testing for the next few days: (If it does not crash)

https://museum.syssrc.com/docent

Privacy - You will need to give microphone and video permissions. The video is not in use, but I have not figured out how to make the WEBRTC component not ask for camera privileges. Your voice will NOT be saved, but I will have a speech to text transcript of your speaking, and the answers the AI gave. I won't know who you are except your IP address will show up in my logs.

I'd love any feedback! So far, I am amazed at what it can do, but remember it is an AI, and can and does make stuff up.

Tech Stack: Pipecat (Framework) DeepGram: Text To Speach Cartesia: Speech To Text OpenAI: Large Language Model Python scripts that I wrote based on Pipecat examples

I'll also be curious to see what my bills are! If it goes crazy, I will need to shut it off!

Bob Roswell broswell@syssrc.com 410-771-5544 ext 4336

Computer Museum Highlights<http://museum.syssrc.com/>

[cid:image001.png@01DC7B64.5C0834E0]

------------------------------

Message: 2 Date: Fri, 2 Jan 2026 03:57:28 -0500 From: Jeffrey Jonas <jeffrey.scott.jonas@gmail.com> To: vcf-midatlantic <vcf-midatlantic@lists.vcfed.org> Subject: [vcf-midatlantic] who was first ... Message-ID: <CAOCNM4SVNENsYTk4+1vTDg32pKqc5Rg7cLkVoUAa-RXEV7vDkw@mail.gmail.com> Content-Type: text/plain; charset="UTF-8"

An interesting article: https://computerhistory.org/blog/the-neverending-quest-for-firsts/ links to https://computerhistory.org/blog/programming-the-eniac-an-example-of-why-com...

Visitors to the Computer History Museum frequently want to know: what was the first computer? The first personal computer? In his recent post, ?Programming the ENIAC: An Example of Why Computer History is Hard <https://computerhistory.org/blog/programming-the-eniac-an-example-of-why-computer-history-is-hard/>,? Computer History Museum Board Chair Len Shustek notes that one of the difficulties of computer history, and indeed, history of technology in general, is the question of ?firsts?: what was the ?first? X, or in alternate form, ?who [first] invented X?? The problem is that for many of these ?firsts,? there is no simple answer, because as Len pointed out, ?What ?first? means depends on precise definitions of fuzzy concepts.? What seem to be questions with easy, factual answers quickly devolve into debates over semantics ...

End of vcf-midatlantic Digest, Vol 92, Issue 2 **********************************************